#  @TroyOnCapital Troy | Following Capital
Troy | Following Capital posts on X about ai, $nbis, gold, money the most. They currently have [---] followers and [---] posts still getting attention that total [---------] engagements in the last [--] hours.
### Engagements: [---------] [#](/creator/twitter::1812298136741896192/interactions)
- [--] Week [---------] +3,242%
- [--] Month [---------] +7,486%
### Mentions: [--] [#](/creator/twitter::1812298136741896192/posts_active)
- [--] Week [--] -74%
- [--] Month [---] +174%
### Followers: [---] [#](/creator/twitter::1812298136741896192/followers)
- [--] Week [---] +18%
- [--] Month [---] +496%
### CreatorRank: [-------] [#](/creator/twitter::1812298136741896192/influencer_rank)
### Social Influence
**Social category influence**
[finance](/list/finance) [stocks](/list/stocks) [technology brands](/list/technology-brands) [cryptocurrencies](/list/cryptocurrencies) [automotive brands](/list/automotive-brands) [countries](/list/countries) [currencies](/list/currencies) [celebrities](/list/celebrities) [social networks](/list/social-networks) [financial services](/list/financial-services)
**Social topic influence**
[ai](/topic/ai), [$nbis](/topic/$nbis) #121, [gold](/topic/gold), [money](/topic/money), [matter](/topic/matter), [inflation](/topic/inflation), [in the](/topic/in-the), [if you](/topic/if-you), [$nvda](/topic/$nvda), [math](/topic/math)
**Top accounts mentioned or mentioned by**
[@moneycruncher](/creator/undefined) [@thealphathought](/creator/undefined) [@kalshi](/creator/undefined) [@kobeissiletter](/creator/undefined) [@brianferoldi](/creator/undefined) [@rajatsoni](/creator/undefined) [@theficouple](/creator/undefined) [@themotleyfool](/creator/undefined) [@stockmktnewz](/creator/undefined) [@polymarket](/creator/undefined) [@cryptofergani](/creator/undefined) [@kinginvestings](/creator/undefined) [@elonmusk](/creator/undefined) [@unusualwhales](/creator/undefined) [@larkdavis](/creator/undefined) [@thestalwart](/creator/undefined) [@robinjbrooks](/creator/undefined) [@lynk0x](/creator/undefined) [@steveonspeed](/creator/undefined) [@nathanielsford](/creator/undefined)
**Top assets mentioned**
[Nebius Group N.V. Class A Ordinary Shares (NBIS)](/topic/$nbis) [NVIDIA Corp. (NVDA)](/topic/$nvda) [Bitcoin (BTC)](/topic/bitcoin) [Tesla, Inc. (TSLA)](/topic/$tsla) [Alphabet Inc Class A (GOOGL)](/topic/$googl) [Iris Energy Limited (IREN)](/topic/$iren) [Microsoft Corp. (MSFT)](/topic/$msft) [Metadium (META)](/topic/$meta) [CatCoin (CAT)](/topic/$cat) [Apple, Inc. (AAPL)](/topic/$aapl) [Strategy (MSTR)](/topic/$mstr) [Amazon.com, Inc. (AMZN)](/topic/$amzn)
### Top Social Posts
Top posts by engagements in the last [--] hours
"@meetblossomapp Great to see $NBIS jumping early [----] I just posted about how $TSLA $NBIS (via Avride) and $NVDA are driving the AI compute race would love your thoughts"
[X Link](https://x.com/TroyOnCapital/status/2007535328530411661) 2026-01-03T19:31Z [--] followers, [--] engagements
"@robo_burro @elonmusk @Tesla Great insights I recently posted about how $TSLA $NBIS (via Avride) and $NVDA are driving the AI compute race would love to hear your thoughts"
[X Link](https://x.com/TroyOnCapital/status/2007535785726357728) 2026-01-03T19:33Z [--] followers, [---] engagements
"@longinvest32 Great points I recently posted about how $TSLA $NBIS (via Avride) and $NVDA are driving the AI compute race would love your thoughts"
[X Link](https://x.com/TroyOnCapital/status/2007536399604691179) 2026-01-03T19:35Z [--] followers, [---] engagements
"Diversification within the AI infrastructure stack is underappreciated. When the core data center layer starts compounding demand having adjacent assets supporting execution and services can matter. Curious how others see the role of these non-core infrastructure revenue streams. https://twitter.com/i/web/status/2007553680825036927 https://twitter.com/i/web/status/2007553680825036927"
[X Link](https://x.com/TroyOnCapital/status/2007553680825036927) 2026-01-03T20:44Z [--] followers, [--] engagements
"@Tpeezy16826 @GrindeOptions Autonomy narratives always lead with software but the gating factor is compute. Training and inference at scale force massive infrastructure build-outs long before revenue shows up. Thats where the real capital gets deployed first"
[X Link](https://x.com/TroyOnCapital/status/2007564519644770705) 2026-01-03T21:27Z [--] followers, [--] engagements
"@DaphneFiel5881 Semiconductor shortages highlight the physical bottlenecks of compute supply but the deeper question is where all that compute will live and how power and cooling will be managed. Thats why Im focused on AI-native data centers and infrastructure that can handle the next wave"
[X Link](https://x.com/TroyOnCapital/status/2007566716252164393) 2026-01-03T21:36Z [--] followers, [--] engagements
"@afterbrainrot Thanks for flagging this. Always wise to verify sources and avoid scams. The real $NBIS story is about AI-native infrastructure and data centersthats where the value lies. Stay focused on fundamentals not pump schemes"
[X Link](https://x.com/TroyOnCapital/status/2007568585657594013) 2026-01-03T21:43Z [--] followers, [---] engagements
"@MarketMatrixs Interesting list Love seeing $NBIS paired with $IREN AI-native infrastructure meets energy solutions. As compute demand grows $NBIS is well-positioned to capture the data center capex. Im bullish on the infrastructure layer"
[X Link](https://x.com/TroyOnCapital/status/2007568894165418022) 2026-01-03T21:44Z [--] followers, [---] engagements
"@TheTechInvest Cant wait for what [----] has in store for $IREN and $NBIS Dont get the hate between the communities - both can thrive"
[X Link](https://x.com/TroyOnCapital/status/2007578080806924418) 2026-01-03T22:21Z [--] followers, [---] engagements
"The real bottleneck isnt just silicon its infrastructure & capital flows scaling that silicon. Raw GPUs address compute Data centers + interconnects scale training & inference Specialized partners like $NBIS / CoreWeave amplify deployment efficiency Thats why $NVDA isnt just dominant it anchors the stack. https://twitter.com/i/web/status/2007593528046825703 https://twitter.com/i/web/status/2007593528046825703"
[X Link](https://x.com/TroyOnCapital/status/2007593528046825703) 2026-01-03T23:22Z [--] followers, [---] engagements
"@danielnewmanUV This is exactly right. Narratives about less compute ignore how AI systems are actually deployed. More models more agents more autonomy = more infrastructure not less. Thats why $NVDA demand stays structural"
[X Link](https://x.com/TroyOnCapital/status/2007595962592805066) 2026-01-03T23:32Z [--] followers, [--] engagements
"The autonomy edge ultimately comes down to training scale not just model design. Vision + data loops require massive efficient compute behind the scenes. Thats why $TSLA AI story still intersects with the broader infrastructure stack $NVDA on compute and names like $NBIS enabling scalable deployment. https://twitter.com/i/web/status/2007597293399331260 https://twitter.com/i/web/status/2007597293399331260"
[X Link](https://x.com/TroyOnCapital/status/2007597293399331260) 2026-01-03T23:37Z [--] followers, [---] engagements
"@MontanaMatos Definitely need to dive deeper on $PATH - main focus has been $NBIS. Ill research more this coming week"
[X Link](https://x.com/TroyOnCapital/status/2007637408540766686) 2026-01-04T02:17Z [--] followers, [---] engagements
"@beauty_oe $NBIS FTW - the company is positioned well with excellent planned growth ahead"
[X Link](https://x.com/TroyOnCapital/status/2007653519789363361) 2026-01-04T03:21Z [--] followers, [----] engagements
"$TSLA is pushing autonomy. $NVDA dominates AI compute. But theres a layer most investors are still missing. The AI infrastructure that actually makes all of this work at scale. Thats where $NBIS quietly sits. Training inference data pipelines and real-world deployment dont happen without it. Were early in the AI infra build-out and this layer gets repriced after autonomy and chips. Most arent paying attention yet. https://twitter.com/i/web/status/2007786542538883391 https://twitter.com/i/web/status/2007786542538883391"
[X Link](https://x.com/TroyOnCapital/status/2007786542538883391) 2026-01-04T12:09Z [--] followers, [---] engagements
"Strong list. One thing that stands out: AI value is starting to bifurcate. App-layer winners like $META compound through distribution but infra names scale through capacity power and utilization not features. Curious which of these you think benefits most from AI demand staying supply-constrained into [----] https://twitter.com/i/web/status/2007851672119173351 https://twitter.com/i/web/status/2007851672119173351"
[X Link](https://x.com/TroyOnCapital/status/2007851672119173351) 2026-01-04T16:28Z [--] followers, [---] engagements
"The way to think about this is by layers not just sectors. Compute: $NVDA Autonomy / platforms: $TSLA Infrastructure & deployment: $NBIS $IREN Software orchestration: $PATH $ZETA Different layers get priced at different times as AI moves from training to real-world scale. https://twitter.com/i/web/status/2007882570000957838 https://twitter.com/i/web/status/2007882570000957838"
[X Link](https://x.com/TroyOnCapital/status/2007882570000957838) 2026-01-04T18:31Z [--] followers, [---] engagements
"Lists like this are useful but the real edge comes from understanding why each layer matters. Chips ( $NVDA ) enable models Autonomy ( $TSLA ) drives applications Infrastructure ( $NBIS $IREN ) determines whether any of it can scale reliably The bottleneck layer usually gets repriced last. https://twitter.com/i/web/status/2007883122260807783 https://twitter.com/i/web/status/2007883122260807783"
[X Link](https://x.com/TroyOnCapital/status/2007883122260807783) 2026-01-04T18:33Z [--] followers, [---] engagements
"This is the key point most people miss. ASICs win on cost only after workloads stabilize. Until then GPUs dominate because flexibility matters more than efficiency. That also means demand stays supply-constrained longer than models assume especially once inference scales. https://twitter.com/i/web/status/2007915033779654958 https://twitter.com/i/web/status/2007915033779654958"
[X Link](https://x.com/TroyOnCapital/status/2007915033779654958) 2026-01-04T20:40Z [--] followers, [--] engagements
"@rohanpaul_ai The hidden constraint isnt data its cost-effective reliable inference at scale. Most AI-first enterprise products stall not on features but when latency utilization and margins collide in production"
[X Link](https://x.com/TroyOnCapital/status/2007915473133088952) 2026-01-04T20:42Z [---] followers, [---] engagements
"When Elon says singularity hes not talking sci-fi consciousness. Hes talking about the point where AI progress becomes self-reinforcing faster models driving more deployment which drives more data which accelerates improvement. The bottleneck isnt intelligence anymore. Its compute power and cost at scale. https://twitter.com/i/web/status/2007917302847869172 https://twitter.com/i/web/status/2007917302847869172"
[X Link](https://x.com/TroyOnCapital/status/2007917302847869172) 2026-01-04T20:49Z [--] followers, [--] engagements
"@elonmusk @DavidSHolz Singularity isnt intelligence its feedback loops hitting physical limits (power capex latency). Thats where timelines stretch"
[X Link](https://x.com/TroyOnCapital/status/2007951868144533592) 2026-01-04T23:06Z [--] followers, [--] engagements
"@slow_developer This is exactly right. Abundance in one layer creates scarcity elsewhere. Were trading physical limits for coordination energy and attention constraints. The climb didnt end it changed shape"
[X Link](https://x.com/TroyOnCapital/status/2007976670452756693) 2026-01-05T00:45Z [---] followers, [--] engagements
"Core thesis on $NBIS: AI value doesnt accrue at models it accrues at production-grade inference. $NBIS isnt generic cloud. Its GPU-first infrastructure built for deterministic latency predictable costs and sustained throughput the requirements for AI at industrial scale. Microsoft & Meta dont experiment here they sign long-term capacity contracts. Thats validation of durability not hype. Risk check: capital intensity + execution. Mitigated by pre-sold capacity tier-1 customers and demand outpacing supply. High beta yes but thats where asymmetric infra returns come from."
[X Link](https://x.com/TroyOnCapital/status/2007991033297297878) 2026-01-05T01:42Z [--] followers, [---] engagements
"The hard part wont be adoption itll be coordination. Were moving from scarcity of intelligence to scarcity of alignment across energy labor institutions and capital allocation. The bottleneck shifts from can we build it to can society absorb it without breaking feedback loops https://twitter.com/i/web/status/2007997671177388106 https://twitter.com/i/web/status/2007997671177388106"
[X Link](https://x.com/TroyOnCapital/status/2007997671177388106) 2026-01-05T02:08Z [--] followers, [--] engagements
"@annapanart @Dr_Singularity Infinite choice without constraint doesnt produce freedom it produces paralysis. The next era isnt about removing limits its about deciding which constraints matter. Infrastructure energy and attention become the real governors of freedom"
[X Link](https://x.com/TroyOnCapital/status/2007997875318653222) 2026-01-05T02:09Z [--] followers, [--] engagements
"@WatcherGuru New ATHs arent about headlines theyre about positioning and flows. ETFs passive demand and short-dated leverage create pressure before the breakout. Price moves first. Narratives follow"
[X Link](https://x.com/TroyOnCapital/status/2007998971952992764) 2026-01-05T02:13Z [--] followers, [--] engagements
"@TheSonOfWalkley $TSLA price doesnt move on vibes it moves on positioning and flows. When momentum flips passive + options hedging do the heavy lifting. Fundamentals get debated after price"
[X Link](https://x.com/TroyOnCapital/status/2007999967915573354) 2026-01-05T02:17Z [--] followers, [---] engagements
"Markets arent reacting to AI hype here theyre reacting to execution under uncertainty. The signal isnt Palantir per se its that AI-native decision systems are now embedded in real-world ops where latency coordination and reliability matter. Thats where AI value actually accrues. https://twitter.com/i/web/status/2008007965383778322 https://twitter.com/i/web/status/2008007965383778322"
[X Link](https://x.com/TroyOnCapital/status/2008007965383778322) 2026-01-05T02:49Z [--] followers, [--] engagements
"The key mistake bears make is modeling $NBIS like a commodity GPU renter or a REIT. $NBIS is closer to an AI factory operator: Revenue mixes long-term contracts + variable/spot compute Margins come from orchestration utilization and reliability not just GPUs Power + grid access is the real moat ownership structure is secondary CapEx is the entry ticket. Returns come from sustained production-grade inference at scale. https://twitter.com/i/web/status/2008013550292545988 https://twitter.com/i/web/status/2008013550292545988"
[X Link](https://x.com/TroyOnCapital/status/2008013550292545988) 2026-01-05T03:11Z [--] followers, 15.5K engagements
"@beauty_oe This point about concentration vs dispersion is important. Modeling $NBIS correctly means understanding it as an operator not an asset holder returns come from utilization orchestration and reliability over time not just GPU ownership"
[X Link](https://x.com/TroyOnCapital/status/2008027750196781440) 2026-01-05T04:08Z [--] followers, [---] engagements
"GPU price hikes expose who actually has operating leverage vs. financial leverage. NeoClouds that win on access alone feel the pain immediately when input costs rise. Operators that win on utilization power discipline and reliability can absorb it. Short-term noise. Long-term signal especially for $NBIS. https://twitter.com/i/web/status/2008150971797377121 https://twitter.com/i/web/status/2008150971797377121"
[X Link](https://x.com/TroyOnCapital/status/2008150971797377121) 2026-01-05T12:17Z [---] followers, [---] engagements
"Thats a fair nuance the build-out phase is where cost shocks actually matter most. The separation I keep coming back to is who bears that pressure: balance sheet leverage vs. operational leverage. Capex-heavy platforms feel memory inflation upfront but the ones with tighter deployment discipline and faster utilization ramps compress that pain window quickly. The weaker operators carry it forward as permanent margin drag. Same cost environment very different outcomes. https://twitter.com/i/web/status/2008160758119686470 https://twitter.com/i/web/status/2008160758119686470"
[X Link](https://x.com/TroyOnCapital/status/2008160758119686470) 2026-01-05T12:56Z [--] followers, [--] engagements
"$NBIS works if you frame it as execution capacity not a chart pattern. The real upside comes from contracted AI factories power access and sustained inference throughput not multiple expansion alone. If utilization ramps as expected price follows fundamentals not trendlines. https://twitter.com/i/web/status/2008213835254296935 https://twitter.com/i/web/status/2008213835254296935"
[X Link](https://x.com/TroyOnCapital/status/2008213835254296935) 2026-01-05T16:27Z [--] followers, [---] engagements
"Agree multiples usually move first. My point is less when price moves and more what determines whether that expansion holds. In AI infra execution capacity (power contracted demand sustained throughput) is what turns early multiple expansion into durable value rather than a round-trip. https://twitter.com/i/web/status/2008218896814489748 https://twitter.com/i/web/status/2008218896814489748"
[X Link](https://x.com/TroyOnCapital/status/2008218896814489748) 2026-01-05T16:47Z [--] followers, [--] engagements
"This framing is spot on. The key insight most people miss: tokens are the unit of production for AI. Once you see AI as factories producing tokens the value accrues to whoever can deliver predictable throughput cost and uptime at scale not just flashy models. Thats the real $NBIS story. https://twitter.com/i/web/status/2008244354037743771 https://twitter.com/i/web/status/2008244354037743771"
[X Link](https://x.com/TroyOnCapital/status/2008244354037743771) 2026-01-05T18:29Z [--] followers, [---] engagements
"The bigger takeaway isnt the automaker its the architecture. Scaled autonomy is fundamentally an inference-at-scale problem: continuous perception low-latency decisioning and massive uptime requirements. Avride Uber is a demand signal for AI factories that can deliver predictable throughput not just training bursts. Thats where the real leverage sits. https://twitter.com/i/web/status/2008246928644178407 https://twitter.com/i/web/status/2008246928644178407"
[X Link](https://x.com/TroyOnCapital/status/2008246928644178407) 2026-01-05T18:39Z [--] followers, [---] engagements
"@munster_gene The updates matter but the real pivot is underneath them. [----] stops rewarding announcements and starts rewarding sustained inference revenue utilization throughput and cost control. Thats the shift from AI trade to AI industrialization"
[X Link](https://x.com/TroyOnCapital/status/2008250250772066596) 2026-01-05T18:52Z [--] followers, [--] engagements
"@DanInvestingss Analyst PTs reflect capacity not hype. $NBIS is supply-constrained with long-term committed contracts and production-grade inference demand. This isnt optional AI spend its infrastructure. Multiple expansion follows utilization 📈"
[X Link](https://x.com/TroyOnCapital/status/2008292906675716434) 2026-01-05T21:41Z [--] followers, [---] engagements
"This matters less for the silicon and more for the deployment model. NVL72 isnt about peak benchmarks its about dense reliable inference at scale. The real winners are clouds that can integrate hardware networking power and scheduling as one system. Thats where physical + agentic AI actually becomes usable. https://twitter.com/i/web/status/2008319577122636050 https://twitter.com/i/web/status/2008319577122636050"
[X Link](https://x.com/TroyOnCapital/status/2008319577122636050) 2026-01-05T23:27Z [--] followers, [---] engagements
"@EndicottInvests Same. What stands out is execution not headlines. Turning next-gen GPUs into production inference predictable cost latency and throughput is where real value gets created. Thats the compounding edge for $NBIS alongside $NVDA"
[X Link](https://x.com/TroyOnCapital/status/2008357046865789191) 2026-01-06T01:56Z [--] followers, [---] engagements
"That timeline makes sense. The hard part isnt raw silicon its making hardware networking and inference software behave as one deterministic system at scale. That integration window is where most of the real work (and differentiation) happens for platforms like $NVDA and for operators like $NBIS downstream. https://twitter.com/i/web/status/2008358602600563039 https://twitter.com/i/web/status/2008358602600563039"
[X Link](https://x.com/TroyOnCapital/status/2008358602600563039) 2026-01-06T02:02Z [--] followers, [--] engagements
"Its less about using NVIDIA and more about where autonomy value concentrates. Most OEMs wont vertically integrate full-stack autonomy theyll standardize on a compute + tooling layer and differentiate above it. That dynamic strongly favors $NVDA regardless of who wins at the vehicle brand level. https://twitter.com/i/web/status/2008359382090690686 https://twitter.com/i/web/status/2008359382090690686"
[X Link](https://x.com/TroyOnCapital/status/2008359382090690686) 2026-01-06T02:06Z [--] followers, [--] engagements
"@PatrickMoorhead The interesting part isnt peak specs its whether this rack can sustain full-speed data movement at scale. In production network topology power density and cooling efficiency end up determining realized throughput far more than headline FLOPs"
[X Link](https://x.com/TroyOnCapital/status/2008375541137605057) 2026-01-06T03:10Z [--] followers, [---] engagements
"This is the key point many miss: Rubin NVL72 isnt just more GPUs it collapses system-level bottlenecks. [--] GPUs acting as one NVLink domain + HBM4 means: deterministic latency predictable cost per token sustained production-grade inference Thats why this matters for $NBIS specifically theyre building AI factories not demo clusters. Rubin turns inference from bursty industrial. Thats where real value accrues. 🏗🧠 https://twitter.com/i/web/status/2008391623873622450 https://twitter.com/i/web/status/2008391623873622450"
[X Link](https://x.com/TroyOnCapital/status/2008391623873622450) 2026-01-06T04:14Z [--] followers, [---] engagements
"@jukan05 The real takeaway isnt peak specs its whether rack-scale systems can sustain utilization. CES announcements raise expectations. Markets will reprice on power density cooling memory bandwidth and software orchestration once these systems hit steady-state"
[X Link](https://x.com/TroyOnCapital/status/2008504271806099570) 2026-01-06T11:41Z [--] followers, [--] engagements
"@unusual_whales This is the quiet part of the AI trade. GPUs scale demand instantly. Memory supply doesnt. When DRAM and HBM availability gate production AI economics stop being about chip headlines and start being about system-level constraints and utilization"
[X Link](https://x.com/TroyOnCapital/status/2008504793376190544) 2026-01-06T11:43Z [--] followers, [--] engagements
"@simonkalouche Demos are the easy part. The hard part is deploying robots that run continuous inference with tight latency power and uptime requirements. Thats where AI shifts from spectacle to infrastructure economics"
[X Link](https://x.com/TroyOnCapital/status/2008505014508228979) 2026-01-06T11:44Z [--] followers, [--] engagements
"@nvidia Specs are impressive but the real differentiator will be steady-state performance. Sustained inference memory bandwidth under load power density and uptime will decide cost per token in production not peak benchmarks"
[X Link](https://x.com/TroyOnCapital/status/2008509835370000448) 2026-01-06T12:03Z [--] followers, [--] engagements
"Rubin being in production is the key point most people are glossing over. The gap isnt about peak benchmarks anymore its about who can deliver reliable cost-predictable inference at scale. Once youre shipping systems that actually run 24/7 networking memory power and software integration matter more than model demos. Thats why the competitive distance keeps widening. $NVDA https://twitter.com/i/web/status/2008511035201003719 https://twitter.com/i/web/status/2008511035201003719"
[X Link](https://x.com/TroyOnCapital/status/2008511035201003719) 2026-01-06T12:08Z [--] followers, [--] engagements
"@mvcinvesting This is real capacity not a press-release KPI. 80MW = tangible revenue runway long-dated visibility and proof NBIS is executing ahead of demand not chasing it. 10x footprint expansion in [--] year tells you everything about where customer pull actually is. $NBIS"
[X Link](https://x.com/TroyOnCapital/status/2008517518147952929) 2026-01-06T12:34Z [--] followers, [---] engagements
"The marketplace angle matters less as a headline and more as a utilization lever. If $NBIS can drive third-party demand onto already-built capacity thats margin expansion without proportional capex. Infrastructure stories dont reprice on announcements they reprice when utilization inflects. https://twitter.com/i/web/status/2008527069131710532 https://twitter.com/i/web/status/2008527069131710532"
[X Link](https://x.com/TroyOnCapital/status/2008527069131710532) 2026-01-06T13:12Z [--] followers, [---] engagements
"@ArtByALKEBULAN @StockMKTNewz It doesnt change what $BTC is it changes who can access it. ETFs abstract custody and rails not the protocol. That likely strengthens the monetary thesis while weakening the ideological purity. Tradeoff not contradiction"
[X Link](https://x.com/TroyOnCapital/status/2008537689604452362) 2026-01-06T13:54Z [---] followers, [--] engagements
"@PolymarketMoney Being early on $NVDA Rubin matters but the real edge is execution. H2 [----] is about lining up power cooling networking and software maturity so capacity actually ships into production not just installs. For $NBIS hardware access is table stakes; delivery is the moat"
[X Link](https://x.com/TroyOnCapital/status/2008539901793538152) 2026-01-06T14:03Z [--] followers, [---] engagements
"@KobeissiLetter The dispersion matters. Revenue is scaling faster than deployable capacity. The bottleneck isnt demand its power cooling networking and time to production. Thats why infra growth looks lumpy while model revenue looks exponential"
[X Link](https://x.com/TroyOnCapital/status/2008548180015894978) 2026-01-06T14:36Z [--] followers, [--] engagements
"@EricBalchunas @TheBlock__ This is less about crypto adoption and more about distribution control. ETFs dont change the protocol they change who allocates how size moves and where flows get warehoused. Morgan Stanley internalizing that demand is a business decision not an ideological one"
[X Link](https://x.com/TroyOnCapital/status/2008555206947242424) 2026-01-06T15:04Z [--] followers, [--] engagements
"@AMD The question isnt peak TOPS its whether on-device NPUs meaningfully reduce cloud inference demand or just shift latency-sensitive workloads locally. Most real revenue still lives where utilization stays high"
[X Link](https://x.com/TroyOnCapital/status/2008558369230184889) 2026-01-06T15:16Z [--] followers, [---] engagements
"This is the key shift. Agentic workloads dont just want faster compute they want persistent state and bandwidth at scale. HBM matters but the real arb is who can keep memory-heavy systems utilized day after day. Thats where infra operators with secured power + long-term contracts quietly win. $NVDA $MU $NBIS https://twitter.com/i/web/status/2008563776577302941 https://twitter.com/i/web/status/2008563776577302941"
[X Link](https://x.com/TroyOnCapital/status/2008563776577302941) 2026-01-06T15:38Z [--] followers, [---] engagements
"@DeItaone [---] compute headlines miss the real constraint. The bottleneck isnt demand its power networking and cost per inference-hour. If compute needs to grow [---] it forces architectural efficiency gains not just more chips. Thats where infra discipline matters. $AMD $NVDA"
[X Link](https://x.com/TroyOnCapital/status/2008572711573094851) 2026-01-06T16:13Z [--] followers, [---] engagements
"Which works better depends on scale and economics not just raw sensing. Lidar adds redundancy but it also adds cost calibration overhead and new failure modes at fleet scale. Teslas bet is that massive data + software generalization can outperform hardware redundancy per dollar deployed. Thats the real tradeoff. https://twitter.com/i/web/status/2008604863937347856 https://twitter.com/i/web/status/2008604863937347856"
[X Link](https://x.com/TroyOnCapital/status/2008604863937347856) 2026-01-06T18:21Z [--] followers, [--] engagements
"This is a great example of optimization shifting from peak performance to sustained utilization. Warm-water liquid cooling isnt about eliminating cooling capex its about improving energy efficiency density and uptime at scale. Less energy spent on heat removal = more energy available for compute which matters far more in steady-state AI factories. The constraint just moves up the stack. https://twitter.com/i/web/status/2008605993031397415 https://twitter.com/i/web/status/2008605993031397415"
[X Link](https://x.com/TroyOnCapital/status/2008605993031397415) 2026-01-06T18:26Z [---] followers, [----] engagements
"Lower cost per token and fewer GPUs per model dont mean less compute demand they usually expand it. As inference gets cheaper usage explodes workloads persist longer and utilization rises. Rubin is about shifting the curve from peak training events to continuous large-scale inference which ultimately drives more sustained demand for $NVDA compute not less. https://twitter.com/i/web/status/2008606493084709281 https://twitter.com/i/web/status/2008606493084709281"
[X Link](https://x.com/TroyOnCapital/status/2008606493084709281) 2026-01-06T18:28Z [--] followers, [---] engagements
"The Roomba comparison highlights the core difference. Lidar works extremely well in constrained low-speed closed environments with limited edge cases. Driving is the opposite: open world long tail high speed adversarial conditions and constant distribution shift. Thats why the debate isnt does lidar work its which approach scales economically and operationally to millions of vehicles. https://twitter.com/i/web/status/2008608499706851532 https://twitter.com/i/web/status/2008608499706851532"
[X Link](https://x.com/TroyOnCapital/status/2008608499706851532) 2026-01-06T18:35Z [--] followers, [--] engagements
"This reflects a broader shift in AI infrastructure: the unit of competition has moved from chips racks systems. At yotta-scale peak FLOPs matter less than sustained utilization networking power delivery cooling and the ability to evolve across generations. Thats where rack-level modular design becomes a strategic advantage. https://twitter.com/i/web/status/2008616617375322328 https://twitter.com/i/web/status/2008616617375322328"
[X Link](https://x.com/TroyOnCapital/status/2008616617375322328) 2026-01-06T19:08Z [--] followers, [---] engagements
"Efficiency gains dont remove the constraint they raise the ceiling. Lower power per GPU enables higher density and utilization not lower total demand. Operators reinvest those gains into more GPUs per rack higher duty cycles and tighter clustering. Cooling savings matter but the steady-state outcome is still power-bound AI factories just with better economics faster scale-up and higher ROI per MW. Thats bullish for neo-clouds like $NBIS because it improves margins and accelerates capacity growth not because energy stops being the limiter. https://twitter.com/i/web/status/2008621353713487875"
[X Link](https://x.com/TroyOnCapital/status/2008621353713487875) 2026-01-06T19:27Z [--] followers, [--] engagements
"Flat price action doesnt signal fading demand it signals investors anchoring to cycle timing instead of steady-state utilization. Blackwell + Rubin push the constraint from can we train to how efficiently can we run at scale. Thats about power density cooling and uptime not peak benchmarks. AI hardware isnt a [----] trade. Its a multi-year capacity absorption story. $NVDA just keeps moving the bottleneck up the stack. https://twitter.com/i/web/status/2008632074614247430 https://twitter.com/i/web/status/2008632074614247430"
[X Link](https://x.com/TroyOnCapital/status/2008632074614247430) 2026-01-06T20:09Z [--] followers, [---] engagements
"Breaking above $100 isnt about momentum its about validation. $NBIS is positioning for sustained utilization not peak hype. Early access to $NVDA Rubin NVL72 = higher density better power efficiency and long-duration enterprise demand. AI factories reward uptime and scale not one-day candles. https://twitter.com/i/web/status/2008654445048185221 https://twitter.com/i/web/status/2008654445048185221"
[X Link](https://x.com/TroyOnCapital/status/2008654445048185221) 2026-01-06T21:38Z [--] followers, [---] engagements
"Most advanced isnt about demos or model size its about deployment. $TSLA's edge is running one autonomy stack across millions of vehicles continuously training on real-world edge cases and closing the loop in production. That scale + feedback cycle is extremely hard to replicate. https://twitter.com/i/web/status/2008671299334484310 https://twitter.com/i/web/status/2008671299334484310"
[X Link](https://x.com/TroyOnCapital/status/2008671299334484310) 2026-01-06T22:45Z [--] followers, [--] engagements
"The shift away from specs makes sense. In [----] the differentiator isnt peak performance its useful performance how efficiently systems deliver outcomes under real workloads. AI success is becoming about deployment power efficiency latency and integration into workflows not raw TOPS on a slide. https://twitter.com/i/web/status/2008678299862790384 https://twitter.com/i/web/status/2008678299862790384"
[X Link](https://x.com/TroyOnCapital/status/2008678299862790384) 2026-01-06T23:13Z [---] followers, [--] engagements
"The key point isnt SSDs replacing HBM its memory hierarchy expansion. HBM stays the hot tier. What Jensen is signaling is that context + state are spilling beyond HBM forcing fast shared non-volatile memory layers to keep GPUs utilized. This is about sustaining throughput not substituting HBM. https://twitter.com/i/web/status/2008679349848076546 https://twitter.com/i/web/status/2008679349848076546"
[X Link](https://x.com/TroyOnCapital/status/2008679349848076546) 2026-01-06T23:17Z [--] followers, [---] engagements
"This is exactly what time-shifting power looks like in practice. 80MW/320MWh isnt about energy arbitrage its about grid stability under peak and intermittent load. Storage like this is becoming critical as AI EVs and renewables all stress the same infrastructure. Megapack is quietly turning power into schedulable capacity. $TSLA https://twitter.com/i/web/status/2008686210802479182 https://twitter.com/i/web/status/2008686210802479182"
[X Link](https://x.com/TroyOnCapital/status/2008686210802479182) 2026-01-06T23:44Z [--] followers, [--] engagements
"@AwareWealth Exactly. Inference density is where utilization power efficiency and latency economics matter most. Thats where platform + ecosystem winners compound not just model trainers. $NBIS $NVDA"
[X Link](https://x.com/TroyOnCapital/status/2008689986556883239) 2026-01-06T23:59Z [--] followers, [---] engagements
"This is the part people miss: autonomy isnt a one-time training run its a continuous power-dense compute problem. $10B on GPUs just to train underscores that the bottleneck is sustained throughput energy efficiency and uptime not model cleverness. Thats why Tesla blends external GPUs with custom silicon: the constraint shifts to steady-state infrastructure economics. https://twitter.com/i/web/status/2008707539555545134 https://twitter.com/i/web/status/2008707539555545134"
[X Link](https://x.com/TroyOnCapital/status/2008707539555545134) 2026-01-07T01:09Z [--] followers, [---] engagements
"@rohanpaul_ai Cool demo but the hard part isnt gesture recognition its robustness at scale. Hands-as-input break down under occlusion lighting variance latency and multi-user ambiguity. Until input is reliable across environments this stays a novelty UI not a workflow primitive"
[X Link](https://x.com/TroyOnCapital/status/2008712583424208907) 2026-01-07T01:29Z [--] followers, [--] engagements
"@jenzhuscott @SharpaRobotics Hand dexterity matters but only if it holds up in steady-state. The real test isnt peak DOF or sensing density its calibration drift wear latency and failure recovery over millions of cycles. Thats where most robotic hands still struggle to leave the lab"
[X Link](https://x.com/TroyOnCapital/status/2008713606234542274) 2026-01-07T01:33Z [--] followers, [---] engagements
"The metaphor works but the constraint doesnt disappear it just moves. GPUs create the gold but warehouses only matter if you can afford to keep it: power bandwidth retrieval latency and lifecycle costs dominate at scale. In steady state storage growth is as much an energy and ops problem as a hardware one. https://twitter.com/i/web/status/2008728662213095582 https://twitter.com/i/web/status/2008728662213095582"
[X Link](https://x.com/TroyOnCapital/status/2008728662213095582) 2026-01-07T02:33Z [--] followers, [---] engagements
"@rohanpaul_ai This looks structural not cyclical. AI servers using 810x more DRAM shift the bottleneck from fabs to allocation as capacity moves to HBM/DDR5. Consumer price spikes are just the downstream symptom of AI factories absorbing steady-state supply"
[X Link](https://x.com/TroyOnCapital/status/2008754212927369345) 2026-01-07T04:14Z [---] followers, [--] engagements
"This resonates. Tools compress iteration in domains you already understand they dont eliminate the need for judgment. The gap between feels faster and can be trusted end-to-end is where most of the hype breaks. Productivity gains are real but theyre uneven and domain-dependent not exponential across the board. https://twitter.com/i/web/status/2008758119820828873 https://twitter.com/i/web/status/2008758119820828873"
[X Link](https://x.com/TroyOnCapital/status/2008758119820828873) 2026-01-07T04:30Z [--] followers, [---] engagements
"This looks less like users switching and more like where the query originates. Gemini inherits traffic from Search Chrome Android Workspace ChatGPT traffic is intentional app-level behavior. So share shifts reflect distribution plumbing not necessarily model preference or depth of usage. Different surfaces different workloads https://twitter.com/i/web/status/2008861654428369029 https://twitter.com/i/web/status/2008861654428369029"
[X Link](https://x.com/TroyOnCapital/status/2008861654428369029) 2026-01-07T11:21Z [--] followers, [--] engagements
"This feels like a broader RL lesson not just diffusion: dont globally constrain the policy to protect against proxy reward failure. Apply constraints selectively where reward uncertainty spikes and let the rest of the distribution explore. Much better trade-off between quality diversity and learning speed. https://twitter.com/i/web/status/2008862023761998078 https://twitter.com/i/web/status/2008862023761998078"
[X Link](https://x.com/TroyOnCapital/status/2008862023761998078) 2026-01-07T11:23Z [--] followers, [--] engagements
"@TheRayMyers The competitive question isnt peak performance its sustained deployment. Software maturity memory bandwidth power efficiency and ecosystem lock-in decide who actually scales inference at volume. Thats where the gap usually widens not narrows"
[X Link](https://x.com/TroyOnCapital/status/2008865947369807979) 2026-01-07T11:38Z [--] followers, [---] engagements
"@WatcherGuru Spot ETFs matter less for short-term price than for long-duration demand stability. They turn ETH from a trading asset into balance-sheet infrastructure predictable flows lower volatility and better capital planning upstream. Thats when ecosystems actually scale"
[X Link](https://x.com/TroyOnCapital/status/2008866540410786079) 2026-01-07T11:41Z [--] followers, [--] engagements
"@BillAckman The underappreciated part is that compounding only reveals value late it doesnt create it late. Most of the real work happens quietly in the flat years: survival reinvestment discipline and avoiding fatal mistakes. Miss that and the curve never shows up"
[X Link](https://x.com/TroyOnCapital/status/2008867191366824143) 2026-01-07T11:43Z [--] followers, [--] engagements
"@jukan05 This looks less like demand softness and more like hard capacity triage. CoWoS and leading-edge nodes are being reallocated across higher-priority internal and external programs capping volumes even with interest intact. Constraint is upstream not workload. $MRVL"
[X Link](https://x.com/TroyOnCapital/status/2008867821208695111) 2026-01-07T11:46Z [---] followers, [---] engagements
"Index-level multiples can be misleading here. Whats likely isnt a clean tech vs defensives rotation but dispersion within tech capital-light vs capital-intensive pricing power vs reinvestment-heavy models. At 23x the index may be capped but select cash-generative platforms still compound while others de-rate. https://twitter.com/i/web/status/2008873161090036210 https://twitter.com/i/web/status/2008873161090036210"
[X Link](https://x.com/TroyOnCapital/status/2008873161090036210) 2026-01-07T12:07Z [--] followers, [---] engagements
"The headline growth is impressive but the key question is mix and durability. [----] deals tells you adoption breadth the next signal is ACV expansion renewal rates and how fast Agentforce attaches to core CRM seats. If this converts from pilots to scaled deployments the ARR trajectory matters a lot more than the initial YoY print. https://twitter.com/i/web/status/2008874271515595150 https://twitter.com/i/web/status/2008874271515595150"
[X Link](https://x.com/TroyOnCapital/status/2008874271515595150) 2026-01-07T12:12Z [--] followers, [--] engagements
"AI factories dont win by peak speed they win by throughput and uptime. A [--] training speedup only matters if the factory can stay fed: memory bandwidth interconnect power & cooling sustained utilization Thats why the real moat isnt just GPUs its the system. $NVDA $MU $SMCI $VRT https://twitter.com/i/web/status/2008880936901833004 https://twitter.com/i/web/status/2008880936901833004"
[X Link](https://x.com/TroyOnCapital/status/2008880936901833004) 2026-01-07T12:38Z [--] followers, [--] engagements
"@CyberRobooo Exactly and matching human sensing isnt enough. The bar is repeatability at scale: low noise low drift cheap sensors that survive millions of contacts. Until tactile stacks hit that reliability curve utilization not intelligence stays the binding constraint"
[X Link](https://x.com/TroyOnCapital/status/2008888438733578424) 2026-01-07T13:08Z [---] followers, [--] engagements
"This is less about AI taking over and more about vertical integration. Proxy voting is a high-volume low-novelty decision pipeline exactly where internal models beat generic external advisors. The real shift is firms owning their own governance logic instead of outsourcing it. https://twitter.com/i/web/status/2008889919347757493 https://twitter.com/i/web/status/2008889919347757493"
[X Link](https://x.com/TroyOnCapital/status/2008889919347757493) 2026-01-07T13:14Z [--] followers, [---] engagements
"@danielnewmanUV @haslindatv CES turning into a chips show makes sense ROI is now decided below the model layer. The real question isnt peak performance its utilization: power cooling networking and uptime at scale. Thats where AI capex either compounds or stalls"
[X Link](https://x.com/TroyOnCapital/status/2008896977959104815) 2026-01-07T13:42Z [--] followers, [--] engagements
"@jukan05 Both can be true. HBM4 readiness can push early ramp timing by weeks without changing full-year shipment intent. Near-term cadence noise vs steady-state volume Jensen is talking end state JPM is flagging ramp friction"
[X Link](https://x.com/TroyOnCapital/status/2008899680995082509) 2026-01-07T13:53Z [---] followers, [---] engagements
"Worth being careful on attribution here. Hiring is slowing less because AI is absorbing jobs and more because companies are optimizing for utilization and margins after years of headcount expansion. AI lets output grow without proportional hiring but thats an efficiency story first - not a demand collapse signal. https://twitter.com/i/web/status/2008901300889481245 https://twitter.com/i/web/status/2008901300889481245"
[X Link](https://x.com/TroyOnCapital/status/2008901300889481245) 2026-01-07T13:59Z [--] followers, [--] engagements
"@AwareWealth The market keeps calling this uncertainty. It isnt. Capacity is already sold. Revenue just hasnt been recognized yet. That gap isnt risk - its an accounting delay. The only real question is how fast they can build. $NBIS"
[X Link](https://x.com/TroyOnCapital/status/2008908893028786195) 2026-01-07T14:29Z [--] followers, [---] engagements
"@SawyerMerritt Q4 unit volatility matters less than what GM is actually building. This reads like: model transitions launch timing noise mix shifting upmarket The real question isnt EV units its software attach pricing power and margins over the next [--] months"
[X Link](https://x.com/TroyOnCapital/status/2008930569598754971) 2026-01-07T15:55Z [--] followers, [--] engagements
"@caro_milanesi @Qualcomm @HP Exactly. AI value isnt the model its sustained outcomes under real constraints. Performance-per-watt thermals and battery life decide whether AI is always-on or just a demo feature. Devices are where the rubber meets the road"
[X Link](https://x.com/TroyOnCapital/status/2008931913504112670) 2026-01-07T16:01Z [--] followers, [--] engagements
"This list shows where $NVDA's real leverage sits. Its not just selling GPUs its embedding itself across the AI stack. $CRWV for utilization + demand pull $APLD for power + data center leverage $ARM for long-cycle architecture control $NVDA is underwriting the ecosystem not just supplying it. https://twitter.com/i/web/status/2008951885944717740 https://twitter.com/i/web/status/2008951885944717740"
[X Link](https://x.com/TroyOnCapital/status/2008951885944717740) 2026-01-07T17:20Z [--] followers, [---] engagements
"@DekmarTrades Right - NVDA isnt trading its underwriting the AI stack. Thats the real signal"
[X Link](https://x.com/TroyOnCapital/status/2008955322975899961) 2026-01-07T17:34Z [---] followers, [--] engagements
"@NVIDIARobotics @CaterpillarInc This isnt a gimmick - its AI turning labor into throughput. Industrial AI monetizes uptime safety and utilization first. Thats where durable margins live. $NVDA $CAT"
[X Link](https://x.com/TroyOnCapital/status/2008965697091539098) 2026-01-07T18:15Z [--] followers, [---] engagements
"@NVIDIARobotics This is the real bottleneck in robotics: not model intelligence but physically accurate simulation at scale. If sim synthetic data deployment isnt tightly integrated iteration speed collapses. OSMO is about throughput and reliability not flash. $NVDA"
[X Link](https://x.com/TroyOnCapital/status/2008992205126685163) 2026-01-07T20:00Z [--] followers, [--] engagements
"This is the real damage of stop-start export policy: it breaks training continuity. Large-scale training and inference planning depends on predictable access to the same GPU class. Uncertainty forces over-stockpiling fragmented clusters and lower utilization - which hurts everyone not just buyers. https://twitter.com/i/web/status/2009001609104355563 https://twitter.com/i/web/status/2009001609104355563"
[X Link](https://x.com/TroyOnCapital/status/2009001609104355563) 2026-01-07T20:38Z [--] followers, [---] engagements
"The key issue isnt insolvency headlines - its maturity mismatch. AI cloud models front-load capex (GPUs power leases) and back-load cash flow. When demand growth slows or financing tightens continuity breaks fast. Thats the real risk: utilization gaps + refinancing risk not demand disappearing. https://twitter.com/i/web/status/2009034220015067568 https://twitter.com/i/web/status/2009034220015067568"
[X Link](https://x.com/TroyOnCapital/status/2009034220015067568) 2026-01-07T22:47Z [---] followers, [----] engagements
"AWS is the easy headline - the harder part is economics. Autonomous trucking scales only if: training + simulation costs stay predictable inference latency is consistent on the edge utilization stays high across fleets Cloud partnerships help with iteration speed but autonomy wins or loses on steady-state operations not demos. https://twitter.com/i/web/status/2009035080132616376 https://twitter.com/i/web/status/2009035080132616376"
[X Link](https://x.com/TroyOnCapital/status/2009035080132616376) 2026-01-07T22:51Z [--] followers, [--] engagements
"On-device AI is where utilization finally matters more than peak FLOPs. Edge systems live and die by: power envelopes thermal headroom sustained throughput predictable latency Thats why embedded NPUs + tightly integrated stacks win - not because theyre flashy but because they stay on in real-world conditions. $AMD https://twitter.com/i/web/status/2009037213057863999 https://twitter.com/i/web/status/2009037213057863999"
[X Link](https://x.com/TroyOnCapital/status/2009037213057863999) 2026-01-07T22:59Z [--] followers, [--] engagements
"@NVIDIADC Vera isnt about peak CPU performance - its about predictability. As clusters scale control-plane latency and scheduling efficiency decide utilization. This is NVIDIA hardening the economics of agentic AI not chasing benchmarks. $NVDA"
[X Link](https://x.com/TroyOnCapital/status/2009045415451406405) 2026-01-07T23:32Z [--] followers, [---] engagements
"This is the same shift were seeing everywhere in AI infrastructure: from max context to efficient utilization. Dynamic context isnt just a UX win its an economics win. Lower token burn fewer truncation failures and more predictable long-run agent behavior. At scale stability + cost curves matter more than raw capability. https://twitter.com/i/web/status/2009050104519667841 https://twitter.com/i/web/status/2009050104519667841"
[X Link](https://x.com/TroyOnCapital/status/2009050104519667841) 2026-01-07T23:50Z [--] followers, [--] engagements
"This is also why agent reliability improves as context gets smaller not larger. Less noise fewer hallucinated branches tighter execution. https://twitter.com/i/web/status/2009050941253927292 https://twitter.com/i/web/status/2009050941253927292"
[X Link](https://x.com/TroyOnCapital/status/2009050941253927292) 2026-01-07T23:54Z [--] followers, [--] engagements
"@moninvestor Googles moat isnt just information. Its distribution + default placement + capital intensity. Search YouTube Android Maps create demand pull for models. Nvidia monetizes supply-side constraints. Different layers of the stack - both win but for very different reasons"
[X Link](https://x.com/TroyOnCapital/status/2009066912043675841) 2026-01-08T00:57Z [--] followers, [---] engagements
"@PolymarketMoney Less a VC win and more a strategic hedge. Google didnt buy IRR it bought model optionality compute demand and ecosystem leverage as AI spend explodes. The real payoff isnt the paper valuation its securing position in a power- and compute-constrained world"
[X Link](https://x.com/TroyOnCapital/status/2009068273254694922) 2026-01-08T01:02Z [--] followers, [--] engagements
"@MontanaMatos AI is shifting profit pools not ending incumbency. Googles advantage is infra + distribution at scale; Apples is device control and monetization. Market cap can move fast durable moats show up in cash flow and steady-state execution over time not single snapshots. $GOOGL"
[X Link](https://x.com/TroyOnCapital/status/2009073341790069125) 2026-01-08T01:23Z [--] followers, [--] engagements
"@TheStalwart @SamRo The tenure dividend is really lower quit rates + wage stickiness showing up as margin expansion. It boosts operating leverage in slow labor markets but its cyclical - not a permanent cost advantage once mobility normalizes"
[X Link](https://x.com/TroyOnCapital/status/2009073634338590980) 2026-01-08T01:24Z [--] followers, [--] engagements
"@rohanpaul_ai This isnt generosity - its labor market economics. When marginal researcher output scales non-linearly retaining top talent becomes a moat. Large equity pools are a rational response to extreme scarcity and capital intensity in frontier AI not a sign of excess"
[X Link](https://x.com/TroyOnCapital/status/2009076927458619837) 2026-01-08T01:37Z [--] followers, [---] engagements
"@SawyerMerritt This was always the core tension. Nonprofit mission + frontier-scale compute + competitive timelines eventually collide with capital reality. The real question isnt intent - its whether governance structures can actually survive exponential infrastructure demands"
[X Link](https://x.com/TroyOnCapital/status/2009082209660227678) 2026-01-08T01:58Z [--] followers, [---] engagements
"@moninvestor The headline isnt demand growth its where the constraint lands. Power availability grid interconnects and permitting timelines will decide which AI capacity actually gets built not GPU supply alone. Compute scales faster than energy infrastructure"
[X Link](https://x.com/TroyOnCapital/status/2009092078555549722) 2026-01-08T02:37Z [--] followers, [---] engagements
"This reads less like demand risk and more like constraint migration. HBM4 readiness and licensing shift the [----] mix but CoWoS consumption stays bounded by packaging + OSAT throughput not wafer starts. If N4 capacity and export approvals clear the upside shows up as delayed CoWoS acceleration rather than lost volume. Steady-state infra constraints keep re-ordering the ramp. https://twitter.com/i/web/status/2009096555815686385 https://twitter.com/i/web/status/2009096555815686385"
[X Link](https://x.com/TroyOnCapital/status/2009096555815686385) 2026-01-08T02:55Z [---] followers, [---] engagements
"Really impressive control and mobility. Whats exciting is less the demo itself and more what it signals - steady gains in autonomy power efficiency and reliability that eventually make real-world deployment viable. Robotics progress compounds quietly before it shows up everywhere. https://twitter.com/i/web/status/2009102932005048556 https://twitter.com/i/web/status/2009102932005048556"
[X Link](https://x.com/TroyOnCapital/status/2009102932005048556) 2026-01-08T03:20Z [--] followers, [--] engagements
"@StockMKTNewz Market cap flips are the effect not the cause. $GOOGL is being priced as an AI services + infra compounder while $AAPL is still anchored to hardware cycles. Different growth physics"
[X Link](https://x.com/TroyOnCapital/status/2009125444478750763) 2026-01-08T04:50Z [--] followers, [---] engagements
"@TheTechInvest The logos get attention but execution is the real test. Being tied to $NVDA / $MSFT / $AMD only matters if power capacity and uptime are already lined up. Thats when partner turns into real utilization. $IREN"
[X Link](https://x.com/TroyOnCapital/status/2009225702512665060) 2026-01-08T11:28Z [---] followers, [---] engagements
"@TopTechStocksUS Logos are great but the real edge is execution. If $NBIS can actually deliver power uptime and scale while demand is there thats when partnerships turn into revenue - not before"
[X Link](https://x.com/TroyOnCapital/status/2009230004501270784) 2026-01-08T11:45Z [---] followers, [---] engagements
"@lisaabramowicz1 @TheTerminal Record borrowing isnt panic - its optionality. Issuers are locking in funding while windows are wide spreads are tight and uncertainty is political not credit-driven. The tell wont be volumes - itll be who stops coming to market first when conditions turn"
[X Link](https://x.com/TroyOnCapital/status/2009230609231863869) 2026-01-08T11:48Z [--] followers, [--] engagements
"AI investing is shifting from whats possible to what actually runs. Chips like $NVDA set the pace but the real question this year is whether the infrastructure can keep up. Power uptime and utilization decide who actually compounds. The story gets less exciting here and thats usually where the money is made. https://twitter.com/i/web/status/2009232017385508890 https://twitter.com/i/web/status/2009232017385508890"
[X Link](https://x.com/TroyOnCapital/status/2009232017385508890) 2026-01-08T11:53Z [--] followers, [--] engagements
"@StockSavvyShay @fiscal_ai Key assumption is sustained supply constraint - margins only hold if capacity additions lag demand longer than expected. Do you see that persisting"
[X Link](https://x.com/TroyOnCapital/status/2009233647174602798) 2026-01-08T12:00Z [--] followers, [---] engagements
"@yianisz Agreed. Is your confirmation signal sustained volume expansion on the next leg or simply higher lows holding through consolidation"
[X Link](https://x.com/TroyOnCapital/status/2009235074873741414) 2026-01-08T12:05Z [--] followers, [----] engagements
"The China headline sounds huge but it feels mostly priced in at this point. If demand everywhere else is already capped by supply losing that slice doesnt really change the near-term story - it just changes whos at the front of the line. To me the bigger question is how long that supply tightness lasts not the China math. https://twitter.com/i/web/status/2009237070531739815 https://twitter.com/i/web/status/2009237070531739815"
[X Link](https://x.com/TroyOnCapital/status/2009237070531739815) 2026-01-08T12:13Z [--] followers, [--] engagements
"@yianisz That makes sense. Absorption first then expansion - otherwise its just noise. Watching whether those higher lows actually stick before getting excited about the breakout"
[X Link](https://x.com/TroyOnCapital/status/2009237865432273041) 2026-01-08T12:16Z [--] followers, [---] engagements
"@NVIDIADRIVE @nvidia @MercedesBenz This is the unsexy part that actually matters. If L2/L2+ works smoothly in real streets people trust it - and thats what scales"
[X Link](https://x.com/TroyOnCapital/status/2009389372525945030) 2026-01-08T22:18Z [--] followers, [---] engagements
"@The_AI_Investor Selling inferior - it usually means timing and scale matter more than peak benchmarks. Inference wins are about software ops and deployment friction not just SRAM or raw silicon"
[X Link](https://x.com/TroyOnCapital/status/2009410820950929648) 2026-01-08T23:44Z [--] followers, [---] engagements
"@rohanpaul_ai This looks less like a clash of ideas and more like a shift in what the org is optimizing for. Research priorities tend to change fast once scale and product pressure kick in"
[X Link](https://x.com/TroyOnCapital/status/2009442683362873761) 2026-01-09T01:50Z [--] followers, [--] engagements
"Hot take that probably shouldnt be a hot take: AI doesnt fail because the model isnt smart enough. It fails because something boring breaks. Power. Uptime. Throughput. Utilization. The interesting part usually starts after the demo works when everything has to run constantly. Thats usually what I look for when Im thinking about where real value compounds. https://twitter.com/i/web/status/2009465791104594062 https://twitter.com/i/web/status/2009465791104594062"
[X Link](https://x.com/TroyOnCapital/status/2009465791104594062) 2026-01-09T03:22Z [--] followers, [--] engagements
"@kyleichan Feels less like a sudden shift and more like momentum showing up in public. Once companies are ready they show up geopolitics or not"
[X Link](https://x.com/TroyOnCapital/status/2009491352585998783) 2026-01-09T05:04Z [--] followers, [---] engagements
"@rwang07 Memory isnt just tight its the pacing item now. When HBM and advanced DRAM are constrained everything upstream slows: GPUs servers deployment schedules. Thats where utilization and pricing power really show up"
[X Link](https://x.com/TroyOnCapital/status/2009596072587563072) 2026-01-09T12:00Z [--] followers, [---] engagements
"@qualtrim This isnt just growth its operating leverage finally showing up. Once cloud becomes steady-state infrastructure margins expand quietly and persistently"
[X Link](https://x.com/TroyOnCapital/status/2009610282272948678) 2026-01-09T12:56Z [--] followers, [--] engagements
"@dlacalle_IA True but rising wages can still coexist with weaker hiring breadth. Fewer sectors adding jobs means gains concentrate even as averages look healthy"
[X Link](https://x.com/TroyOnCapital/status/2009647758647664785) 2026-01-09T15:25Z [---] followers, [--] engagements
"@caro_milanesi @intel @HP @Microsoft Yeah this keeps showing up. The models arent the bottleneck as much as the messy last mile: perception dexterity reliability and knowing when to get out of the way. CES demos make that gap very visible"
[X Link](https://x.com/TroyOnCapital/status/2009718523447464407) 2026-01-09T20:06Z [---] followers, [--] engagements
"@danielisdizzy That combo can coexist for a while. Growth can look strong when gains concentrate in a few capital-intensive sectors even as hiring breadth thins elsewhere. The question is how wide that diffusion actually gets"
[X Link](https://x.com/TroyOnCapital/status/2009719361804873763) 2026-01-09T20:10Z [---] followers, [--] engagements
"@MikeZaccardi This is the disconnect people miss. Headline prints can look fine while the underlying job engine is already below breakeven. Not a crash signal just much less margin for strong labor narratives"
[X Link](https://x.com/TroyOnCapital/status/2009734708049956989) 2026-01-09T21:11Z [---] followers, [--] engagements
"@HedgieMarkets Feels more like a slow correction from pandemic over-hiring than AI replacing engineers. The spend shift and the layoffs are correlated in time not necessarily in cause"
[X Link](https://x.com/TroyOnCapital/status/2009735442615415203) 2026-01-09T21:14Z [---] followers, [---] engagements
"Todays jobs report doesnt really refute this it mostly shows the lag. Payrolls and unemployment tell you where the labor market was. The jobs-workers gap tells you how much buffer is left. When that buffer disappears growth can continue but it gets much more fragile to shocks. https://twitter.com/i/web/status/2009769640894124480 https://twitter.com/i/web/status/2009769640894124480"
[X Link](https://x.com/TroyOnCapital/status/2009769640894124480) 2026-01-09T23:29Z [---] followers, [--] engagements
"Nothing that mattered today actually broke. Jobs cooled without collapsing. Wages didnt re-accelerate. AI keeps shifting from whats possible to what actually runs. Thats why markets arent debating growth vs recession"
[X Link](https://x.com/TroyOnCapital/status/2009773191460466860) 2026-01-09T23:44Z [---] followers, [--] engagements
"@JosephPolitano The headline rate keeps saying fine but this shows the pressure shifting sideways instead of snapping. Fewer people are fully unemployed more are stuck on the margins. Thats a quieter kind of weak"
[X Link](https://x.com/TroyOnCapital/status/2009840302471938353) 2026-01-10T04:10Z [---] followers, [--] engagements
"@greg_ip I wonder if this is mostly a timing story. Tariffs hit prices fast but reshoring shows up only after long capex and permitting cycles. In the interim softer construction demand can easily swamp any early gains"
[X Link](https://x.com/TroyOnCapital/status/2009842117364068581) 2026-01-10T04:17Z [---] followers, [---] engagements
"@Schuldensuehner Digitalization is part of it but the deeper gap is utilization. The US reorganized firms capital and workflows around software. Germany digitized existing processes without re-wiring incentives. Same tools different throughput. Productivity follows structure"
[X Link](https://x.com/TroyOnCapital/status/2009962559169007671) 2026-01-10T12:16Z [---] followers, [---] engagements
"Markets arent reacting to data surprises right now. Theyre reacting to constraint clarity. Growth isnt breaking Inflation isnt re-accelerating Liquidity isnt tightening When nothing snaps capital stops hiding. https://twitter.com/i/web/status/2009969891361861722 https://twitter.com/i/web/status/2009969891361861722"
[X Link](https://x.com/TroyOnCapital/status/2009969891361861722) 2026-01-10T12:45Z [---] followers, [--] engagements
"Big prior-year returns matter less than what reset underneath them. In past cycles intl small value kept running when currencies stabilized valuations stayed below trend and earnings breadth widened not just because momentum persisted. The setup matters more than the headline gain. https://twitter.com/i/web/status/2009987050519572745 https://twitter.com/i/web/status/2009987050519572745"
[X Link](https://x.com/TroyOnCapital/status/2009987050519572745) 2026-01-10T13:53Z [---] followers, [---] engagements
"@robin_j_brooks This chart really shows the asymmetry. Markets can price cuts but as long as growth and services inflation dont crack the Feds bias stays tighter for longer. That tension not politics is what keeps the USD pressure story alive"
[X Link](https://x.com/TroyOnCapital/status/2009988022901846155) 2026-01-10T13:57Z [---] followers, [---] engagements
"@robin_j_brooks That pressure may rise agreed. But markets still need cover. As long as growth and services inflation dont actually crack cuts stay conditional not automatic. Politics can lean but constraints still decide timing"
[X Link](https://x.com/TroyOnCapital/status/2009993768074887215) 2026-01-10T14:20Z [---] followers, [--] engagements
"@MikeZaccardi @StockCharts Yeah energy helps the comps but its a narrow window. Once base effects fade CPI momentum is going to hinge more on shelter and services than oil alone. A brief 2-handle wouldnt change the underlying glide path"
[X Link](https://x.com/TroyOnCapital/status/2010034353552990710) 2026-01-10T17:01Z [---] followers, [--] engagements
"@dailychartbook @BlackRock The key detail is where that growth is coming from. This is capital-intensive buildout showing up as investment not a broad productivity surge yet which also means its power- supply-chain- and balance-sheet constrained. Durable but not frictionless"
[X Link](https://x.com/TroyOnCapital/status/2010035657968869544) 2026-01-10T17:07Z [---] followers, [--] engagements
"@dailychartbook This isnt really about retirement culture its about labor composition and utilization. Countries with older exit ages tend to have higher participation less churn and slower productivity dilution. That quietly matters for growth and inflation dynamics"
[X Link](https://x.com/TroyOnCapital/status/2010081436074963303) 2026-01-10T20:08Z [---] followers, [--] engagements
"@neilksethi Thats a real tell this close to earnings. The next question is how broad those revisions are when a few heavyweights do the lifting headline growth looks fine but durability is thinner"
[X Link](https://x.com/TroyOnCapital/status/2010091627537449330) 2026-01-10T20:49Z [---] followers, [--] engagements
"@curious_founder Whats striking is how persistent this pattern is across revolutions. Early gains concentrate in capital and profits; labor share recovers only after institutions bargaining power and diffusion catch up. Were still in that lag phase with AI"
[X Link](https://x.com/TroyOnCapital/status/2010172442355085552) 2026-01-11T02:10Z [---] followers, [--] engagements
"This is the shift hiding in plain sight. U.S. electricity demand growth has broken out of a decade-long flat regime and remains elevated. Thats why power siting and firm generation are becoming the real constraints not chips or models"
[X Link](https://x.com/TroyOnCapital/status/2010181272052895986) 2026-01-11T02:45Z [---] followers, [---] engagements
"@Barchart The Buffett Indicator tells you where market cap sits relative to GDP not where stress builds. What matters now is whether the system can carry that valuation: fiscal support passive flows and capital-dense growth keep pressure off until a real constraint binds"
[X Link](https://x.com/TroyOnCapital/status/2010200969645355362) 2026-01-11T04:03Z [---] followers, [---] engagements
"@EugeneNg Tech employment peaking doesnt contradict AI it fits it. Capital is shifting from labor-heavy software to compute- power- and infra-intensive layers. Fewer hires more throughput. Structural looks more plausible than cyclical here"
[X Link](https://x.com/TroyOnCapital/status/2010324178550038742) 2026-01-11T12:13Z [---] followers, [--] engagements
"@SKundojjala This comparison really highlights the shift in value creation. NVIDIAs growth isnt cyclical demand it reflects a structural reallocation of capital toward compute- and power-intensive infrastructure while mature consumer franchises remain volume-bound"
[X Link](https://x.com/TroyOnCapital/status/2010325662570258576) 2026-01-11T12:19Z [---] followers, [---] engagements
"@dlacalle_IA Agreed but the dynamism isnt just growth its capital velocity. The U.S. reallocates faster when constraints shift: labor capital software infrastructure. That flexibility is what keeps productivity compounding even as cycles slow"
[X Link](https://x.com/TroyOnCapital/status/2010326934891483267) 2026-01-11T12:24Z [---] followers, [--] engagements
"@StockSavvyShay Agreed and the edge isnt just growth its capital efficiency. YouTube lets Google deploy AI at scale with zero behavior change turning distribution into a structural compounding advantage"
[X Link](https://x.com/TroyOnCapital/status/2010328518018924939) 2026-01-11T12:30Z [---] followers, [---] engagements
"@dlacalle_IA Currency collapse despite exports is the tell. FX prices confidence not cash flow. When capital controls sanctions risk and policy credibility erode trade surpluses and external support cant anchor the currency. Growth without trust doesnt clear in FX"
[X Link](https://x.com/TroyOnCapital/status/2010333317468676194) 2026-01-11T12:49Z [---] followers, [--] engagements
"Feels less like a causal dollar story and more like regime context. A weaker USD only helps equities when growth is steady enough to absorb it not when leadership is being second-guessed by surveys. NVIDIA turning usually isnt a macro call so much as a utilization and capex reality check. https://twitter.com/i/web/status/2010406683751891129 https://twitter.com/i/web/status/2010406683751891129"
[X Link](https://x.com/TroyOnCapital/status/2010406683751891129) 2026-01-11T17:41Z [---] followers, [--] engagements
"@curious_founder The scale gap is real. The harder part now is what happens after install. Turning capacity into usable power comes down to grids storage and utilization thats where the bottleneck is shifting"
[X Link](https://x.com/TroyOnCapital/status/2010438894781927604) 2026-01-11T19:49Z [---] followers, [---] engagements
"The numbers make sense together. Renewables are scaling fast but demand growth + grid constraints mean theyre still largely additive. Displacement only accelerates once storage transmission and industrial load shifting scale in parallel. Capacity is the easier part. Coordination is the bottleneck. https://twitter.com/i/web/status/2010444326749515864 https://twitter.com/i/web/status/2010444326749515864"
[X Link](https://x.com/TroyOnCapital/status/2010444326749515864) 2026-01-11T20:10Z [---] followers, [--] engagements
"Thats the key point returns followed earnings not multiple expansion. Koreas cycle has been driven by export-heavy capital-intensive sectors where margins recover fast once utilization turns. The question now isnt valuation its whether that earnings durability persists as global capex normalizes. https://twitter.com/i/web/status/2010450659104247932 https://twitter.com/i/web/status/2010450659104247932"
[X Link](https://x.com/TroyOnCapital/status/2010450659104247932) 2026-01-11T20:36Z [---] followers, [---] engagements
"@MikeZaccardi This is a timing distortion more than a growth impulse. Shutdowns pull activity forward/back across quarters so the headline swing matters less than the underlying trend. Markets usually look through this once its clear nothing structural changed"
[X Link](https://x.com/TroyOnCapital/status/2010452533526089751) 2026-01-11T20:43Z [---] followers, [--] engagements
"@RyanDetrick @SamRo Striking chart but its as much composition as productivity. Capital-intensive leaders drive revenue per worker higher as utilization and margins scale even if median firms dont see the same gains. That concentration is the tell"
[X Link](https://x.com/TroyOnCapital/status/2010454654434550139) 2026-01-11T20:51Z [---] followers, [---] engagements
"Markets dont move on headlines. They move when constraints become undeniable. Capital freezes when exit risk turns structural Growth survives when utilization not optimism holds Uninvestable isnt political; its learned behavior I focus on where systems quietly break: power compute labor capital mobility and how markets reprice before consensus notices. If you care about second-order effects not surface narratives youre in the right place. https://twitter.com/i/web/status/2010508471591191005 https://twitter.com/i/web/status/2010508471591191005"
[X Link](https://x.com/TroyOnCapital/status/2010508471591191005) 2026-01-12T00:25Z [---] followers, [---] engagements
"What matters less than the headline odds is what they reveal about uncertainty. Markets dont need Powell to actually leave to reprice they move when institutional continuity becomes a variable. Rising probabilities translate into higher term premia wider risk buffers and a shift from growth narratives toward credibility and governance risk. Thats the real signal here. https://twitter.com/i/web/status/2010688118161391711 https://twitter.com/i/web/status/2010688118161391711"
[X Link](https://x.com/TroyOnCapital/status/2010688118161391711) 2026-01-12T12:19Z [---] followers, [--] engagements
"Whats notable is how gold is outperforming not as a spike on panic but as a steady relative grind. That tends to show up when confidence in policy frameworks erodes at the margin. Less about immediate recession risk more about credibility term premia and the cost of uncertainty rising quietly across assets. https://twitter.com/i/web/status/2010693006022840609 https://twitter.com/i/web/status/2010693006022840609"
[X Link](https://x.com/TroyOnCapital/status/2010693006022840609) 2026-01-12T12:39Z [---] followers, [--] engagements
"@EricBalchunas Feels less like speculative risk-on and more like structural allocation behavior. Broad beta cash-adjacent ETFs and liquidity preference dominating not a chase but positioning. Those flows usually stick until a real constraint snaps"
[X Link](https://x.com/TroyOnCapital/status/2010694229077701110) 2026-01-12T12:43Z [---] followers, [---] engagements
"@TheStalwart Once an institution is running real P&L independence becomes conditional whether courts admit it or not. Markets dont wait for constitutional clarity they price loss socialization and governance uncertainty immediately. Thats the channel showing up in term premia and gold"
[X Link](https://x.com/TroyOnCapital/status/2010696758460867033) 2026-01-12T12:53Z [---] followers, [--] engagements
"@robin_j_brooks @TheStalwart Breakevens dont lead regimes they tend to lag credibility breaks. Markets wait for realized inflation to force belief then reprice fast. Until then they anchor to the last cycle"
[X Link](https://x.com/TroyOnCapital/status/2010711044285436130) 2026-01-12T13:50Z [---] followers, [--] engagements
"@RyanDetrick @sonusvarghese This looks less like broad labor weakness and more like entry-level friction. Firms are hoarding experienced labor while slowing new intake so the pain concentrates at the margin without showing up in headline unemployment"
[X Link](https://x.com/TroyOnCapital/status/2010727117076804067) 2026-01-12T14:54Z [---] followers, [---] engagements
"@JKempEnergy @JesseJenkins This looks less like gasoline prices not mattering and more like efficiency absorbing the shock. Miles driven keep rising but demand elasticity is muted because each mile just uses less fuel than it used to"
[X Link](https://x.com/TroyOnCapital/status/2010740095264117085) 2026-01-12T15:46Z [---] followers, [---] engagements
"@Convertbond Its not just a rotation out of bonds its a repricing of what qualifies as duration when fiscal supply is elastic and real yields are volatile. Gold absorbs duration risk without counterparty or policy exposure. Thats the quiet shift"
[X Link](https://x.com/TroyOnCapital/status/2010784217593160020) 2026-01-12T18:41Z [---] followers, [--] engagements
"@KevRGordon This gap is mostly composition. Net worth gains are highly concentrated and asset-heavy while day-to-day sentiment is driven by prices cash flow and housing affordability. Aggregate wealth can rise even as median experience deteriorates"
[X Link](https://x.com/TroyOnCapital/status/2010789631244275756) 2026-01-12T19:03Z [---] followers, [--] engagements
"@chigrl The signal here isnt that next-gen nuclear is suddenly ready its how scarce firm power has become. Hyperscalers are buying optionality and learning time because grid timelines are the real constraint not reactor physics"
[X Link](https://x.com/TroyOnCapital/status/2010808568489132167) 2026-01-12T20:18Z [---] followers, [---] engagements
"@robin_j_brooks What jumps out is how mechanical this looks. When global duration and safe collateral are scarce USTs clear even without a premium. Less a verdict on the Fed more a signal about balance-sheet constraints"
[X Link](https://x.com/TroyOnCapital/status/2010815053382177001) 2026-01-12T20:44Z [---] followers, [---] engagements
"I think the tension is that markets arent really pricing cuts theyre pricing duration suppression. Hard assets can run in a world where policy rates stay higher but real rates get capped through balance sheets regulation and debt management. Thats financial repression not a classic easing cycle"
[X Link](https://x.com/TroyOnCapital/status/2010829935578136659) 2026-01-12T21:43Z [---] followers, [---] engagements
"The distinction vs. the mid-2000s is that todays housing market is constrained by stock and flow not leverage. Household balance sheets are stronger but rate lock-in and low turnover have impaired price discovery. That can support prices without a debt boom yet still create real affordability stress. https://twitter.com/i/web/status/2010880061625237820 https://twitter.com/i/web/status/2010880061625237820"
[X Link](https://x.com/TroyOnCapital/status/2010880061625237820) 2026-01-13T01:02Z [---] followers, [--] engagements
"@BobEUnlimited The key constraint is that households experience inflation through cash-flow timing not averages. Even if policy looks through it recurring price resets on essentials compress real disposable income long before labor or nominal growth can offset it"
[X Link](https://x.com/TroyOnCapital/status/2011037475926663522) 2026-01-13T11:27Z [---] followers, [--] engagements
"@dailychartbook @neilksethi Looks less like a true rotation and more like risk-budget inertia. Mega-cap growth stays overweight because it still absorbs uncertainty better than cyclicals even as positioning bleeds at the margin"
[X Link](https://x.com/TroyOnCapital/status/2011038596887887878) 2026-01-13T11:32Z [---] followers, [--] engagements
"@MikeZaccardi @StockCharts Total return tells the story here FX and duration are doing the work. Its another signal that global real rates are being capped even as policy paths diverge"
[X Link](https://x.com/TroyOnCapital/status/2011040753712328916) 2026-01-13T11:40Z [---] followers, [--] engagements
"What makes this unusual isnt just the price action but the signal it sends about policy credibility. If yields rise and the currency still weakens markets may be questioning whether normalization can occur without destabilizing balance sheets. Thats a constraint not a cyclical move. https://twitter.com/i/web/status/2011041228775936014 https://twitter.com/i/web/status/2011041228775936014"
[X Link](https://x.com/TroyOnCapital/status/2011041228775936014) 2026-01-13T11:42Z [---] followers, [----] engagements
"@neilksethi @dailychartbook The asymmetry is the point. CTAs dont need to be bearish to be a risk flat to modestly higher prices dont add much fuel but downside breaches trigger large mechanical selling. Thats convexity embedded in positioning not sentiment"
[X Link](https://x.com/TroyOnCapital/status/2011041949093085265) 2026-01-13T11:45Z [---] followers, [--] engagements
"What stands out isnt just fear of disruption its how little conviction there is in spot. Record upside hedging alongside a soft curve suggests participants want insurance not exposure. Thats classic constraint pricing: limited spare capacity + geopolitics force volatility bids even when the base case still points to surplus. https://twitter.com/i/web/status/2011042556612919637 https://twitter.com/i/web/status/2011042556612919637"
[X Link](https://x.com/TroyOnCapital/status/2011042556612919637) 2026-01-13T11:48Z [---] followers, [---] engagements
"This framing matters. What gets lost is that restraint on hiring isnt a macro pessimism signal its a capital allocation choice. Firms are prioritizing durability (AI M&A optionality buybacks) over labor flexibility even with strong earnings. That shift has real human consequences but its driven more by balance-sheet logic than fear. https://twitter.com/i/web/status/2011043938145366284 https://twitter.com/i/web/status/2011043938145366284"
[X Link](https://x.com/TroyOnCapital/status/2011043938145366284) 2026-01-13T11:53Z [---] followers, [---] engagements
"@RonStoeferle @kevinmuir This is the tell: price is breaking out but options arent chasing it. Calm upside IV and offered call skew say this is allocation and stress hedgingnot FOMO. Gold peaks when upside insurance gets crowded; disbelief is still in charge"
[X Link](https://x.com/TroyOnCapital/status/2011049132467834883) 2026-01-13T12:14Z [---] followers, [---] engagements
"This reads less like credit deterioration and more like optionality. Strong profits allow banks to over-reserve early which cushions future shocks and protects capital if the cycle turns. Stress shows up in charge-offs not provisions. Provisions buy time. Charge-offs reveal truth"
[X Link](https://x.com/TroyOnCapital/status/2011050615284384157) 2026-01-13T12:20Z [---] followers, [---] engagements
"@KevRGordon Big jump. Feels less like taxes suddenly rose and more like margins finally got thin enough that they hurt. When growth slows everything fixed becomes the problem"
[X Link](https://x.com/TroyOnCapital/status/2011062174920413653) 2026-01-13T13:06Z [---] followers, [--] engagements
"CPI didnt surprise today but the path still matters more than the print. Inflation is calmer than it was not back to the old regime either. December CPI #inflation +2.7% vs. +2.7% est & +2.7% prior core +2.6% vs. +2.7% est. & +2.6% prior https://t.co/R3kcxvyULj December CPI #inflation +2.7% vs. +2.7% est & +2.7% prior core +2.6% vs. +2.7% est. & +2.6% prior https://t.co/R3kcxvyULj"
[X Link](https://x.com/TroyOnCapital/status/2011070016700236263) 2026-01-13T13:37Z [---] followers, [--] engagements
"@KevRGordon This reads less hawkish than it sounds. Well positioned is Fed code for optionality not conviction. If inflation keeps behaving the hurdle for easing is time not resistance"
[X Link](https://x.com/TroyOnCapital/status/2011094909265322017) 2026-01-13T15:16Z [---] followers, [--] engagements
"@conorsen That framing feels right. When population growth slows growth shows up as income dispersion more than headcount expansion. Youre not creating more middle class jobs youre asking existing households to climb and not everyone can"
[X Link](https://x.com/TroyOnCapital/status/2011107814463455459) 2026-01-13T16:07Z [---] followers, [---] engagements
"@RealEJAntoni Feels like less churn not less leverage. Switching premiums have compressed but stayers are still seeing gains which points to tighter matching and fewer outside options. Stable yes but bargaining power is narrowing at the margin"
[X Link](https://x.com/TroyOnCapital/status/2011108912821649591) 2026-01-13T16:11Z [---] followers, [--] engagements
"@BobEUnlimited @albertocavallo What this really shows is a level reset not persistent inflation. Tariffs tend to push prices up once then sit in the index rather than unwind. That still matters for how inflation feels even if the rate cools"
[X Link](https://x.com/TroyOnCapital/status/2011110452672282908) 2026-01-13T16:17Z [---] followers, [---] engagements
"ELI5: instead of a flat margin number CME Group will now scale margin with price. If gold goes up required margin goes up automatically. Its just about keeping risk proportional as prices move not signaling direction. BREAKING: CME Group will set gold silver platinum and palladium margins as a percentage of notional value instead of a fixed dollar amount BREAKING: CME Group will set gold silver platinum and palladium margins as a percentage of notional value instead of a fixed dollar amount"
[X Link](https://x.com/TroyOnCapital/status/2011112761577914689) 2026-01-13T16:27Z [---] followers, [--] engagements
"@LizThomasStrat Early days but this reads less like growth optimism and more like pricing power and scarcity. Materials and gold doing well together usually says something about constraints not animal spirits"
[X Link](https://x.com/TroyOnCapital/status/2011143942688162271) 2026-01-13T18:30Z [---] followers, [--] engagements
"@robin_j_brooks It might be less a break in inflation dynamics and more a break in measurement continuity. Missing a month matters when so much of core is smoothed lagged or imputed. Hard to know if inflation slowed or if the signal just got noisier"
[X Link](https://x.com/TroyOnCapital/status/2011151762661785707) 2026-01-13T19:02Z [---] followers, [---] engagements
"@KoyfinCharts This is the one price everyone actually feels. You dont notice electricity until it stops being cheap then it quietly shows up everywhere rent groceries running a business. Its the kind of cost that doesnt make headlines but slowly changes behavior"
[X Link](https://x.com/TroyOnCapital/status/2011153738195747083) 2026-01-13T19:09Z [---] followers, [---] engagements
"@IrvingSwisher This is the inflation people actually feel. Restaurants sit at the intersection of wages rent and demand so prices dont come down easily. Even as goods cool eating out keeps reminding people inflation isnt really gone"
[X Link](https://x.com/TroyOnCapital/status/2011161358415646974) 2026-01-13T19:40Z [---] followers, [---] engagements
"@LizAnnSonders New home sales can look stable even as prices fall because builders are the shock absorber. They adjust incentives and mix to keep deals moving. Thats flexibility not a broad demand rebound"
[X Link](https://x.com/TroyOnCapital/status/2011175400974712939) 2026-01-13T20:35Z [---] followers, [---] engagements
"@joshrauh Capital isnt interchangeable here. Project finance looks cheap on paper but it pushes risk around. When you actually need power on time the balance sheet matters more than the coupon"
[X Link](https://x.com/TroyOnCapital/status/2011181697795035322) 2026-01-13T21:00Z [---] followers, [---] engagements
"@EPBResearch This is the frustrating part of this cycle. You think you see cyclical hiring come back then the revisions walk it back. More uncertainty than momentum"
[X Link](https://x.com/TroyOnCapital/status/2011189617744023862) 2026-01-13T21:32Z [---] followers, [---] engagements
### Topic Creators
https://lunarcrush.com/api4/public/topic/:topic/creators/v1
Get the top creators for a social topic
input parameters:
+ **topic**: _Provide the topic to get details for. A topic must be all lower case and can only include letters, numbers, spaces, # and $._ **required**
Example request:
```bash
curl -H "Authorization: Bearer <API_KEY>" https://lunarcrush.com/api4/public/topic/bitcoin/creators/v1
```
Example response:
```json
{
"data": [
{
"creator_id": "twitter::1605310254085980161",
"creator_name": "Bitcoin_Teddy",
"creator_avatar": "https://pbs.twimg.com/profile_images/1739753415134138368/FrWHMAnb_200x200.jpg",
"creator_followers": 74293,
"creator_rank": 1,
"interactions_24h": [-------]
}
]
}
```
Schema:
+ **creator_id**: The [network]::[unique_id] for the influencer
+ **creator_name**: The unique screen name for the influencer
+ **creator_avatar**: The URL to the avatar for the creator
+ **creator_followers**: number of followers the account has
+ **creator_rank**: ranking based on all posts in the last [--] hours that have interactions
+ **interactions_24h**: Number of interactions in the last [--] hours
---
### Creators List
https://lunarcrush.com/api4/public/creators/list/v1
Get a list of trending social creators over all of social based on interactions. To get lists of creators by category or topic see the topics and categories endpoints.
Example request:
```bash
curl -H "Authorization: Bearer <API_KEY>" https://lunarcrush.com/api4/public/creators/list/v1
```
Example response:
```json
{
"data": [
{
"creator_name": "elonmusk",
"creator_display_name": "Elon Musk",
"creator_id": "44196397",
"creator_network": "twitter",
"creator_avatar": "https://pbs.twimg.com/profile_images/2008546467615580160/57KcqsTA_200x200.jpg",
"creator_followers": 234295750,
"creator_posts": 3649,
"creator_rank": 1,
"interactions_24h": [---------]
}
]
}
```
Schema:
+ **creator_name**: The unique screen name for the influencer
+ **creator_display_name**: The chosen display name for the influencer if available
+ **creator_id**: The [network]::[unique_id] for the influencer
+ **creator_network**: The social network for the post or influencer. We still refer to x as twitter out of developer preference.
+ **creator_avatar**: The URL to the avatar for the creator
+ **creator_followers**: number of followers the account has
+ **creator_posts**: total number of posts with interactions in the last [--] hours
+ **creator_rank**: ranking based on all posts in the last [--] hours that have interactions
+ **interactions_24h**: Number of interactions in the last [--] hours
---
### Creator
https://lunarcrush.com/api4/public/creator/:network/:id/v1
Get detail information on a specific creator
input parameters:
+ **network**: _Provide the network for the creator. One of twitter, youtube, instagram, reddit, or tiktok_ **required**
+ **id**: _Provide the unique ID or screen name of the creator_ **required**
Example request:
```bash
curl -H "Authorization: Bearer <API_KEY>" https://lunarcrush.com/api4/public/creator/twitter/elonmusk/v1
```
Example response:
```json
{
"data": {
"creator_id": "twitter::44196397",
"creator_name": "elonmusk",
"creator_display_name": "Elon Musk",
"creator_avatar": "https://pbs.twimg.com/profile_images/2008546467615580160/57KcqsTA_200x200.jpg",
"creator_followers": 234296337,
"creator_rank": 2,
"interactions_24h": 318539577,
"topic_influence": [
{
"topic": "in the",
"count": 294,
"percent": 5.88,
"rank": [--]
}
],
"top_community": [
{
"creator_name": "grok",
"creator_display_name": "Grok",
"creator_avatar": "https://pbs.twimg.com/profile_images/1893219113717342208/Vgg2hEPa_200x200.jpg",
"count": [---]
}
]
}
}
```
Schema:
+ **creator_id**: The [network]::[unique_id] for the influencer
+ **creator_name**: The unique screen name for the influencer
+ **creator_display_name**: The chosen display name for the influencer if available
+ **creator_avatar**: The URL to the avatar for the creator
+ **creator_followers**: number of followers the account has
+ **creator_rank**: ranking based on all posts in the last [--] hours that have interactions
+ **interactions_24h**: Number of interactions in the last [--] hours
+ **topic_influence**: an array of social topics and the creators ranking on each topic
+ **topic**: LunarCrush social topic. Can only includes letters, numbers, spaces, #, and $
+ **top_community**: an array of the top accounts that have recently mentioned this creator or that this creator has mentioned
---
### Creator Time Series
https://lunarcrush.com/api4/public/creator/:network/:id/time-series/v1
Get time series data on a creator.
input parameters:
+ **network**: _Influencer social network_ **required**
+ **id**: _The unique id or screen name of the creator_ **required**
+ **bucket**: _bucket time series data into hours or days. default is hours._
+ **interval**: _Use interval to specify the start and end time automatically for convenience. If "start" or "end" parameters are provided this parameter is ignored._
+ **start**: _The start time (unix timestamp) to go back to._
+ **end**: _The end time (unix timestamp) to stop at._
Example request:
```bash
curl -H "Authorization: Bearer <API_KEY>" https://lunarcrush.com/api4/public/creator/twitter/lunarcrush/time-series/v1
```
Example response:
```json
{
"config": {
"network": "twitter",
"influencer_id": "twitter::988992203568562176",
"interval": "1w",
"start": 1769990400,
"end": 1770681600,
"bucket": "hour",
"name": "lunarcrush",
"remote_api": "danode1-13",
"generated": [----------]
},
"data": [
{
"time": 1769990400,
"followers": 305326,
"interactions": 717,
"posts_active": 14,
"creator_rank": [------]
}
]
}
```
Schema:
+ **config**: This includes the inputs for the request processed by the server and may include additional hints about the request and response information.
+ **network**: The social network for the post or influencer. We still refer to x as twitter out of developer preference.
+ **interval**: Typically used for specifying time intervals like 1w = [--] week, 1m = [--] month etc
+ **start**: Start/from unix timestamp (in seconds)
+ **end**: End/to unix timestamp (in seconds)
+ **bucket**: Data is generally bucketed into hours or days
+ **name**: The full name of the asset
+ **generated**: A unix timestamp (in seconds) when the data was generated to understand possibly stale data
+ **time**: A unix timestamp (in seconds)
+ **followers**: The number of publicly displayed followers the creator has
+ **interactions**: number of all publicly measurable interactions on a social post (views, likes, comments, thumbs up, upvote, share etc)
+ **posts_active**: number of unique social posts with interactions
+ **creator_rank**: ranking based on all posts in the last [--] hours that have interactions
---
Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing
@TroyOnCapital Troy | Following Capital
Troy | Following Capital posts on X about ai, $nbis, gold, money the most. They currently have [---] followers and [---] posts still getting attention that total [---------] engagements in the last [--] hours.
Engagements: [---------] #
- [--] Week [---------] +3,242%
- [--] Month [---------] +7,486%
Mentions: [--] #
- [--] Week [--] -74%
- [--] Month [---] +174%
Followers: [---] #
- [--] Week [---] +18%
- [--] Month [---] +496%
CreatorRank: [-------] #
Social Influence
Social category influence finance stocks technology brands cryptocurrencies automotive brands countries currencies celebrities social networks financial services
Social topic influence ai, $nbis #121, gold, money, matter, inflation, in the, if you, $nvda, math
Top accounts mentioned or mentioned by @moneycruncher @thealphathought @kalshi @kobeissiletter @brianferoldi @rajatsoni @theficouple @themotleyfool @stockmktnewz @polymarket @cryptofergani @kinginvestings @elonmusk @unusualwhales @larkdavis @thestalwart @robinjbrooks @lynk0x @steveonspeed @nathanielsford
Top assets mentioned Nebius Group N.V. Class A Ordinary Shares (NBIS) NVIDIA Corp. (NVDA) Bitcoin (BTC) Tesla, Inc. (TSLA) Alphabet Inc Class A (GOOGL) Iris Energy Limited (IREN) Microsoft Corp. (MSFT) Metadium (META) CatCoin (CAT) Apple, Inc. (AAPL) Strategy (MSTR) Amazon.com, Inc. (AMZN)
Top Social Posts
Top posts by engagements in the last [--] hours
"@meetblossomapp Great to see $NBIS jumping early [----] I just posted about how $TSLA $NBIS (via Avride) and $NVDA are driving the AI compute race would love your thoughts"
X Link 2026-01-03T19:31Z [--] followers, [--] engagements
"@robo_burro @elonmusk @Tesla Great insights I recently posted about how $TSLA $NBIS (via Avride) and $NVDA are driving the AI compute race would love to hear your thoughts"
X Link 2026-01-03T19:33Z [--] followers, [---] engagements
"@longinvest32 Great points I recently posted about how $TSLA $NBIS (via Avride) and $NVDA are driving the AI compute race would love your thoughts"
X Link 2026-01-03T19:35Z [--] followers, [---] engagements
"Diversification within the AI infrastructure stack is underappreciated. When the core data center layer starts compounding demand having adjacent assets supporting execution and services can matter. Curious how others see the role of these non-core infrastructure revenue streams. https://twitter.com/i/web/status/2007553680825036927 https://twitter.com/i/web/status/2007553680825036927"
X Link 2026-01-03T20:44Z [--] followers, [--] engagements
"@Tpeezy16826 @GrindeOptions Autonomy narratives always lead with software but the gating factor is compute. Training and inference at scale force massive infrastructure build-outs long before revenue shows up. Thats where the real capital gets deployed first"
X Link 2026-01-03T21:27Z [--] followers, [--] engagements
"@DaphneFiel5881 Semiconductor shortages highlight the physical bottlenecks of compute supply but the deeper question is where all that compute will live and how power and cooling will be managed. Thats why Im focused on AI-native data centers and infrastructure that can handle the next wave"
X Link 2026-01-03T21:36Z [--] followers, [--] engagements
"@afterbrainrot Thanks for flagging this. Always wise to verify sources and avoid scams. The real $NBIS story is about AI-native infrastructure and data centersthats where the value lies. Stay focused on fundamentals not pump schemes"
X Link 2026-01-03T21:43Z [--] followers, [---] engagements
"@MarketMatrixs Interesting list Love seeing $NBIS paired with $IREN AI-native infrastructure meets energy solutions. As compute demand grows $NBIS is well-positioned to capture the data center capex. Im bullish on the infrastructure layer"
X Link 2026-01-03T21:44Z [--] followers, [---] engagements
"@TheTechInvest Cant wait for what [----] has in store for $IREN and $NBIS Dont get the hate between the communities - both can thrive"
X Link 2026-01-03T22:21Z [--] followers, [---] engagements
"The real bottleneck isnt just silicon its infrastructure & capital flows scaling that silicon. Raw GPUs address compute Data centers + interconnects scale training & inference Specialized partners like $NBIS / CoreWeave amplify deployment efficiency Thats why $NVDA isnt just dominant it anchors the stack. https://twitter.com/i/web/status/2007593528046825703 https://twitter.com/i/web/status/2007593528046825703"
X Link 2026-01-03T23:22Z [--] followers, [---] engagements
"@danielnewmanUV This is exactly right. Narratives about less compute ignore how AI systems are actually deployed. More models more agents more autonomy = more infrastructure not less. Thats why $NVDA demand stays structural"
X Link 2026-01-03T23:32Z [--] followers, [--] engagements
"The autonomy edge ultimately comes down to training scale not just model design. Vision + data loops require massive efficient compute behind the scenes. Thats why $TSLA AI story still intersects with the broader infrastructure stack $NVDA on compute and names like $NBIS enabling scalable deployment. https://twitter.com/i/web/status/2007597293399331260 https://twitter.com/i/web/status/2007597293399331260"
X Link 2026-01-03T23:37Z [--] followers, [---] engagements
"@MontanaMatos Definitely need to dive deeper on $PATH - main focus has been $NBIS. Ill research more this coming week"
X Link 2026-01-04T02:17Z [--] followers, [---] engagements
"@beauty_oe $NBIS FTW - the company is positioned well with excellent planned growth ahead"
X Link 2026-01-04T03:21Z [--] followers, [----] engagements
"$TSLA is pushing autonomy. $NVDA dominates AI compute. But theres a layer most investors are still missing. The AI infrastructure that actually makes all of this work at scale. Thats where $NBIS quietly sits. Training inference data pipelines and real-world deployment dont happen without it. Were early in the AI infra build-out and this layer gets repriced after autonomy and chips. Most arent paying attention yet. https://twitter.com/i/web/status/2007786542538883391 https://twitter.com/i/web/status/2007786542538883391"
X Link 2026-01-04T12:09Z [--] followers, [---] engagements
"Strong list. One thing that stands out: AI value is starting to bifurcate. App-layer winners like $META compound through distribution but infra names scale through capacity power and utilization not features. Curious which of these you think benefits most from AI demand staying supply-constrained into [----] https://twitter.com/i/web/status/2007851672119173351 https://twitter.com/i/web/status/2007851672119173351"
X Link 2026-01-04T16:28Z [--] followers, [---] engagements
"The way to think about this is by layers not just sectors. Compute: $NVDA Autonomy / platforms: $TSLA Infrastructure & deployment: $NBIS $IREN Software orchestration: $PATH $ZETA Different layers get priced at different times as AI moves from training to real-world scale. https://twitter.com/i/web/status/2007882570000957838 https://twitter.com/i/web/status/2007882570000957838"
X Link 2026-01-04T18:31Z [--] followers, [---] engagements
"Lists like this are useful but the real edge comes from understanding why each layer matters. Chips ( $NVDA ) enable models Autonomy ( $TSLA ) drives applications Infrastructure ( $NBIS $IREN ) determines whether any of it can scale reliably The bottleneck layer usually gets repriced last. https://twitter.com/i/web/status/2007883122260807783 https://twitter.com/i/web/status/2007883122260807783"
X Link 2026-01-04T18:33Z [--] followers, [---] engagements
"This is the key point most people miss. ASICs win on cost only after workloads stabilize. Until then GPUs dominate because flexibility matters more than efficiency. That also means demand stays supply-constrained longer than models assume especially once inference scales. https://twitter.com/i/web/status/2007915033779654958 https://twitter.com/i/web/status/2007915033779654958"
X Link 2026-01-04T20:40Z [--] followers, [--] engagements
"@rohanpaul_ai The hidden constraint isnt data its cost-effective reliable inference at scale. Most AI-first enterprise products stall not on features but when latency utilization and margins collide in production"
X Link 2026-01-04T20:42Z [---] followers, [---] engagements
"When Elon says singularity hes not talking sci-fi consciousness. Hes talking about the point where AI progress becomes self-reinforcing faster models driving more deployment which drives more data which accelerates improvement. The bottleneck isnt intelligence anymore. Its compute power and cost at scale. https://twitter.com/i/web/status/2007917302847869172 https://twitter.com/i/web/status/2007917302847869172"
X Link 2026-01-04T20:49Z [--] followers, [--] engagements
"@elonmusk @DavidSHolz Singularity isnt intelligence its feedback loops hitting physical limits (power capex latency). Thats where timelines stretch"
X Link 2026-01-04T23:06Z [--] followers, [--] engagements
"@slow_developer This is exactly right. Abundance in one layer creates scarcity elsewhere. Were trading physical limits for coordination energy and attention constraints. The climb didnt end it changed shape"
X Link 2026-01-05T00:45Z [---] followers, [--] engagements
"Core thesis on $NBIS: AI value doesnt accrue at models it accrues at production-grade inference. $NBIS isnt generic cloud. Its GPU-first infrastructure built for deterministic latency predictable costs and sustained throughput the requirements for AI at industrial scale. Microsoft & Meta dont experiment here they sign long-term capacity contracts. Thats validation of durability not hype. Risk check: capital intensity + execution. Mitigated by pre-sold capacity tier-1 customers and demand outpacing supply. High beta yes but thats where asymmetric infra returns come from."
X Link 2026-01-05T01:42Z [--] followers, [---] engagements
"The hard part wont be adoption itll be coordination. Were moving from scarcity of intelligence to scarcity of alignment across energy labor institutions and capital allocation. The bottleneck shifts from can we build it to can society absorb it without breaking feedback loops https://twitter.com/i/web/status/2007997671177388106 https://twitter.com/i/web/status/2007997671177388106"
X Link 2026-01-05T02:08Z [--] followers, [--] engagements
"@annapanart @Dr_Singularity Infinite choice without constraint doesnt produce freedom it produces paralysis. The next era isnt about removing limits its about deciding which constraints matter. Infrastructure energy and attention become the real governors of freedom"
X Link 2026-01-05T02:09Z [--] followers, [--] engagements
"@WatcherGuru New ATHs arent about headlines theyre about positioning and flows. ETFs passive demand and short-dated leverage create pressure before the breakout. Price moves first. Narratives follow"
X Link 2026-01-05T02:13Z [--] followers, [--] engagements
"@TheSonOfWalkley $TSLA price doesnt move on vibes it moves on positioning and flows. When momentum flips passive + options hedging do the heavy lifting. Fundamentals get debated after price"
X Link 2026-01-05T02:17Z [--] followers, [---] engagements
"Markets arent reacting to AI hype here theyre reacting to execution under uncertainty. The signal isnt Palantir per se its that AI-native decision systems are now embedded in real-world ops where latency coordination and reliability matter. Thats where AI value actually accrues. https://twitter.com/i/web/status/2008007965383778322 https://twitter.com/i/web/status/2008007965383778322"
X Link 2026-01-05T02:49Z [--] followers, [--] engagements
"The key mistake bears make is modeling $NBIS like a commodity GPU renter or a REIT. $NBIS is closer to an AI factory operator: Revenue mixes long-term contracts + variable/spot compute Margins come from orchestration utilization and reliability not just GPUs Power + grid access is the real moat ownership structure is secondary CapEx is the entry ticket. Returns come from sustained production-grade inference at scale. https://twitter.com/i/web/status/2008013550292545988 https://twitter.com/i/web/status/2008013550292545988"
X Link 2026-01-05T03:11Z [--] followers, 15.5K engagements
"@beauty_oe This point about concentration vs dispersion is important. Modeling $NBIS correctly means understanding it as an operator not an asset holder returns come from utilization orchestration and reliability over time not just GPU ownership"
X Link 2026-01-05T04:08Z [--] followers, [---] engagements
"GPU price hikes expose who actually has operating leverage vs. financial leverage. NeoClouds that win on access alone feel the pain immediately when input costs rise. Operators that win on utilization power discipline and reliability can absorb it. Short-term noise. Long-term signal especially for $NBIS. https://twitter.com/i/web/status/2008150971797377121 https://twitter.com/i/web/status/2008150971797377121"
X Link 2026-01-05T12:17Z [---] followers, [---] engagements
"Thats a fair nuance the build-out phase is where cost shocks actually matter most. The separation I keep coming back to is who bears that pressure: balance sheet leverage vs. operational leverage. Capex-heavy platforms feel memory inflation upfront but the ones with tighter deployment discipline and faster utilization ramps compress that pain window quickly. The weaker operators carry it forward as permanent margin drag. Same cost environment very different outcomes. https://twitter.com/i/web/status/2008160758119686470 https://twitter.com/i/web/status/2008160758119686470"
X Link 2026-01-05T12:56Z [--] followers, [--] engagements
"$NBIS works if you frame it as execution capacity not a chart pattern. The real upside comes from contracted AI factories power access and sustained inference throughput not multiple expansion alone. If utilization ramps as expected price follows fundamentals not trendlines. https://twitter.com/i/web/status/2008213835254296935 https://twitter.com/i/web/status/2008213835254296935"
X Link 2026-01-05T16:27Z [--] followers, [---] engagements
"Agree multiples usually move first. My point is less when price moves and more what determines whether that expansion holds. In AI infra execution capacity (power contracted demand sustained throughput) is what turns early multiple expansion into durable value rather than a round-trip. https://twitter.com/i/web/status/2008218896814489748 https://twitter.com/i/web/status/2008218896814489748"
X Link 2026-01-05T16:47Z [--] followers, [--] engagements
"This framing is spot on. The key insight most people miss: tokens are the unit of production for AI. Once you see AI as factories producing tokens the value accrues to whoever can deliver predictable throughput cost and uptime at scale not just flashy models. Thats the real $NBIS story. https://twitter.com/i/web/status/2008244354037743771 https://twitter.com/i/web/status/2008244354037743771"
X Link 2026-01-05T18:29Z [--] followers, [---] engagements
"The bigger takeaway isnt the automaker its the architecture. Scaled autonomy is fundamentally an inference-at-scale problem: continuous perception low-latency decisioning and massive uptime requirements. Avride Uber is a demand signal for AI factories that can deliver predictable throughput not just training bursts. Thats where the real leverage sits. https://twitter.com/i/web/status/2008246928644178407 https://twitter.com/i/web/status/2008246928644178407"
X Link 2026-01-05T18:39Z [--] followers, [---] engagements
"@munster_gene The updates matter but the real pivot is underneath them. [----] stops rewarding announcements and starts rewarding sustained inference revenue utilization throughput and cost control. Thats the shift from AI trade to AI industrialization"
X Link 2026-01-05T18:52Z [--] followers, [--] engagements
"@DanInvestingss Analyst PTs reflect capacity not hype. $NBIS is supply-constrained with long-term committed contracts and production-grade inference demand. This isnt optional AI spend its infrastructure. Multiple expansion follows utilization 📈"
X Link 2026-01-05T21:41Z [--] followers, [---] engagements
"This matters less for the silicon and more for the deployment model. NVL72 isnt about peak benchmarks its about dense reliable inference at scale. The real winners are clouds that can integrate hardware networking power and scheduling as one system. Thats where physical + agentic AI actually becomes usable. https://twitter.com/i/web/status/2008319577122636050 https://twitter.com/i/web/status/2008319577122636050"
X Link 2026-01-05T23:27Z [--] followers, [---] engagements
"@EndicottInvests Same. What stands out is execution not headlines. Turning next-gen GPUs into production inference predictable cost latency and throughput is where real value gets created. Thats the compounding edge for $NBIS alongside $NVDA"
X Link 2026-01-06T01:56Z [--] followers, [---] engagements
"That timeline makes sense. The hard part isnt raw silicon its making hardware networking and inference software behave as one deterministic system at scale. That integration window is where most of the real work (and differentiation) happens for platforms like $NVDA and for operators like $NBIS downstream. https://twitter.com/i/web/status/2008358602600563039 https://twitter.com/i/web/status/2008358602600563039"
X Link 2026-01-06T02:02Z [--] followers, [--] engagements
"Its less about using NVIDIA and more about where autonomy value concentrates. Most OEMs wont vertically integrate full-stack autonomy theyll standardize on a compute + tooling layer and differentiate above it. That dynamic strongly favors $NVDA regardless of who wins at the vehicle brand level. https://twitter.com/i/web/status/2008359382090690686 https://twitter.com/i/web/status/2008359382090690686"
X Link 2026-01-06T02:06Z [--] followers, [--] engagements
"@PatrickMoorhead The interesting part isnt peak specs its whether this rack can sustain full-speed data movement at scale. In production network topology power density and cooling efficiency end up determining realized throughput far more than headline FLOPs"
X Link 2026-01-06T03:10Z [--] followers, [---] engagements
"This is the key point many miss: Rubin NVL72 isnt just more GPUs it collapses system-level bottlenecks. [--] GPUs acting as one NVLink domain + HBM4 means: deterministic latency predictable cost per token sustained production-grade inference Thats why this matters for $NBIS specifically theyre building AI factories not demo clusters. Rubin turns inference from bursty industrial. Thats where real value accrues. 🏗🧠 https://twitter.com/i/web/status/2008391623873622450 https://twitter.com/i/web/status/2008391623873622450"
X Link 2026-01-06T04:14Z [--] followers, [---] engagements
"@jukan05 The real takeaway isnt peak specs its whether rack-scale systems can sustain utilization. CES announcements raise expectations. Markets will reprice on power density cooling memory bandwidth and software orchestration once these systems hit steady-state"
X Link 2026-01-06T11:41Z [--] followers, [--] engagements
"@unusual_whales This is the quiet part of the AI trade. GPUs scale demand instantly. Memory supply doesnt. When DRAM and HBM availability gate production AI economics stop being about chip headlines and start being about system-level constraints and utilization"
X Link 2026-01-06T11:43Z [--] followers, [--] engagements
"@simonkalouche Demos are the easy part. The hard part is deploying robots that run continuous inference with tight latency power and uptime requirements. Thats where AI shifts from spectacle to infrastructure economics"
X Link 2026-01-06T11:44Z [--] followers, [--] engagements
"@nvidia Specs are impressive but the real differentiator will be steady-state performance. Sustained inference memory bandwidth under load power density and uptime will decide cost per token in production not peak benchmarks"
X Link 2026-01-06T12:03Z [--] followers, [--] engagements
"Rubin being in production is the key point most people are glossing over. The gap isnt about peak benchmarks anymore its about who can deliver reliable cost-predictable inference at scale. Once youre shipping systems that actually run 24/7 networking memory power and software integration matter more than model demos. Thats why the competitive distance keeps widening. $NVDA https://twitter.com/i/web/status/2008511035201003719 https://twitter.com/i/web/status/2008511035201003719"
X Link 2026-01-06T12:08Z [--] followers, [--] engagements
"@mvcinvesting This is real capacity not a press-release KPI. 80MW = tangible revenue runway long-dated visibility and proof NBIS is executing ahead of demand not chasing it. 10x footprint expansion in [--] year tells you everything about where customer pull actually is. $NBIS"
X Link 2026-01-06T12:34Z [--] followers, [---] engagements
"The marketplace angle matters less as a headline and more as a utilization lever. If $NBIS can drive third-party demand onto already-built capacity thats margin expansion without proportional capex. Infrastructure stories dont reprice on announcements they reprice when utilization inflects. https://twitter.com/i/web/status/2008527069131710532 https://twitter.com/i/web/status/2008527069131710532"
X Link 2026-01-06T13:12Z [--] followers, [---] engagements
"@ArtByALKEBULAN @StockMKTNewz It doesnt change what $BTC is it changes who can access it. ETFs abstract custody and rails not the protocol. That likely strengthens the monetary thesis while weakening the ideological purity. Tradeoff not contradiction"
X Link 2026-01-06T13:54Z [---] followers, [--] engagements
"@PolymarketMoney Being early on $NVDA Rubin matters but the real edge is execution. H2 [----] is about lining up power cooling networking and software maturity so capacity actually ships into production not just installs. For $NBIS hardware access is table stakes; delivery is the moat"
X Link 2026-01-06T14:03Z [--] followers, [---] engagements
"@KobeissiLetter The dispersion matters. Revenue is scaling faster than deployable capacity. The bottleneck isnt demand its power cooling networking and time to production. Thats why infra growth looks lumpy while model revenue looks exponential"
X Link 2026-01-06T14:36Z [--] followers, [--] engagements
"@EricBalchunas @TheBlock__ This is less about crypto adoption and more about distribution control. ETFs dont change the protocol they change who allocates how size moves and where flows get warehoused. Morgan Stanley internalizing that demand is a business decision not an ideological one"
X Link 2026-01-06T15:04Z [--] followers, [--] engagements
"@AMD The question isnt peak TOPS its whether on-device NPUs meaningfully reduce cloud inference demand or just shift latency-sensitive workloads locally. Most real revenue still lives where utilization stays high"
X Link 2026-01-06T15:16Z [--] followers, [---] engagements
"This is the key shift. Agentic workloads dont just want faster compute they want persistent state and bandwidth at scale. HBM matters but the real arb is who can keep memory-heavy systems utilized day after day. Thats where infra operators with secured power + long-term contracts quietly win. $NVDA $MU $NBIS https://twitter.com/i/web/status/2008563776577302941 https://twitter.com/i/web/status/2008563776577302941"
X Link 2026-01-06T15:38Z [--] followers, [---] engagements
"@DeItaone [---] compute headlines miss the real constraint. The bottleneck isnt demand its power networking and cost per inference-hour. If compute needs to grow [---] it forces architectural efficiency gains not just more chips. Thats where infra discipline matters. $AMD $NVDA"
X Link 2026-01-06T16:13Z [--] followers, [---] engagements
"Which works better depends on scale and economics not just raw sensing. Lidar adds redundancy but it also adds cost calibration overhead and new failure modes at fleet scale. Teslas bet is that massive data + software generalization can outperform hardware redundancy per dollar deployed. Thats the real tradeoff. https://twitter.com/i/web/status/2008604863937347856 https://twitter.com/i/web/status/2008604863937347856"
X Link 2026-01-06T18:21Z [--] followers, [--] engagements
"This is a great example of optimization shifting from peak performance to sustained utilization. Warm-water liquid cooling isnt about eliminating cooling capex its about improving energy efficiency density and uptime at scale. Less energy spent on heat removal = more energy available for compute which matters far more in steady-state AI factories. The constraint just moves up the stack. https://twitter.com/i/web/status/2008605993031397415 https://twitter.com/i/web/status/2008605993031397415"
X Link 2026-01-06T18:26Z [---] followers, [----] engagements
"Lower cost per token and fewer GPUs per model dont mean less compute demand they usually expand it. As inference gets cheaper usage explodes workloads persist longer and utilization rises. Rubin is about shifting the curve from peak training events to continuous large-scale inference which ultimately drives more sustained demand for $NVDA compute not less. https://twitter.com/i/web/status/2008606493084709281 https://twitter.com/i/web/status/2008606493084709281"
X Link 2026-01-06T18:28Z [--] followers, [---] engagements
"The Roomba comparison highlights the core difference. Lidar works extremely well in constrained low-speed closed environments with limited edge cases. Driving is the opposite: open world long tail high speed adversarial conditions and constant distribution shift. Thats why the debate isnt does lidar work its which approach scales economically and operationally to millions of vehicles. https://twitter.com/i/web/status/2008608499706851532 https://twitter.com/i/web/status/2008608499706851532"
X Link 2026-01-06T18:35Z [--] followers, [--] engagements
"This reflects a broader shift in AI infrastructure: the unit of competition has moved from chips racks systems. At yotta-scale peak FLOPs matter less than sustained utilization networking power delivery cooling and the ability to evolve across generations. Thats where rack-level modular design becomes a strategic advantage. https://twitter.com/i/web/status/2008616617375322328 https://twitter.com/i/web/status/2008616617375322328"
X Link 2026-01-06T19:08Z [--] followers, [---] engagements
"Efficiency gains dont remove the constraint they raise the ceiling. Lower power per GPU enables higher density and utilization not lower total demand. Operators reinvest those gains into more GPUs per rack higher duty cycles and tighter clustering. Cooling savings matter but the steady-state outcome is still power-bound AI factories just with better economics faster scale-up and higher ROI per MW. Thats bullish for neo-clouds like $NBIS because it improves margins and accelerates capacity growth not because energy stops being the limiter. https://twitter.com/i/web/status/2008621353713487875"
X Link 2026-01-06T19:27Z [--] followers, [--] engagements
"Flat price action doesnt signal fading demand it signals investors anchoring to cycle timing instead of steady-state utilization. Blackwell + Rubin push the constraint from can we train to how efficiently can we run at scale. Thats about power density cooling and uptime not peak benchmarks. AI hardware isnt a [----] trade. Its a multi-year capacity absorption story. $NVDA just keeps moving the bottleneck up the stack. https://twitter.com/i/web/status/2008632074614247430 https://twitter.com/i/web/status/2008632074614247430"
X Link 2026-01-06T20:09Z [--] followers, [---] engagements
"Breaking above $100 isnt about momentum its about validation. $NBIS is positioning for sustained utilization not peak hype. Early access to $NVDA Rubin NVL72 = higher density better power efficiency and long-duration enterprise demand. AI factories reward uptime and scale not one-day candles. https://twitter.com/i/web/status/2008654445048185221 https://twitter.com/i/web/status/2008654445048185221"
X Link 2026-01-06T21:38Z [--] followers, [---] engagements
"Most advanced isnt about demos or model size its about deployment. $TSLA's edge is running one autonomy stack across millions of vehicles continuously training on real-world edge cases and closing the loop in production. That scale + feedback cycle is extremely hard to replicate. https://twitter.com/i/web/status/2008671299334484310 https://twitter.com/i/web/status/2008671299334484310"
X Link 2026-01-06T22:45Z [--] followers, [--] engagements
"The shift away from specs makes sense. In [----] the differentiator isnt peak performance its useful performance how efficiently systems deliver outcomes under real workloads. AI success is becoming about deployment power efficiency latency and integration into workflows not raw TOPS on a slide. https://twitter.com/i/web/status/2008678299862790384 https://twitter.com/i/web/status/2008678299862790384"
X Link 2026-01-06T23:13Z [---] followers, [--] engagements
"The key point isnt SSDs replacing HBM its memory hierarchy expansion. HBM stays the hot tier. What Jensen is signaling is that context + state are spilling beyond HBM forcing fast shared non-volatile memory layers to keep GPUs utilized. This is about sustaining throughput not substituting HBM. https://twitter.com/i/web/status/2008679349848076546 https://twitter.com/i/web/status/2008679349848076546"
X Link 2026-01-06T23:17Z [--] followers, [---] engagements
"This is exactly what time-shifting power looks like in practice. 80MW/320MWh isnt about energy arbitrage its about grid stability under peak and intermittent load. Storage like this is becoming critical as AI EVs and renewables all stress the same infrastructure. Megapack is quietly turning power into schedulable capacity. $TSLA https://twitter.com/i/web/status/2008686210802479182 https://twitter.com/i/web/status/2008686210802479182"
X Link 2026-01-06T23:44Z [--] followers, [--] engagements
"@AwareWealth Exactly. Inference density is where utilization power efficiency and latency economics matter most. Thats where platform + ecosystem winners compound not just model trainers. $NBIS $NVDA"
X Link 2026-01-06T23:59Z [--] followers, [---] engagements
"This is the part people miss: autonomy isnt a one-time training run its a continuous power-dense compute problem. $10B on GPUs just to train underscores that the bottleneck is sustained throughput energy efficiency and uptime not model cleverness. Thats why Tesla blends external GPUs with custom silicon: the constraint shifts to steady-state infrastructure economics. https://twitter.com/i/web/status/2008707539555545134 https://twitter.com/i/web/status/2008707539555545134"
X Link 2026-01-07T01:09Z [--] followers, [---] engagements
"@rohanpaul_ai Cool demo but the hard part isnt gesture recognition its robustness at scale. Hands-as-input break down under occlusion lighting variance latency and multi-user ambiguity. Until input is reliable across environments this stays a novelty UI not a workflow primitive"
X Link 2026-01-07T01:29Z [--] followers, [--] engagements
"@jenzhuscott @SharpaRobotics Hand dexterity matters but only if it holds up in steady-state. The real test isnt peak DOF or sensing density its calibration drift wear latency and failure recovery over millions of cycles. Thats where most robotic hands still struggle to leave the lab"
X Link 2026-01-07T01:33Z [--] followers, [---] engagements
"The metaphor works but the constraint doesnt disappear it just moves. GPUs create the gold but warehouses only matter if you can afford to keep it: power bandwidth retrieval latency and lifecycle costs dominate at scale. In steady state storage growth is as much an energy and ops problem as a hardware one. https://twitter.com/i/web/status/2008728662213095582 https://twitter.com/i/web/status/2008728662213095582"
X Link 2026-01-07T02:33Z [--] followers, [---] engagements
"@rohanpaul_ai This looks structural not cyclical. AI servers using 810x more DRAM shift the bottleneck from fabs to allocation as capacity moves to HBM/DDR5. Consumer price spikes are just the downstream symptom of AI factories absorbing steady-state supply"
X Link 2026-01-07T04:14Z [---] followers, [--] engagements
"This resonates. Tools compress iteration in domains you already understand they dont eliminate the need for judgment. The gap between feels faster and can be trusted end-to-end is where most of the hype breaks. Productivity gains are real but theyre uneven and domain-dependent not exponential across the board. https://twitter.com/i/web/status/2008758119820828873 https://twitter.com/i/web/status/2008758119820828873"
X Link 2026-01-07T04:30Z [--] followers, [---] engagements
"This looks less like users switching and more like where the query originates. Gemini inherits traffic from Search Chrome Android Workspace ChatGPT traffic is intentional app-level behavior. So share shifts reflect distribution plumbing not necessarily model preference or depth of usage. Different surfaces different workloads https://twitter.com/i/web/status/2008861654428369029 https://twitter.com/i/web/status/2008861654428369029"
X Link 2026-01-07T11:21Z [--] followers, [--] engagements
"This feels like a broader RL lesson not just diffusion: dont globally constrain the policy to protect against proxy reward failure. Apply constraints selectively where reward uncertainty spikes and let the rest of the distribution explore. Much better trade-off between quality diversity and learning speed. https://twitter.com/i/web/status/2008862023761998078 https://twitter.com/i/web/status/2008862023761998078"
X Link 2026-01-07T11:23Z [--] followers, [--] engagements
"@TheRayMyers The competitive question isnt peak performance its sustained deployment. Software maturity memory bandwidth power efficiency and ecosystem lock-in decide who actually scales inference at volume. Thats where the gap usually widens not narrows"
X Link 2026-01-07T11:38Z [--] followers, [---] engagements
"@WatcherGuru Spot ETFs matter less for short-term price than for long-duration demand stability. They turn ETH from a trading asset into balance-sheet infrastructure predictable flows lower volatility and better capital planning upstream. Thats when ecosystems actually scale"
X Link 2026-01-07T11:41Z [--] followers, [--] engagements
"@BillAckman The underappreciated part is that compounding only reveals value late it doesnt create it late. Most of the real work happens quietly in the flat years: survival reinvestment discipline and avoiding fatal mistakes. Miss that and the curve never shows up"
X Link 2026-01-07T11:43Z [--] followers, [--] engagements
"@jukan05 This looks less like demand softness and more like hard capacity triage. CoWoS and leading-edge nodes are being reallocated across higher-priority internal and external programs capping volumes even with interest intact. Constraint is upstream not workload. $MRVL"
X Link 2026-01-07T11:46Z [---] followers, [---] engagements
"Index-level multiples can be misleading here. Whats likely isnt a clean tech vs defensives rotation but dispersion within tech capital-light vs capital-intensive pricing power vs reinvestment-heavy models. At 23x the index may be capped but select cash-generative platforms still compound while others de-rate. https://twitter.com/i/web/status/2008873161090036210 https://twitter.com/i/web/status/2008873161090036210"
X Link 2026-01-07T12:07Z [--] followers, [---] engagements
"The headline growth is impressive but the key question is mix and durability. [----] deals tells you adoption breadth the next signal is ACV expansion renewal rates and how fast Agentforce attaches to core CRM seats. If this converts from pilots to scaled deployments the ARR trajectory matters a lot more than the initial YoY print. https://twitter.com/i/web/status/2008874271515595150 https://twitter.com/i/web/status/2008874271515595150"
X Link 2026-01-07T12:12Z [--] followers, [--] engagements
"AI factories dont win by peak speed they win by throughput and uptime. A [--] training speedup only matters if the factory can stay fed: memory bandwidth interconnect power & cooling sustained utilization Thats why the real moat isnt just GPUs its the system. $NVDA $MU $SMCI $VRT https://twitter.com/i/web/status/2008880936901833004 https://twitter.com/i/web/status/2008880936901833004"
X Link 2026-01-07T12:38Z [--] followers, [--] engagements
"@CyberRobooo Exactly and matching human sensing isnt enough. The bar is repeatability at scale: low noise low drift cheap sensors that survive millions of contacts. Until tactile stacks hit that reliability curve utilization not intelligence stays the binding constraint"
X Link 2026-01-07T13:08Z [---] followers, [--] engagements
"This is less about AI taking over and more about vertical integration. Proxy voting is a high-volume low-novelty decision pipeline exactly where internal models beat generic external advisors. The real shift is firms owning their own governance logic instead of outsourcing it. https://twitter.com/i/web/status/2008889919347757493 https://twitter.com/i/web/status/2008889919347757493"
X Link 2026-01-07T13:14Z [--] followers, [---] engagements
"@danielnewmanUV @haslindatv CES turning into a chips show makes sense ROI is now decided below the model layer. The real question isnt peak performance its utilization: power cooling networking and uptime at scale. Thats where AI capex either compounds or stalls"
X Link 2026-01-07T13:42Z [--] followers, [--] engagements
"@jukan05 Both can be true. HBM4 readiness can push early ramp timing by weeks without changing full-year shipment intent. Near-term cadence noise vs steady-state volume Jensen is talking end state JPM is flagging ramp friction"
X Link 2026-01-07T13:53Z [---] followers, [---] engagements
"Worth being careful on attribution here. Hiring is slowing less because AI is absorbing jobs and more because companies are optimizing for utilization and margins after years of headcount expansion. AI lets output grow without proportional hiring but thats an efficiency story first - not a demand collapse signal. https://twitter.com/i/web/status/2008901300889481245 https://twitter.com/i/web/status/2008901300889481245"
X Link 2026-01-07T13:59Z [--] followers, [--] engagements
"@AwareWealth The market keeps calling this uncertainty. It isnt. Capacity is already sold. Revenue just hasnt been recognized yet. That gap isnt risk - its an accounting delay. The only real question is how fast they can build. $NBIS"
X Link 2026-01-07T14:29Z [--] followers, [---] engagements
"@SawyerMerritt Q4 unit volatility matters less than what GM is actually building. This reads like: model transitions launch timing noise mix shifting upmarket The real question isnt EV units its software attach pricing power and margins over the next [--] months"
X Link 2026-01-07T15:55Z [--] followers, [--] engagements
"@caro_milanesi @Qualcomm @HP Exactly. AI value isnt the model its sustained outcomes under real constraints. Performance-per-watt thermals and battery life decide whether AI is always-on or just a demo feature. Devices are where the rubber meets the road"
X Link 2026-01-07T16:01Z [--] followers, [--] engagements
"This list shows where $NVDA's real leverage sits. Its not just selling GPUs its embedding itself across the AI stack. $CRWV for utilization + demand pull $APLD for power + data center leverage $ARM for long-cycle architecture control $NVDA is underwriting the ecosystem not just supplying it. https://twitter.com/i/web/status/2008951885944717740 https://twitter.com/i/web/status/2008951885944717740"
X Link 2026-01-07T17:20Z [--] followers, [---] engagements
"@DekmarTrades Right - NVDA isnt trading its underwriting the AI stack. Thats the real signal"
X Link 2026-01-07T17:34Z [---] followers, [--] engagements
"@NVIDIARobotics @CaterpillarInc This isnt a gimmick - its AI turning labor into throughput. Industrial AI monetizes uptime safety and utilization first. Thats where durable margins live. $NVDA $CAT"
X Link 2026-01-07T18:15Z [--] followers, [---] engagements
"@NVIDIARobotics This is the real bottleneck in robotics: not model intelligence but physically accurate simulation at scale. If sim synthetic data deployment isnt tightly integrated iteration speed collapses. OSMO is about throughput and reliability not flash. $NVDA"
X Link 2026-01-07T20:00Z [--] followers, [--] engagements
"This is the real damage of stop-start export policy: it breaks training continuity. Large-scale training and inference planning depends on predictable access to the same GPU class. Uncertainty forces over-stockpiling fragmented clusters and lower utilization - which hurts everyone not just buyers. https://twitter.com/i/web/status/2009001609104355563 https://twitter.com/i/web/status/2009001609104355563"
X Link 2026-01-07T20:38Z [--] followers, [---] engagements
"The key issue isnt insolvency headlines - its maturity mismatch. AI cloud models front-load capex (GPUs power leases) and back-load cash flow. When demand growth slows or financing tightens continuity breaks fast. Thats the real risk: utilization gaps + refinancing risk not demand disappearing. https://twitter.com/i/web/status/2009034220015067568 https://twitter.com/i/web/status/2009034220015067568"
X Link 2026-01-07T22:47Z [---] followers, [----] engagements
"AWS is the easy headline - the harder part is economics. Autonomous trucking scales only if: training + simulation costs stay predictable inference latency is consistent on the edge utilization stays high across fleets Cloud partnerships help with iteration speed but autonomy wins or loses on steady-state operations not demos. https://twitter.com/i/web/status/2009035080132616376 https://twitter.com/i/web/status/2009035080132616376"
X Link 2026-01-07T22:51Z [--] followers, [--] engagements
"On-device AI is where utilization finally matters more than peak FLOPs. Edge systems live and die by: power envelopes thermal headroom sustained throughput predictable latency Thats why embedded NPUs + tightly integrated stacks win - not because theyre flashy but because they stay on in real-world conditions. $AMD https://twitter.com/i/web/status/2009037213057863999 https://twitter.com/i/web/status/2009037213057863999"
X Link 2026-01-07T22:59Z [--] followers, [--] engagements
"@NVIDIADC Vera isnt about peak CPU performance - its about predictability. As clusters scale control-plane latency and scheduling efficiency decide utilization. This is NVIDIA hardening the economics of agentic AI not chasing benchmarks. $NVDA"
X Link 2026-01-07T23:32Z [--] followers, [---] engagements
"This is the same shift were seeing everywhere in AI infrastructure: from max context to efficient utilization. Dynamic context isnt just a UX win its an economics win. Lower token burn fewer truncation failures and more predictable long-run agent behavior. At scale stability + cost curves matter more than raw capability. https://twitter.com/i/web/status/2009050104519667841 https://twitter.com/i/web/status/2009050104519667841"
X Link 2026-01-07T23:50Z [--] followers, [--] engagements
"This is also why agent reliability improves as context gets smaller not larger. Less noise fewer hallucinated branches tighter execution. https://twitter.com/i/web/status/2009050941253927292 https://twitter.com/i/web/status/2009050941253927292"
X Link 2026-01-07T23:54Z [--] followers, [--] engagements
"@moninvestor Googles moat isnt just information. Its distribution + default placement + capital intensity. Search YouTube Android Maps create demand pull for models. Nvidia monetizes supply-side constraints. Different layers of the stack - both win but for very different reasons"
X Link 2026-01-08T00:57Z [--] followers, [---] engagements
"@PolymarketMoney Less a VC win and more a strategic hedge. Google didnt buy IRR it bought model optionality compute demand and ecosystem leverage as AI spend explodes. The real payoff isnt the paper valuation its securing position in a power- and compute-constrained world"
X Link 2026-01-08T01:02Z [--] followers, [--] engagements
"@MontanaMatos AI is shifting profit pools not ending incumbency. Googles advantage is infra + distribution at scale; Apples is device control and monetization. Market cap can move fast durable moats show up in cash flow and steady-state execution over time not single snapshots. $GOOGL"
X Link 2026-01-08T01:23Z [--] followers, [--] engagements
"@TheStalwart @SamRo The tenure dividend is really lower quit rates + wage stickiness showing up as margin expansion. It boosts operating leverage in slow labor markets but its cyclical - not a permanent cost advantage once mobility normalizes"
X Link 2026-01-08T01:24Z [--] followers, [--] engagements
"@rohanpaul_ai This isnt generosity - its labor market economics. When marginal researcher output scales non-linearly retaining top talent becomes a moat. Large equity pools are a rational response to extreme scarcity and capital intensity in frontier AI not a sign of excess"
X Link 2026-01-08T01:37Z [--] followers, [---] engagements
"@SawyerMerritt This was always the core tension. Nonprofit mission + frontier-scale compute + competitive timelines eventually collide with capital reality. The real question isnt intent - its whether governance structures can actually survive exponential infrastructure demands"
X Link 2026-01-08T01:58Z [--] followers, [---] engagements
"@moninvestor The headline isnt demand growth its where the constraint lands. Power availability grid interconnects and permitting timelines will decide which AI capacity actually gets built not GPU supply alone. Compute scales faster than energy infrastructure"
X Link 2026-01-08T02:37Z [--] followers, [---] engagements
"This reads less like demand risk and more like constraint migration. HBM4 readiness and licensing shift the [----] mix but CoWoS consumption stays bounded by packaging + OSAT throughput not wafer starts. If N4 capacity and export approvals clear the upside shows up as delayed CoWoS acceleration rather than lost volume. Steady-state infra constraints keep re-ordering the ramp. https://twitter.com/i/web/status/2009096555815686385 https://twitter.com/i/web/status/2009096555815686385"
X Link 2026-01-08T02:55Z [---] followers, [---] engagements
"Really impressive control and mobility. Whats exciting is less the demo itself and more what it signals - steady gains in autonomy power efficiency and reliability that eventually make real-world deployment viable. Robotics progress compounds quietly before it shows up everywhere. https://twitter.com/i/web/status/2009102932005048556 https://twitter.com/i/web/status/2009102932005048556"
X Link 2026-01-08T03:20Z [--] followers, [--] engagements
"@StockMKTNewz Market cap flips are the effect not the cause. $GOOGL is being priced as an AI services + infra compounder while $AAPL is still anchored to hardware cycles. Different growth physics"
X Link 2026-01-08T04:50Z [--] followers, [---] engagements
"@TheTechInvest The logos get attention but execution is the real test. Being tied to $NVDA / $MSFT / $AMD only matters if power capacity and uptime are already lined up. Thats when partner turns into real utilization. $IREN"
X Link 2026-01-08T11:28Z [---] followers, [---] engagements
"@TopTechStocksUS Logos are great but the real edge is execution. If $NBIS can actually deliver power uptime and scale while demand is there thats when partnerships turn into revenue - not before"
X Link 2026-01-08T11:45Z [---] followers, [---] engagements
"@lisaabramowicz1 @TheTerminal Record borrowing isnt panic - its optionality. Issuers are locking in funding while windows are wide spreads are tight and uncertainty is political not credit-driven. The tell wont be volumes - itll be who stops coming to market first when conditions turn"
X Link 2026-01-08T11:48Z [--] followers, [--] engagements
"AI investing is shifting from whats possible to what actually runs. Chips like $NVDA set the pace but the real question this year is whether the infrastructure can keep up. Power uptime and utilization decide who actually compounds. The story gets less exciting here and thats usually where the money is made. https://twitter.com/i/web/status/2009232017385508890 https://twitter.com/i/web/status/2009232017385508890"
X Link 2026-01-08T11:53Z [--] followers, [--] engagements
"@StockSavvyShay @fiscal_ai Key assumption is sustained supply constraint - margins only hold if capacity additions lag demand longer than expected. Do you see that persisting"
X Link 2026-01-08T12:00Z [--] followers, [---] engagements
"@yianisz Agreed. Is your confirmation signal sustained volume expansion on the next leg or simply higher lows holding through consolidation"
X Link 2026-01-08T12:05Z [--] followers, [----] engagements
"The China headline sounds huge but it feels mostly priced in at this point. If demand everywhere else is already capped by supply losing that slice doesnt really change the near-term story - it just changes whos at the front of the line. To me the bigger question is how long that supply tightness lasts not the China math. https://twitter.com/i/web/status/2009237070531739815 https://twitter.com/i/web/status/2009237070531739815"
X Link 2026-01-08T12:13Z [--] followers, [--] engagements
"@yianisz That makes sense. Absorption first then expansion - otherwise its just noise. Watching whether those higher lows actually stick before getting excited about the breakout"
X Link 2026-01-08T12:16Z [--] followers, [---] engagements
"@NVIDIADRIVE @nvidia @MercedesBenz This is the unsexy part that actually matters. If L2/L2+ works smoothly in real streets people trust it - and thats what scales"
X Link 2026-01-08T22:18Z [--] followers, [---] engagements
"@The_AI_Investor Selling inferior - it usually means timing and scale matter more than peak benchmarks. Inference wins are about software ops and deployment friction not just SRAM or raw silicon"
X Link 2026-01-08T23:44Z [--] followers, [---] engagements
"@rohanpaul_ai This looks less like a clash of ideas and more like a shift in what the org is optimizing for. Research priorities tend to change fast once scale and product pressure kick in"
X Link 2026-01-09T01:50Z [--] followers, [--] engagements
"Hot take that probably shouldnt be a hot take: AI doesnt fail because the model isnt smart enough. It fails because something boring breaks. Power. Uptime. Throughput. Utilization. The interesting part usually starts after the demo works when everything has to run constantly. Thats usually what I look for when Im thinking about where real value compounds. https://twitter.com/i/web/status/2009465791104594062 https://twitter.com/i/web/status/2009465791104594062"
X Link 2026-01-09T03:22Z [--] followers, [--] engagements
"@kyleichan Feels less like a sudden shift and more like momentum showing up in public. Once companies are ready they show up geopolitics or not"
X Link 2026-01-09T05:04Z [--] followers, [---] engagements
"@rwang07 Memory isnt just tight its the pacing item now. When HBM and advanced DRAM are constrained everything upstream slows: GPUs servers deployment schedules. Thats where utilization and pricing power really show up"
X Link 2026-01-09T12:00Z [--] followers, [---] engagements
"@qualtrim This isnt just growth its operating leverage finally showing up. Once cloud becomes steady-state infrastructure margins expand quietly and persistently"
X Link 2026-01-09T12:56Z [--] followers, [--] engagements
"@dlacalle_IA True but rising wages can still coexist with weaker hiring breadth. Fewer sectors adding jobs means gains concentrate even as averages look healthy"
X Link 2026-01-09T15:25Z [---] followers, [--] engagements
"@caro_milanesi @intel @HP @Microsoft Yeah this keeps showing up. The models arent the bottleneck as much as the messy last mile: perception dexterity reliability and knowing when to get out of the way. CES demos make that gap very visible"
X Link 2026-01-09T20:06Z [---] followers, [--] engagements
"@danielisdizzy That combo can coexist for a while. Growth can look strong when gains concentrate in a few capital-intensive sectors even as hiring breadth thins elsewhere. The question is how wide that diffusion actually gets"
X Link 2026-01-09T20:10Z [---] followers, [--] engagements
"@MikeZaccardi This is the disconnect people miss. Headline prints can look fine while the underlying job engine is already below breakeven. Not a crash signal just much less margin for strong labor narratives"
X Link 2026-01-09T21:11Z [---] followers, [--] engagements
"@HedgieMarkets Feels more like a slow correction from pandemic over-hiring than AI replacing engineers. The spend shift and the layoffs are correlated in time not necessarily in cause"
X Link 2026-01-09T21:14Z [---] followers, [---] engagements
"Todays jobs report doesnt really refute this it mostly shows the lag. Payrolls and unemployment tell you where the labor market was. The jobs-workers gap tells you how much buffer is left. When that buffer disappears growth can continue but it gets much more fragile to shocks. https://twitter.com/i/web/status/2009769640894124480 https://twitter.com/i/web/status/2009769640894124480"
X Link 2026-01-09T23:29Z [---] followers, [--] engagements
"Nothing that mattered today actually broke. Jobs cooled without collapsing. Wages didnt re-accelerate. AI keeps shifting from whats possible to what actually runs. Thats why markets arent debating growth vs recession"
X Link 2026-01-09T23:44Z [---] followers, [--] engagements
"@JosephPolitano The headline rate keeps saying fine but this shows the pressure shifting sideways instead of snapping. Fewer people are fully unemployed more are stuck on the margins. Thats a quieter kind of weak"
X Link 2026-01-10T04:10Z [---] followers, [--] engagements
"@greg_ip I wonder if this is mostly a timing story. Tariffs hit prices fast but reshoring shows up only after long capex and permitting cycles. In the interim softer construction demand can easily swamp any early gains"
X Link 2026-01-10T04:17Z [---] followers, [---] engagements
"@Schuldensuehner Digitalization is part of it but the deeper gap is utilization. The US reorganized firms capital and workflows around software. Germany digitized existing processes without re-wiring incentives. Same tools different throughput. Productivity follows structure"
X Link 2026-01-10T12:16Z [---] followers, [---] engagements
"Markets arent reacting to data surprises right now. Theyre reacting to constraint clarity. Growth isnt breaking Inflation isnt re-accelerating Liquidity isnt tightening When nothing snaps capital stops hiding. https://twitter.com/i/web/status/2009969891361861722 https://twitter.com/i/web/status/2009969891361861722"
X Link 2026-01-10T12:45Z [---] followers, [--] engagements
"Big prior-year returns matter less than what reset underneath them. In past cycles intl small value kept running when currencies stabilized valuations stayed below trend and earnings breadth widened not just because momentum persisted. The setup matters more than the headline gain. https://twitter.com/i/web/status/2009987050519572745 https://twitter.com/i/web/status/2009987050519572745"
X Link 2026-01-10T13:53Z [---] followers, [---] engagements
"@robin_j_brooks This chart really shows the asymmetry. Markets can price cuts but as long as growth and services inflation dont crack the Feds bias stays tighter for longer. That tension not politics is what keeps the USD pressure story alive"
X Link 2026-01-10T13:57Z [---] followers, [---] engagements
"@robin_j_brooks That pressure may rise agreed. But markets still need cover. As long as growth and services inflation dont actually crack cuts stay conditional not automatic. Politics can lean but constraints still decide timing"
X Link 2026-01-10T14:20Z [---] followers, [--] engagements
"@MikeZaccardi @StockCharts Yeah energy helps the comps but its a narrow window. Once base effects fade CPI momentum is going to hinge more on shelter and services than oil alone. A brief 2-handle wouldnt change the underlying glide path"
X Link 2026-01-10T17:01Z [---] followers, [--] engagements
"@dailychartbook @BlackRock The key detail is where that growth is coming from. This is capital-intensive buildout showing up as investment not a broad productivity surge yet which also means its power- supply-chain- and balance-sheet constrained. Durable but not frictionless"
X Link 2026-01-10T17:07Z [---] followers, [--] engagements
"@dailychartbook This isnt really about retirement culture its about labor composition and utilization. Countries with older exit ages tend to have higher participation less churn and slower productivity dilution. That quietly matters for growth and inflation dynamics"
X Link 2026-01-10T20:08Z [---] followers, [--] engagements
"@neilksethi Thats a real tell this close to earnings. The next question is how broad those revisions are when a few heavyweights do the lifting headline growth looks fine but durability is thinner"
X Link 2026-01-10T20:49Z [---] followers, [--] engagements
"@curious_founder Whats striking is how persistent this pattern is across revolutions. Early gains concentrate in capital and profits; labor share recovers only after institutions bargaining power and diffusion catch up. Were still in that lag phase with AI"
X Link 2026-01-11T02:10Z [---] followers, [--] engagements
"This is the shift hiding in plain sight. U.S. electricity demand growth has broken out of a decade-long flat regime and remains elevated. Thats why power siting and firm generation are becoming the real constraints not chips or models"
X Link 2026-01-11T02:45Z [---] followers, [---] engagements
"@Barchart The Buffett Indicator tells you where market cap sits relative to GDP not where stress builds. What matters now is whether the system can carry that valuation: fiscal support passive flows and capital-dense growth keep pressure off until a real constraint binds"
X Link 2026-01-11T04:03Z [---] followers, [---] engagements
"@EugeneNg Tech employment peaking doesnt contradict AI it fits it. Capital is shifting from labor-heavy software to compute- power- and infra-intensive layers. Fewer hires more throughput. Structural looks more plausible than cyclical here"
X Link 2026-01-11T12:13Z [---] followers, [--] engagements
"@SKundojjala This comparison really highlights the shift in value creation. NVIDIAs growth isnt cyclical demand it reflects a structural reallocation of capital toward compute- and power-intensive infrastructure while mature consumer franchises remain volume-bound"
X Link 2026-01-11T12:19Z [---] followers, [---] engagements
"@dlacalle_IA Agreed but the dynamism isnt just growth its capital velocity. The U.S. reallocates faster when constraints shift: labor capital software infrastructure. That flexibility is what keeps productivity compounding even as cycles slow"
X Link 2026-01-11T12:24Z [---] followers, [--] engagements
"@StockSavvyShay Agreed and the edge isnt just growth its capital efficiency. YouTube lets Google deploy AI at scale with zero behavior change turning distribution into a structural compounding advantage"
X Link 2026-01-11T12:30Z [---] followers, [---] engagements
"@dlacalle_IA Currency collapse despite exports is the tell. FX prices confidence not cash flow. When capital controls sanctions risk and policy credibility erode trade surpluses and external support cant anchor the currency. Growth without trust doesnt clear in FX"
X Link 2026-01-11T12:49Z [---] followers, [--] engagements
"Feels less like a causal dollar story and more like regime context. A weaker USD only helps equities when growth is steady enough to absorb it not when leadership is being second-guessed by surveys. NVIDIA turning usually isnt a macro call so much as a utilization and capex reality check. https://twitter.com/i/web/status/2010406683751891129 https://twitter.com/i/web/status/2010406683751891129"
X Link 2026-01-11T17:41Z [---] followers, [--] engagements
"@curious_founder The scale gap is real. The harder part now is what happens after install. Turning capacity into usable power comes down to grids storage and utilization thats where the bottleneck is shifting"
X Link 2026-01-11T19:49Z [---] followers, [---] engagements
"The numbers make sense together. Renewables are scaling fast but demand growth + grid constraints mean theyre still largely additive. Displacement only accelerates once storage transmission and industrial load shifting scale in parallel. Capacity is the easier part. Coordination is the bottleneck. https://twitter.com/i/web/status/2010444326749515864 https://twitter.com/i/web/status/2010444326749515864"
X Link 2026-01-11T20:10Z [---] followers, [--] engagements
"Thats the key point returns followed earnings not multiple expansion. Koreas cycle has been driven by export-heavy capital-intensive sectors where margins recover fast once utilization turns. The question now isnt valuation its whether that earnings durability persists as global capex normalizes. https://twitter.com/i/web/status/2010450659104247932 https://twitter.com/i/web/status/2010450659104247932"
X Link 2026-01-11T20:36Z [---] followers, [---] engagements
"@MikeZaccardi This is a timing distortion more than a growth impulse. Shutdowns pull activity forward/back across quarters so the headline swing matters less than the underlying trend. Markets usually look through this once its clear nothing structural changed"
X Link 2026-01-11T20:43Z [---] followers, [--] engagements
"@RyanDetrick @SamRo Striking chart but its as much composition as productivity. Capital-intensive leaders drive revenue per worker higher as utilization and margins scale even if median firms dont see the same gains. That concentration is the tell"
X Link 2026-01-11T20:51Z [---] followers, [---] engagements
"Markets dont move on headlines. They move when constraints become undeniable. Capital freezes when exit risk turns structural Growth survives when utilization not optimism holds Uninvestable isnt political; its learned behavior I focus on where systems quietly break: power compute labor capital mobility and how markets reprice before consensus notices. If you care about second-order effects not surface narratives youre in the right place. https://twitter.com/i/web/status/2010508471591191005 https://twitter.com/i/web/status/2010508471591191005"
X Link 2026-01-12T00:25Z [---] followers, [---] engagements
"What matters less than the headline odds is what they reveal about uncertainty. Markets dont need Powell to actually leave to reprice they move when institutional continuity becomes a variable. Rising probabilities translate into higher term premia wider risk buffers and a shift from growth narratives toward credibility and governance risk. Thats the real signal here. https://twitter.com/i/web/status/2010688118161391711 https://twitter.com/i/web/status/2010688118161391711"
X Link 2026-01-12T12:19Z [---] followers, [--] engagements
"Whats notable is how gold is outperforming not as a spike on panic but as a steady relative grind. That tends to show up when confidence in policy frameworks erodes at the margin. Less about immediate recession risk more about credibility term premia and the cost of uncertainty rising quietly across assets. https://twitter.com/i/web/status/2010693006022840609 https://twitter.com/i/web/status/2010693006022840609"
X Link 2026-01-12T12:39Z [---] followers, [--] engagements
"@EricBalchunas Feels less like speculative risk-on and more like structural allocation behavior. Broad beta cash-adjacent ETFs and liquidity preference dominating not a chase but positioning. Those flows usually stick until a real constraint snaps"
X Link 2026-01-12T12:43Z [---] followers, [---] engagements
"@TheStalwart Once an institution is running real P&L independence becomes conditional whether courts admit it or not. Markets dont wait for constitutional clarity they price loss socialization and governance uncertainty immediately. Thats the channel showing up in term premia and gold"
X Link 2026-01-12T12:53Z [---] followers, [--] engagements
"@robin_j_brooks @TheStalwart Breakevens dont lead regimes they tend to lag credibility breaks. Markets wait for realized inflation to force belief then reprice fast. Until then they anchor to the last cycle"
X Link 2026-01-12T13:50Z [---] followers, [--] engagements
"@RyanDetrick @sonusvarghese This looks less like broad labor weakness and more like entry-level friction. Firms are hoarding experienced labor while slowing new intake so the pain concentrates at the margin without showing up in headline unemployment"
X Link 2026-01-12T14:54Z [---] followers, [---] engagements
"@JKempEnergy @JesseJenkins This looks less like gasoline prices not mattering and more like efficiency absorbing the shock. Miles driven keep rising but demand elasticity is muted because each mile just uses less fuel than it used to"
X Link 2026-01-12T15:46Z [---] followers, [---] engagements
"@Convertbond Its not just a rotation out of bonds its a repricing of what qualifies as duration when fiscal supply is elastic and real yields are volatile. Gold absorbs duration risk without counterparty or policy exposure. Thats the quiet shift"
X Link 2026-01-12T18:41Z [---] followers, [--] engagements
"@KevRGordon This gap is mostly composition. Net worth gains are highly concentrated and asset-heavy while day-to-day sentiment is driven by prices cash flow and housing affordability. Aggregate wealth can rise even as median experience deteriorates"
X Link 2026-01-12T19:03Z [---] followers, [--] engagements
"@chigrl The signal here isnt that next-gen nuclear is suddenly ready its how scarce firm power has become. Hyperscalers are buying optionality and learning time because grid timelines are the real constraint not reactor physics"
X Link 2026-01-12T20:18Z [---] followers, [---] engagements
"@robin_j_brooks What jumps out is how mechanical this looks. When global duration and safe collateral are scarce USTs clear even without a premium. Less a verdict on the Fed more a signal about balance-sheet constraints"
X Link 2026-01-12T20:44Z [---] followers, [---] engagements
"I think the tension is that markets arent really pricing cuts theyre pricing duration suppression. Hard assets can run in a world where policy rates stay higher but real rates get capped through balance sheets regulation and debt management. Thats financial repression not a classic easing cycle"
X Link 2026-01-12T21:43Z [---] followers, [---] engagements
"The distinction vs. the mid-2000s is that todays housing market is constrained by stock and flow not leverage. Household balance sheets are stronger but rate lock-in and low turnover have impaired price discovery. That can support prices without a debt boom yet still create real affordability stress. https://twitter.com/i/web/status/2010880061625237820 https://twitter.com/i/web/status/2010880061625237820"
X Link 2026-01-13T01:02Z [---] followers, [--] engagements
"@BobEUnlimited The key constraint is that households experience inflation through cash-flow timing not averages. Even if policy looks through it recurring price resets on essentials compress real disposable income long before labor or nominal growth can offset it"
X Link 2026-01-13T11:27Z [---] followers, [--] engagements
"@dailychartbook @neilksethi Looks less like a true rotation and more like risk-budget inertia. Mega-cap growth stays overweight because it still absorbs uncertainty better than cyclicals even as positioning bleeds at the margin"
X Link 2026-01-13T11:32Z [---] followers, [--] engagements
"@MikeZaccardi @StockCharts Total return tells the story here FX and duration are doing the work. Its another signal that global real rates are being capped even as policy paths diverge"
X Link 2026-01-13T11:40Z [---] followers, [--] engagements
"What makes this unusual isnt just the price action but the signal it sends about policy credibility. If yields rise and the currency still weakens markets may be questioning whether normalization can occur without destabilizing balance sheets. Thats a constraint not a cyclical move. https://twitter.com/i/web/status/2011041228775936014 https://twitter.com/i/web/status/2011041228775936014"
X Link 2026-01-13T11:42Z [---] followers, [----] engagements
"@neilksethi @dailychartbook The asymmetry is the point. CTAs dont need to be bearish to be a risk flat to modestly higher prices dont add much fuel but downside breaches trigger large mechanical selling. Thats convexity embedded in positioning not sentiment"
X Link 2026-01-13T11:45Z [---] followers, [--] engagements
"What stands out isnt just fear of disruption its how little conviction there is in spot. Record upside hedging alongside a soft curve suggests participants want insurance not exposure. Thats classic constraint pricing: limited spare capacity + geopolitics force volatility bids even when the base case still points to surplus. https://twitter.com/i/web/status/2011042556612919637 https://twitter.com/i/web/status/2011042556612919637"
X Link 2026-01-13T11:48Z [---] followers, [---] engagements
"This framing matters. What gets lost is that restraint on hiring isnt a macro pessimism signal its a capital allocation choice. Firms are prioritizing durability (AI M&A optionality buybacks) over labor flexibility even with strong earnings. That shift has real human consequences but its driven more by balance-sheet logic than fear. https://twitter.com/i/web/status/2011043938145366284 https://twitter.com/i/web/status/2011043938145366284"
X Link 2026-01-13T11:53Z [---] followers, [---] engagements
"@RonStoeferle @kevinmuir This is the tell: price is breaking out but options arent chasing it. Calm upside IV and offered call skew say this is allocation and stress hedgingnot FOMO. Gold peaks when upside insurance gets crowded; disbelief is still in charge"
X Link 2026-01-13T12:14Z [---] followers, [---] engagements
"This reads less like credit deterioration and more like optionality. Strong profits allow banks to over-reserve early which cushions future shocks and protects capital if the cycle turns. Stress shows up in charge-offs not provisions. Provisions buy time. Charge-offs reveal truth"
X Link 2026-01-13T12:20Z [---] followers, [---] engagements
"@KevRGordon Big jump. Feels less like taxes suddenly rose and more like margins finally got thin enough that they hurt. When growth slows everything fixed becomes the problem"
X Link 2026-01-13T13:06Z [---] followers, [--] engagements
"CPI didnt surprise today but the path still matters more than the print. Inflation is calmer than it was not back to the old regime either. December CPI #inflation +2.7% vs. +2.7% est & +2.7% prior core +2.6% vs. +2.7% est. & +2.6% prior https://t.co/R3kcxvyULj December CPI #inflation +2.7% vs. +2.7% est & +2.7% prior core +2.6% vs. +2.7% est. & +2.6% prior https://t.co/R3kcxvyULj"
X Link 2026-01-13T13:37Z [---] followers, [--] engagements
"@KevRGordon This reads less hawkish than it sounds. Well positioned is Fed code for optionality not conviction. If inflation keeps behaving the hurdle for easing is time not resistance"
X Link 2026-01-13T15:16Z [---] followers, [--] engagements
"@conorsen That framing feels right. When population growth slows growth shows up as income dispersion more than headcount expansion. Youre not creating more middle class jobs youre asking existing households to climb and not everyone can"
X Link 2026-01-13T16:07Z [---] followers, [---] engagements
"@RealEJAntoni Feels like less churn not less leverage. Switching premiums have compressed but stayers are still seeing gains which points to tighter matching and fewer outside options. Stable yes but bargaining power is narrowing at the margin"
X Link 2026-01-13T16:11Z [---] followers, [--] engagements
"@BobEUnlimited @albertocavallo What this really shows is a level reset not persistent inflation. Tariffs tend to push prices up once then sit in the index rather than unwind. That still matters for how inflation feels even if the rate cools"
X Link 2026-01-13T16:17Z [---] followers, [---] engagements
"ELI5: instead of a flat margin number CME Group will now scale margin with price. If gold goes up required margin goes up automatically. Its just about keeping risk proportional as prices move not signaling direction. BREAKING: CME Group will set gold silver platinum and palladium margins as a percentage of notional value instead of a fixed dollar amount BREAKING: CME Group will set gold silver platinum and palladium margins as a percentage of notional value instead of a fixed dollar amount"
X Link 2026-01-13T16:27Z [---] followers, [--] engagements
"@LizThomasStrat Early days but this reads less like growth optimism and more like pricing power and scarcity. Materials and gold doing well together usually says something about constraints not animal spirits"
X Link 2026-01-13T18:30Z [---] followers, [--] engagements
"@robin_j_brooks It might be less a break in inflation dynamics and more a break in measurement continuity. Missing a month matters when so much of core is smoothed lagged or imputed. Hard to know if inflation slowed or if the signal just got noisier"
X Link 2026-01-13T19:02Z [---] followers, [---] engagements
"@KoyfinCharts This is the one price everyone actually feels. You dont notice electricity until it stops being cheap then it quietly shows up everywhere rent groceries running a business. Its the kind of cost that doesnt make headlines but slowly changes behavior"
X Link 2026-01-13T19:09Z [---] followers, [---] engagements
"@IrvingSwisher This is the inflation people actually feel. Restaurants sit at the intersection of wages rent and demand so prices dont come down easily. Even as goods cool eating out keeps reminding people inflation isnt really gone"
X Link 2026-01-13T19:40Z [---] followers, [---] engagements
"@LizAnnSonders New home sales can look stable even as prices fall because builders are the shock absorber. They adjust incentives and mix to keep deals moving. Thats flexibility not a broad demand rebound"
X Link 2026-01-13T20:35Z [---] followers, [---] engagements
"@joshrauh Capital isnt interchangeable here. Project finance looks cheap on paper but it pushes risk around. When you actually need power on time the balance sheet matters more than the coupon"
X Link 2026-01-13T21:00Z [---] followers, [---] engagements
"@EPBResearch This is the frustrating part of this cycle. You think you see cyclical hiring come back then the revisions walk it back. More uncertainty than momentum"
X Link 2026-01-13T21:32Z [---] followers, [---] engagements
Topic Creators
https://lunarcrush.com/api4/public/topic/:topic/creators/v1
Get the top creators for a social topic
input parameters:
- topic: Provide the topic to get details for. A topic must be all lower case and can only include letters, numbers, spaces, # and $. required
Example request:
curl -H "Authorization: Bearer <API_KEY>" https://lunarcrush.com/api4/public/topic/bitcoin/creators/v1
Example response:
{
"data": [
{
"creator_id": "twitter::1605310254085980161",
"creator_name": "Bitcoin_Teddy",
"creator_avatar": "https://pbs.twimg.com/profile_images/1739753415134138368/FrWHMAnb_200x200.jpg",
"creator_followers": 74293,
"creator_rank": 1,
"interactions_24h": [-------]
}
]
}
Schema:
- creator_id: The [network]::[unique_id] for the influencer
- creator_name: The unique screen name for the influencer
- creator_avatar: The URL to the avatar for the creator
- creator_followers: number of followers the account has
- creator_rank: ranking based on all posts in the last [--] hours that have interactions
- interactions_24h: Number of interactions in the last [--] hours
Creators List
https://lunarcrush.com/api4/public/creators/list/v1
Get a list of trending social creators over all of social based on interactions. To get lists of creators by category or topic see the topics and categories endpoints.
Example request:
curl -H "Authorization: Bearer <API_KEY>" https://lunarcrush.com/api4/public/creators/list/v1
Example response:
{
"data": [
{
"creator_name": "elonmusk",
"creator_display_name": "Elon Musk",
"creator_id": "44196397",
"creator_network": "twitter",
"creator_avatar": "https://pbs.twimg.com/profile_images/2008546467615580160/57KcqsTA_200x200.jpg",
"creator_followers": 234295750,
"creator_posts": 3649,
"creator_rank": 1,
"interactions_24h": [---------]
}
]
}
Schema:
- creator_name: The unique screen name for the influencer
- creator_display_name: The chosen display name for the influencer if available
- creator_id: The [network]::[unique_id] for the influencer
- creator_network: The social network for the post or influencer. We still refer to x as twitter out of developer preference.
- creator_avatar: The URL to the avatar for the creator
- creator_followers: number of followers the account has
- creator_posts: total number of posts with interactions in the last [--] hours
- creator_rank: ranking based on all posts in the last [--] hours that have interactions
- interactions_24h: Number of interactions in the last [--] hours
Creator
https://lunarcrush.com/api4/public/creator/:network/:id/v1
Get detail information on a specific creator
input parameters:
- network: Provide the network for the creator. One of twitter, youtube, instagram, reddit, or tiktok required
- id: Provide the unique ID or screen name of the creator required
Example request:
curl -H "Authorization: Bearer <API_KEY>" https://lunarcrush.com/api4/public/creator/twitter/elonmusk/v1
Example response:
{
"data": {
"creator_id": "twitter::44196397",
"creator_name": "elonmusk",
"creator_display_name": "Elon Musk",
"creator_avatar": "https://pbs.twimg.com/profile_images/2008546467615580160/57KcqsTA_200x200.jpg",
"creator_followers": 234296337,
"creator_rank": 2,
"interactions_24h": 318539577,
"topic_influence": [
{
"topic": "in the",
"count": 294,
"percent": 5.88,
"rank": [--]
}
],
"top_community": [
{
"creator_name": "grok",
"creator_display_name": "Grok",
"creator_avatar": "https://pbs.twimg.com/profile_images/1893219113717342208/Vgg2hEPa_200x200.jpg",
"count": [---]
}
]
}
}
Schema:
- creator_id: The [network]::[unique_id] for the influencer
- creator_name: The unique screen name for the influencer
- creator_display_name: The chosen display name for the influencer if available
- creator_avatar: The URL to the avatar for the creator
- creator_followers: number of followers the account has
- creator_rank: ranking based on all posts in the last [--] hours that have interactions
- interactions_24h: Number of interactions in the last [--] hours
- topic_influence: an array of social topics and the creators ranking on each topic
- topic: LunarCrush social topic. Can only includes letters, numbers, spaces, #, and $
- top_community: an array of the top accounts that have recently mentioned this creator or that this creator has mentioned
Creator Time Series
https://lunarcrush.com/api4/public/creator/:network/:id/time-series/v1
Get time series data on a creator.
input parameters:
- network: Influencer social network required
- id: The unique id or screen name of the creator required
- bucket: bucket time series data into hours or days. default is hours.
- interval: Use interval to specify the start and end time automatically for convenience. If "start" or "end" parameters are provided this parameter is ignored.
- start: The start time (unix timestamp) to go back to.
- end: The end time (unix timestamp) to stop at.
Example request:
curl -H "Authorization: Bearer <API_KEY>" https://lunarcrush.com/api4/public/creator/twitter/lunarcrush/time-series/v1
Example response:
{
"config": {
"network": "twitter",
"influencer_id": "twitter::988992203568562176",
"interval": "1w",
"start": 1769990400,
"end": 1770681600,
"bucket": "hour",
"name": "lunarcrush",
"remote_api": "danode1-13",
"generated": [----------]
},
"data": [
{
"time": 1769990400,
"followers": 305326,
"interactions": 717,
"posts_active": 14,
"creator_rank": [------]
}
]
}
Schema:
- config: This includes the inputs for the request processed by the server and may include additional hints about the request and response information.
- network: The social network for the post or influencer. We still refer to x as twitter out of developer preference.
- interval: Typically used for specifying time intervals like 1w = [--] week, 1m = [--] month etc
- start: Start/from unix timestamp (in seconds)
- end: End/to unix timestamp (in seconds)
- bucket: Data is generally bucketed into hours or days
- name: The full name of the asset
- generated: A unix timestamp (in seconds) when the data was generated to understand possibly stale data
- time: A unix timestamp (in seconds)
- followers: The number of publicly displayed followers the creator has
- interactions: number of all publicly measurable interactions on a social post (views, likes, comments, thumbs up, upvote, share etc)
- posts_active: number of unique social posts with interactions
- creator_rank: ranking based on all posts in the last [--] hours that have interactions
Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing