[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.] #  @D_K_Rajasekar DKR DKR posts on X about lbt, fab, taiwan, shell the most. They currently have XXX followers and XXX posts still getting attention that total XXXXX engagements in the last XX hours. ### Engagements: XXXXX [#](/creator/twitter::3674371694/interactions)  - X Week XXXXXX +8,444% - X Month XXXXXX +178% - X Months XXXXXXX +408% ### Mentions: XX [#](/creator/twitter::3674371694/posts_active)  ### Followers: XXX [#](/creator/twitter::3674371694/followers)  - X Week XXX +12% - X Month XXX +24% - X Months XXX +252% ### CreatorRank: XXXXXXX [#](/creator/twitter::3674371694/influencer_rank)  ### Social Influence [#](/creator/twitter::3674371694/influence) --- **Social category influence** [stocks](/list/stocks) [technology brands](/list/technology-brands) [countries](/list/countries) [finance](/list/finance) **Social topic influence** [lbt](/topic/lbt), [fab](/topic/fab), [taiwan](/topic/taiwan), [shell](/topic/shell), [$18b](/topic/$18b) #3, [asp](/topic/asp), [tan](/topic/tan), [happened](/topic/happened), [if you](/topic/if-you), [malaysia](/topic/malaysia) **Top assets mentioned** [Intel Corporation (INTC)](/topic/$intc) ### Top Social Posts [#](/creator/twitter::3674371694/posts) --- Top posts by engagements in the last XX hours "@christophauto @Silicon_Fly Intel Foundry is a startup no matter how you cut it. 18A is their test drive node. No previous experience of being an external foundry & recent debacles of delays & struggles on 10nm. Exactly you need the shell first to ramp which is exactly what Pat did"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948847394356879658) 2025-07-25 20:46:41 UTC XXX followers, XXX engagements "Looks like full effect of LBT's headcount reduction has *not hit the P&L yet for Intel"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948475357578805348) 2025-07-24 20:08:20 UTC XXX followers, 1718 engagements "@cristian_m32713 @XYang2023 @johnrheritage @Srasgon Those fabs are already being built out at Ohio. I estimate by 2028 atleast X fab at Ohio will be completed ready for ramping 14A. Intel needs to equip it validate and ramp the node to HVM. LBT is saying he will not do that if 14A does not have external customers"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1949091743938343096) 2025-07-26 12:57:38 UTC XXX followers, XXX engagements "Intel shipped XX% SPR XX% EMR & X% GNR (about 109k units) in Q1'25. No information about SRF. AMD OTOH shipped XX% Genoa XX% Bergamo & XX% Turin (about 232k units). I always noticed GNR's slow ramp may be Intel X is not as mature as N4 in yield cycle for large dies 🤷♂(16/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442338268758347) 2025-07-24 17:57:08 UTC XXX followers, XXX engagements "And thanks to Pat Gelsinger's 5N4Y (18A) & competitive product roadmap (CWF & DMR) Intel is at a better position than before like SPR vs Genoa era. GNR has closed some gap on perf/watt & core count hopefully DMR closes the gap further or even takes leadership. We'll see🤞(14/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442334477123634) 2025-07-24 17:57:07 UTC XXX followers, XXX engagements "@christophauto @Silicon_Fly Rest of the Mods on Oregon were all for R&D required for 5N4Y initiative no production fabs there. If I am not mistaken they worked on some of the nodes simultaneously to derisk the tech (GAA & BSPDN). Which also meant more personnel to support that"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948875967759274351) 2025-07-25 22:40:13 UTC XXX followers, XX engagements "@XYang2023 @Silicon_Fly @JackieC42653470 But I agree with you on one thingturning to Private Equity was a bad mistake on PG's part & I think it was DZ's idea (CFO) of Smart Capital. They are sharks. I even think this Private equity is probably why PG got "retired". They probably pressured BoD getting impatient on ROI"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1949127170242035825) 2025-07-26 15:18:25 UTC XXX followers, XX engagements "Things to note - AMD CPUs commands better ASP (Average Selling Price) than Intel because they had higher core count products. We need to normalize this per core to actually get better picture of ASP imo. Plus AMD's CPUs are better than Intel's (ICLSPREMR vs MilanGenoa).(3/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442303062049076) 2025-07-24 17:57:00 UTC XXX followers, XXX engagements "@XYang2023 @johnrheritage It's pure fantasy that Intel could run profitably with most of the products outsourced to external & keep a portion of Intel Foundry underutilized. Exactly why Intel P&L is in the toilet now. @Srasgon asked this question to which LBT mumbled some generic answers to"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948708226901402056) 2025-07-25 11:33:41 UTC XXX followers, XXX engagements "A classless act by Lip Bu Tan imo. He was on the Intel BoD for X years & only quit when the expansion plans were about be pared down. Maybe "build it they will come" may have been excessive but "I will only build if you commit" is not a sound strategy for a new entrant as well"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948521923496849452) 2025-07-24 23:13:23 UTC XXX followers, 8497 engagements "@XYang2023 @Silicon_Fly @JackieC42653470 You do understand that construction in progress digestion is why they can cut the Capex to $18B this year and "higher than $9B to less than $18B" right Sure PG era front loaded it but it is helping Intel one way or the other"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1949101766663893327) 2025-07-26 13:37:28 UTC XXX followers, XXX engagements "@johnrheritage @XYang2023 @Srasgon It's not only the R&D cost it's also the fixed cost of fabs + depreciation cost of non productive assets. Fabs need to be utilised at very high % or they will have negative gross margins (like Intel has now due with XX% wafers outsourced)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948710643067912530) 2025-07-25 11:43:17 UTC XXX followers, XX engagements "Okay there was an impact but through a $1.9B restructuring charges due to XX% planned head count reduction"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948477884227555819) 2025-07-24 20:18:23 UTC XXX followers, XXX engagements "@realmemes6 Compare that to what Pat said here. Really disappointed with LBT on this. He was on the Intel BoD for X years while all these investments were made & was quiet. He quit only right before they were paring investments back"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948576001400672271) 2025-07-25 02:48:16 UTC XXX followers, XXX engagements "@Mojo_flyin @temann90 Lip Bu can easily make these tall claims because Pat got the 18A tech for Intel to comfortably have X gen of products in future & have completed fab shells needed to expand when customers do show up"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948826167483449698) 2025-07-25 19:22:20 UTC XXX followers, XXX engagements "Now back to Mercury research data there were some discussions about GNR ramp (@Mojo_flyin @FosterTraveler) it looks like Server CPUs are very slow to ramp & it takes multiple Qtrs to ramp up to full volume. Intel's bias towards Enterprise customers means it's even slower (15/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442336205119622) 2025-07-24 17:57:07 UTC XXX followers, XXX engagements "@XYang2023 @Silicon_Fly @JackieC42653470 I also saw your Intel Q2'25 "deleverage" post it's wrong You need to add short term debt & long term debt to arrive at Total debt all that happened in Q2'25 is some portion of long term debt is moved to short term debt as it will be due in a year"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1949124621682290830) 2025-07-26 15:08:17 UTC XXX followers, XXX engagements "@christophauto @Silicon_Fly Lets not forget the CWF & DMR Xeon products on 18A as well. I expect Xeon 7P dies to be larger than desktop ones. Xeon 7E could be smaller as they are E core based. Also these processors needs large base tile on Intel X too with cache"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948880179499335731) 2025-07-25 22:56:57 UTC XXX followers, XXX engagements "@diamondrapids @Silicon_Fly @JackieC42653470 I don't understand what mess you are referring to Intel's internal culture Care to elaborate. I have shown data that Intel's financial troubles are due to market downturn in PC & emergence of AI in DC. I think the internal culture stuff is very exaggerated"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1949135651648974936) 2025-07-26 15:52:07 UTC XXX followers, XXX engagements "@XYang2023 I have some data (CPU unit sales data) about how the emergence of AI did a number on Intel especially in DC Server space. I will post it later when I have time. It also solved a mystery (at least to me) of how AMD gained so much market share in DC so quickly"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1947968065775325329) 2025-07-23 10:32:33 UTC XXX followers, 6592 engagements "@XYang2023 @mayorhardin Sure they will have 18A. At the same time AMD might be on a better than 18A node (say N2P ppa 18A) in TSMC USA (because they mostly use n-1 or n-2 nodes). Now Intel is stuck on an inferior node & business disruption due to relying on Taiwan too"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948706310498697235) 2025-07-25 11:26:04 UTC XXX followers, XXX engagements "@johnrheritage @XYang2023 @Srasgon It works as a stop gap solution but not a sustainable business solution. Intel needs to either become a fabless or commit fully to foundry initiative. Middle ground does not work. If Pat hadn't taken over Intel is likely a fabless company now"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948709952689987611) 2025-07-25 11:40:32 UTC XXX followers, XX engagements "Oh uh This is not good for Intel USA or Semi Industry"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948487151454183726) 2025-07-24 20:55:12 UTC XXX followers, 1471 engagements "This is not just a Post COVID demand drop. We can see that Intel went from averaging 5+Mu server CPU shipment per Qtr to around 3+Mu post AI/cloud server useful life change at cloud vendor. Pic quality is not great but that's all I could find (8/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442318916493816) 2025-07-24 17:57:03 UTC XXX followers, XXX engagements "@Lukateake @BenBajarin They already announced the slow down of Intel Ohio internally a while ago(Naga's memo iirc in oregonian). But construction progressed at steady pace as you tracked. I believe the shell will be completed. Once the foundry takes off it is going to be very important for Intel IMHO"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948481061232468257) 2025-07-24 20:31:00 UTC XXX followers, XXX engagements "Kudos to AMD they're gaining market share but there r nuances in the data that investors & media misses (& exaggerates) imo. One can see covid time server life extension by cloud vendors & rise of AI hasn't been kind to Intel. Intel's unit shipments have been cut in half(2/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442299983430006) 2025-07-24 17:56:59 UTC XXX followers, XXX engagements "@christophauto @Silicon_Fly Fab XX can be ramped up when 18A external customers show up. These are rough ball bark calculations. The older fabs like Intel X & Intel X would have fab the supporting tiles for these products"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948878579628114208) 2025-07-25 22:50:36 UTC XXX followers, XXX engagements "AMD is taking market share (no question about it) but their unit growth has slowed/stalled possibly due to Intel reducing price and fighting for its market position. It could be also due to higher core count Bergamo & Turin/Turin dense (more '25 story imo) contributions. (6/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442312155025801) 2025-07-24 17:57:02 UTC XXX followers, XXX engagements "This got more relevant today than I thought. I was wondering why Intel is losing market share a lot in DC even after launch of GNR (better than SPR vs Genoa) & started looking into it. This info is based on Mercury research but sourced from a BofA report & other sources. NFA(1/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442296556683744) 2025-07-24 17:56:58 UTC XXX followers, 6771 engagements "@XYang2023 PG did what he could with what Intel had at that time (AI PCs & Gaudi). Tech does not grow in trees. Like I said previously all praise to LBT if he manages to bring mid single digit $B in AI cloud accelerator revenue. Only when AI moves to edge/enterprise they make AI money imo"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948545309778608299) 2025-07-25 00:46:18 UTC XXX followers, XXX engagements "@XYang2023 @johnrheritage @Srasgon Does GFS outsource it's business to TSMC and keep its fabs idle"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948708994685469120) 2025-07-25 11:36:44 UTC XXX followers, XXX engagements "@XYang2023 @cou_end @icnewer Considering Xeons are the choice of CPU for AI servers (like Nvidia DGX) I would presume intel's market share is very high in AI head nodes"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948609567765791102) 2025-07-25 05:01:39 UTC XXX followers, XX engagements "How many Low-NA EUV & High-NA EUV machines Intel owns based on BofA estimates on ASML notes (July 1st 2025). So roughly XX Low-NA EUV & X High-NA EUV machines will be owned by Intel by end of this year. Look at the Low-NA EUV numbers for TSMC🤯 & Samsung owns 2x Intel too"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1947827650841530439) 2025-07-23 01:14:35 UTC XXX followers, 5715 engagements "@XYang2023 @cou_end @icnewer No idea what that is supposed to mean Only XX% market share on AI head nodes"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948609039090233535) 2025-07-25 04:59:32 UTC XXX followers, XXX engagements "@mayorhardin It is Intel's problem LBT idea is to stop developing 14A if there is no external customer & outsource leading edge prodicts to TSMC. So Intel's business depends on TSMC in Taiwan. Intel needs to develop 14A for its own use not others"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948700928162275633) 2025-07-25 11:04:41 UTC XXX followers, XXX engagements "@XYang2023 @cou_end @icnewer ARM is X% of the total shipment volume of total servers. It will be even lower if you consider the revenue point of view. Like Insaid he is talking about cloud/hyperscale only where AMD does represent near XX% market share (including Lisa Su's commentary)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948697281936720373) 2025-07-25 10:50:11 UTC XXX followers, XX engagements "When the DC CPU market comes back (CPUs are still needed for AI servers & not everything needs GPUs cloud will need CPUs again) I am guessing Intel will benefit more on the way up like the way down although AMD has gained a foothold now. That'll have an impact on future.(13/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442332870947259) 2025-07-24 17:57:07 UTC XXX followers, XXX engagements "@Silicon_Fly Thank you That's a strategy an established industry player like TSMC should follow not a startup like Intel Foundry"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948840952917098567) 2025-07-25 20:21:05 UTC XXX followers, XXX engagements "Now let's talk about market share % that media and investors like to talk about. Like how the ARM server market share % is skewed by total shipment being down significantly AMD market share also is skewed due to same reason. It's good for media headlines but lacks nuance. (12/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442329460998172) 2025-07-24 17:57:06 UTC XXX followers, XXX engagements "ASP growth & market share gain has propelled AMD's DC CPU revenue higher while Intel lost about $10B from peak most of it due to AI & Covid era server useful life extensions by major cloud players. Let's not forget the post covid PC market slump adding to their misery (7/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442315150033259) 2025-07-24 17:57:02 UTC XXX followers, XXX engagements "Or is he talking about lack of SMT on Lion Cove in Arrow Lake/ Lunar Lake. But those are client products"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948485820425318617) 2025-07-24 20:49:55 UTC XXX followers, XXX engagements "Sometime market forces work against you especially in cyclical & disruption prone industries. If Post-COVID PC market was forecasted correctly by Intel's C-suite & AI disruption hadn't happened I presume all the over-hiring & over-building wouldn't have weighed down Intel (11/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442326348824625) 2025-07-24 17:57:05 UTC XXX followers, XXX engagements "I found that "I will not commit any capex for 14A until external customer commit" to be extremely dangerous for Intel. What happens if Taiwan gets invaded with only the US TSM fabs stuck in n-1 node (but better than 18A) survives. Will Intel get fab capacity at TSM"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948530777399943471) 2025-07-24 23:48:33 UTC XXX followers, 1827 engagements "@IanCutress Considering there is X EUV count against Rapidus in 2025 I guess this "estimate" could be revenue recognition on ASML's P&L rather than deliveries. If that is the case the total number of EUV owned by any of these players could be considerably higher. Thanks for your insights"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948010210947465714) 2025-07-23 13:20:01 UTC XXX followers, XXX engagements "@XYang2023 @johnrheritage @Srasgon Because to replace the wafer volume of Intel products with external customer is no easy task. What DZ said is not a guarantee. If they are that confident about future customers for 18A then they should develop 14A foundry too because that will follow the same trajectory as 18A"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948712451723165997) 2025-07-25 11:50:28 UTC XXX followers, XXX engagements "@Mojo_flyin @temann90 If Pat started with I will only build Fab XX if 18A development proves good yield (even now we are debating 18A yields) customer commits in 2025 X years to construction of fab X yr to equip it validate & ramp it to HVM. The first product comes out in 2029+"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948827015982858375) 2025-07-25 19:25:42 UTC XXX followers, XX engagements "@christophauto @Silicon_Fly Pat was the one put the projects in Germany Israel & Malaysia on hold. So what you are saying is not definitely not true. He built X fab in Ireland for Intel 3/4 X fabs in Arizona for 18A. Israel Fab XX you can say was the one is the overdone. Ohio one was meant to be for 14A"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948875387217322494) 2025-07-25 22:37:55 UTC XXX followers, XXX engagements "So Intel lost more than what AMD gained. There is also ARM servers but those are still very less in number per Mercury & does not explain 12Mu unit difference from '22-'23 for Intel. imo ARM as a visible % is only showing up because total units of x86 CPU are down in '23.(4/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442306845311360) 2025-07-24 17:57:00 UTC XXX followers, XXX engagements "@christophauto @Silicon_Fly I am basing my guess on the TSM USA fabs are 20k wspm capacity per some analyst notes I have seen. Now add another XX Mu of desktop processors which will be slightly bigger die. They can easily fill Fab XX & XX with 18A in next X years"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948878240199967019) 2025-07-25 22:49:15 UTC XXX followers, XX engagements "So demand destruction for DC CPUs due to rise of AI in Q4'22 is the real culprit for Intel's predicament in DC. Intel's server CPU shipment volume dropped nearly XX% in '23 vs AMD's volume drop of XX% in '23. AMD shipment grew @15% in '24 but Intel lost -X% volume in '24. (5/n)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948442310607601711) 2025-07-24 17:57:01 UTC XXX followers, XXX engagements "@XYang2023 @johnrheritage @Srasgon The idea of Foundry breaking even is relying on Intel's outsourced products coming back from TSMC to Intel. If you reverse that the foundry will turn to loss making business again and will drag P&L for the business. They will have significantly shed assets to achieve this"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948708746638594557) 2025-07-25 11:35:45 UTC XXX followers, XX engagements "@christophauto @Silicon_Fly Now Intel ships 150Mu units of notebooks in a given year. The faster they switch their entire product to 18A the better for them in cost. So that translates to 31250 wspm. Fab XX & XX are 20-25k wspm capacity (my guess)"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948877593542906012) 2025-07-25 22:46:41 UTC XXX followers, XXX engagements "Intel 18A is critical node for Intel products for the next X generation of products in client and server market - LBT. Reuters authors 🤡"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948491282818478588) 2025-07-24 21:11:37 UTC XXX followers, 5531 engagements "It is also fully abandoning the Germany and Poland expansion plans (was put on hold last year). Also winddown down the Assembly & Test plant in Costa Rica (consolidate into Vietnam & Malaysia sites). Intel Ohio construction is officially being slowed down to match market demand"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948478588166033723) 2025-07-24 20:21:11 UTC XXX followers, XXX engagements "@johnrheritage @XYang2023 @Srasgon How is that different from today Intel foundry couldn't handle XX% wafer outsourcing with very mature depreciated fabs I can imagine it will be a way worse scenario of outsourcing XX% wafers & depreciating costlier leading edge fabs at the same time"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948739184740438360) 2025-07-25 13:36:42 UTC XXX followers, XXX engagements "@christophauto @Silicon_Fly Now lets assume Intel ships 40-60 million units of PTL in 2026. Considering what we know about PTL compute die 114sq-mm lets say XXX good dies per wafer. That leads to 100k to 150k total wafers in 2026 for PTL alone. That translates to 8333 to 12500 wspm"  [@D_K_Rajasekar](/creator/x/D_K_Rajasekar) on [X](/post/tweet/1948876892091601030) 2025-07-25 22:43:54 UTC XXX followers, XX engagements
[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]
@D_K_Rajasekar DKR
DKR posts on X about lbt, fab, taiwan, shell the most. They currently have XXX followers and XXX posts still getting attention that total XXXXX engagements in the last XX hours.
Engagements: XXXXX #
- X Week XXXXXX +8,444%
- X Month XXXXXX +178%
- X Months XXXXXXX +408%
Mentions: XX #
Followers: XXX #
- X Week XXX +12%
- X Month XXX +24%
- X Months XXX +252%
CreatorRank: XXXXXXX #
Social Influence #
Social category influence stocks technology brands countries finance
Social topic influence lbt, fab, taiwan, shell, $18b #3, asp, tan, happened, if you, malaysia
Top assets mentioned Intel Corporation (INTC)
Top Social Posts #
Top posts by engagements in the last XX hours
"@christophauto @Silicon_Fly Intel Foundry is a startup no matter how you cut it. 18A is their test drive node. No previous experience of being an external foundry & recent debacles of delays & struggles on 10nm. Exactly you need the shell first to ramp which is exactly what Pat did"
@D_K_Rajasekar on X 2025-07-25 20:46:41 UTC XXX followers, XXX engagements
"Looks like full effect of LBT's headcount reduction has *not hit the P&L yet for Intel" @D_K_Rajasekar on X 2025-07-24 20:08:20 UTC XXX followers, 1718 engagements
"@cristian_m32713 @XYang2023 @johnrheritage @Srasgon Those fabs are already being built out at Ohio. I estimate by 2028 atleast X fab at Ohio will be completed ready for ramping 14A. Intel needs to equip it validate and ramp the node to HVM. LBT is saying he will not do that if 14A does not have external customers" @D_K_Rajasekar on X 2025-07-26 12:57:38 UTC XXX followers, XXX engagements
"Intel shipped XX% SPR XX% EMR & X% GNR (about 109k units) in Q1'25. No information about SRF. AMD OTOH shipped XX% Genoa XX% Bergamo & XX% Turin (about 232k units). I always noticed GNR's slow ramp may be Intel X is not as mature as N4 in yield cycle for large dies 🤷♂(16/n)" @D_K_Rajasekar on X 2025-07-24 17:57:08 UTC XXX followers, XXX engagements
"And thanks to Pat Gelsinger's 5N4Y (18A) & competitive product roadmap (CWF & DMR) Intel is at a better position than before like SPR vs Genoa era. GNR has closed some gap on perf/watt & core count hopefully DMR closes the gap further or even takes leadership. We'll see🤞(14/n)" @D_K_Rajasekar on X 2025-07-24 17:57:07 UTC XXX followers, XXX engagements
"@christophauto @Silicon_Fly Rest of the Mods on Oregon were all for R&D required for 5N4Y initiative no production fabs there. If I am not mistaken they worked on some of the nodes simultaneously to derisk the tech (GAA & BSPDN). Which also meant more personnel to support that" @D_K_Rajasekar on X 2025-07-25 22:40:13 UTC XXX followers, XX engagements
"@XYang2023 @Silicon_Fly @JackieC42653470 But I agree with you on one thingturning to Private Equity was a bad mistake on PG's part & I think it was DZ's idea (CFO) of Smart Capital. They are sharks. I even think this Private equity is probably why PG got "retired". They probably pressured BoD getting impatient on ROI" @D_K_Rajasekar on X 2025-07-26 15:18:25 UTC XXX followers, XX engagements
"Things to note - AMD CPUs commands better ASP (Average Selling Price) than Intel because they had higher core count products. We need to normalize this per core to actually get better picture of ASP imo. Plus AMD's CPUs are better than Intel's (ICLSPREMR vs MilanGenoa).(3/n)" @D_K_Rajasekar on X 2025-07-24 17:57:00 UTC XXX followers, XXX engagements
"@XYang2023 @johnrheritage It's pure fantasy that Intel could run profitably with most of the products outsourced to external & keep a portion of Intel Foundry underutilized. Exactly why Intel P&L is in the toilet now. @Srasgon asked this question to which LBT mumbled some generic answers to" @D_K_Rajasekar on X 2025-07-25 11:33:41 UTC XXX followers, XXX engagements
"A classless act by Lip Bu Tan imo. He was on the Intel BoD for X years & only quit when the expansion plans were about be pared down. Maybe "build it they will come" may have been excessive but "I will only build if you commit" is not a sound strategy for a new entrant as well" @D_K_Rajasekar on X 2025-07-24 23:13:23 UTC XXX followers, 8497 engagements
"@XYang2023 @Silicon_Fly @JackieC42653470 You do understand that construction in progress digestion is why they can cut the Capex to $18B this year and "higher than $9B to less than $18B" right Sure PG era front loaded it but it is helping Intel one way or the other" @D_K_Rajasekar on X 2025-07-26 13:37:28 UTC XXX followers, XXX engagements
"@johnrheritage @XYang2023 @Srasgon It's not only the R&D cost it's also the fixed cost of fabs + depreciation cost of non productive assets. Fabs need to be utilised at very high % or they will have negative gross margins (like Intel has now due with XX% wafers outsourced)" @D_K_Rajasekar on X 2025-07-25 11:43:17 UTC XXX followers, XX engagements
"Okay there was an impact but through a $1.9B restructuring charges due to XX% planned head count reduction" @D_K_Rajasekar on X 2025-07-24 20:18:23 UTC XXX followers, XXX engagements
"@realmemes6 Compare that to what Pat said here. Really disappointed with LBT on this. He was on the Intel BoD for X years while all these investments were made & was quiet. He quit only right before they were paring investments back" @D_K_Rajasekar on X 2025-07-25 02:48:16 UTC XXX followers, XXX engagements
"@Mojo_flyin @temann90 Lip Bu can easily make these tall claims because Pat got the 18A tech for Intel to comfortably have X gen of products in future & have completed fab shells needed to expand when customers do show up" @D_K_Rajasekar on X 2025-07-25 19:22:20 UTC XXX followers, XXX engagements
"Now back to Mercury research data there were some discussions about GNR ramp (@Mojo_flyin @FosterTraveler) it looks like Server CPUs are very slow to ramp & it takes multiple Qtrs to ramp up to full volume. Intel's bias towards Enterprise customers means it's even slower (15/n)" @D_K_Rajasekar on X 2025-07-24 17:57:07 UTC XXX followers, XXX engagements
"@XYang2023 @Silicon_Fly @JackieC42653470 I also saw your Intel Q2'25 "deleverage" post it's wrong You need to add short term debt & long term debt to arrive at Total debt all that happened in Q2'25 is some portion of long term debt is moved to short term debt as it will be due in a year" @D_K_Rajasekar on X 2025-07-26 15:08:17 UTC XXX followers, XXX engagements
"@christophauto @Silicon_Fly Lets not forget the CWF & DMR Xeon products on 18A as well. I expect Xeon 7P dies to be larger than desktop ones. Xeon 7E could be smaller as they are E core based. Also these processors needs large base tile on Intel X too with cache" @D_K_Rajasekar on X 2025-07-25 22:56:57 UTC XXX followers, XXX engagements
"@diamondrapids @Silicon_Fly @JackieC42653470 I don't understand what mess you are referring to Intel's internal culture Care to elaborate. I have shown data that Intel's financial troubles are due to market downturn in PC & emergence of AI in DC. I think the internal culture stuff is very exaggerated" @D_K_Rajasekar on X 2025-07-26 15:52:07 UTC XXX followers, XXX engagements
"@XYang2023 I have some data (CPU unit sales data) about how the emergence of AI did a number on Intel especially in DC Server space. I will post it later when I have time. It also solved a mystery (at least to me) of how AMD gained so much market share in DC so quickly" @D_K_Rajasekar on X 2025-07-23 10:32:33 UTC XXX followers, 6592 engagements
"@XYang2023 @mayorhardin Sure they will have 18A. At the same time AMD might be on a better than 18A node (say N2P ppa 18A) in TSMC USA (because they mostly use n-1 or n-2 nodes). Now Intel is stuck on an inferior node & business disruption due to relying on Taiwan too" @D_K_Rajasekar on X 2025-07-25 11:26:04 UTC XXX followers, XXX engagements
"@johnrheritage @XYang2023 @Srasgon It works as a stop gap solution but not a sustainable business solution. Intel needs to either become a fabless or commit fully to foundry initiative. Middle ground does not work. If Pat hadn't taken over Intel is likely a fabless company now" @D_K_Rajasekar on X 2025-07-25 11:40:32 UTC XXX followers, XX engagements
"Oh uh This is not good for Intel USA or Semi Industry" @D_K_Rajasekar on X 2025-07-24 20:55:12 UTC XXX followers, 1471 engagements
"This is not just a Post COVID demand drop. We can see that Intel went from averaging 5+Mu server CPU shipment per Qtr to around 3+Mu post AI/cloud server useful life change at cloud vendor. Pic quality is not great but that's all I could find (8/n)" @D_K_Rajasekar on X 2025-07-24 17:57:03 UTC XXX followers, XXX engagements
"@Lukateake @BenBajarin They already announced the slow down of Intel Ohio internally a while ago(Naga's memo iirc in oregonian). But construction progressed at steady pace as you tracked. I believe the shell will be completed. Once the foundry takes off it is going to be very important for Intel IMHO" @D_K_Rajasekar on X 2025-07-24 20:31:00 UTC XXX followers, XXX engagements
"Kudos to AMD they're gaining market share but there r nuances in the data that investors & media misses (& exaggerates) imo. One can see covid time server life extension by cloud vendors & rise of AI hasn't been kind to Intel. Intel's unit shipments have been cut in half(2/n)" @D_K_Rajasekar on X 2025-07-24 17:56:59 UTC XXX followers, XXX engagements
"@christophauto @Silicon_Fly Fab XX can be ramped up when 18A external customers show up. These are rough ball bark calculations. The older fabs like Intel X & Intel X would have fab the supporting tiles for these products" @D_K_Rajasekar on X 2025-07-25 22:50:36 UTC XXX followers, XXX engagements
"AMD is taking market share (no question about it) but their unit growth has slowed/stalled possibly due to Intel reducing price and fighting for its market position. It could be also due to higher core count Bergamo & Turin/Turin dense (more '25 story imo) contributions. (6/n)" @D_K_Rajasekar on X 2025-07-24 17:57:02 UTC XXX followers, XXX engagements
"This got more relevant today than I thought. I was wondering why Intel is losing market share a lot in DC even after launch of GNR (better than SPR vs Genoa) & started looking into it. This info is based on Mercury research but sourced from a BofA report & other sources. NFA(1/n)" @D_K_Rajasekar on X 2025-07-24 17:56:58 UTC XXX followers, 6771 engagements
"@XYang2023 PG did what he could with what Intel had at that time (AI PCs & Gaudi). Tech does not grow in trees. Like I said previously all praise to LBT if he manages to bring mid single digit $B in AI cloud accelerator revenue. Only when AI moves to edge/enterprise they make AI money imo" @D_K_Rajasekar on X 2025-07-25 00:46:18 UTC XXX followers, XXX engagements
"@XYang2023 @johnrheritage @Srasgon Does GFS outsource it's business to TSMC and keep its fabs idle" @D_K_Rajasekar on X 2025-07-25 11:36:44 UTC XXX followers, XXX engagements
"@XYang2023 @cou_end @icnewer Considering Xeons are the choice of CPU for AI servers (like Nvidia DGX) I would presume intel's market share is very high in AI head nodes" @D_K_Rajasekar on X 2025-07-25 05:01:39 UTC XXX followers, XX engagements
"How many Low-NA EUV & High-NA EUV machines Intel owns based on BofA estimates on ASML notes (July 1st 2025). So roughly XX Low-NA EUV & X High-NA EUV machines will be owned by Intel by end of this year. Look at the Low-NA EUV numbers for TSMC🤯 & Samsung owns 2x Intel too" @D_K_Rajasekar on X 2025-07-23 01:14:35 UTC XXX followers, 5715 engagements
"@XYang2023 @cou_end @icnewer No idea what that is supposed to mean Only XX% market share on AI head nodes" @D_K_Rajasekar on X 2025-07-25 04:59:32 UTC XXX followers, XXX engagements
"@mayorhardin It is Intel's problem LBT idea is to stop developing 14A if there is no external customer & outsource leading edge prodicts to TSMC. So Intel's business depends on TSMC in Taiwan. Intel needs to develop 14A for its own use not others" @D_K_Rajasekar on X 2025-07-25 11:04:41 UTC XXX followers, XXX engagements
"@XYang2023 @cou_end @icnewer ARM is X% of the total shipment volume of total servers. It will be even lower if you consider the revenue point of view. Like Insaid he is talking about cloud/hyperscale only where AMD does represent near XX% market share (including Lisa Su's commentary)" @D_K_Rajasekar on X 2025-07-25 10:50:11 UTC XXX followers, XX engagements
"When the DC CPU market comes back (CPUs are still needed for AI servers & not everything needs GPUs cloud will need CPUs again) I am guessing Intel will benefit more on the way up like the way down although AMD has gained a foothold now. That'll have an impact on future.(13/n)" @D_K_Rajasekar on X 2025-07-24 17:57:07 UTC XXX followers, XXX engagements
"@Silicon_Fly Thank you That's a strategy an established industry player like TSMC should follow not a startup like Intel Foundry" @D_K_Rajasekar on X 2025-07-25 20:21:05 UTC XXX followers, XXX engagements
"Now let's talk about market share % that media and investors like to talk about. Like how the ARM server market share % is skewed by total shipment being down significantly AMD market share also is skewed due to same reason. It's good for media headlines but lacks nuance. (12/n)" @D_K_Rajasekar on X 2025-07-24 17:57:06 UTC XXX followers, XXX engagements
"ASP growth & market share gain has propelled AMD's DC CPU revenue higher while Intel lost about $10B from peak most of it due to AI & Covid era server useful life extensions by major cloud players. Let's not forget the post covid PC market slump adding to their misery (7/n)" @D_K_Rajasekar on X 2025-07-24 17:57:02 UTC XXX followers, XXX engagements
"Or is he talking about lack of SMT on Lion Cove in Arrow Lake/ Lunar Lake. But those are client products" @D_K_Rajasekar on X 2025-07-24 20:49:55 UTC XXX followers, XXX engagements
"Sometime market forces work against you especially in cyclical & disruption prone industries. If Post-COVID PC market was forecasted correctly by Intel's C-suite & AI disruption hadn't happened I presume all the over-hiring & over-building wouldn't have weighed down Intel (11/n)" @D_K_Rajasekar on X 2025-07-24 17:57:05 UTC XXX followers, XXX engagements
"I found that "I will not commit any capex for 14A until external customer commit" to be extremely dangerous for Intel. What happens if Taiwan gets invaded with only the US TSM fabs stuck in n-1 node (but better than 18A) survives. Will Intel get fab capacity at TSM" @D_K_Rajasekar on X 2025-07-24 23:48:33 UTC XXX followers, 1827 engagements
"@IanCutress Considering there is X EUV count against Rapidus in 2025 I guess this "estimate" could be revenue recognition on ASML's P&L rather than deliveries. If that is the case the total number of EUV owned by any of these players could be considerably higher. Thanks for your insights" @D_K_Rajasekar on X 2025-07-23 13:20:01 UTC XXX followers, XXX engagements
"@XYang2023 @johnrheritage @Srasgon Because to replace the wafer volume of Intel products with external customer is no easy task. What DZ said is not a guarantee. If they are that confident about future customers for 18A then they should develop 14A foundry too because that will follow the same trajectory as 18A" @D_K_Rajasekar on X 2025-07-25 11:50:28 UTC XXX followers, XXX engagements
"@Mojo_flyin @temann90 If Pat started with I will only build Fab XX if 18A development proves good yield (even now we are debating 18A yields) customer commits in 2025 X years to construction of fab X yr to equip it validate & ramp it to HVM. The first product comes out in 2029+" @D_K_Rajasekar on X 2025-07-25 19:25:42 UTC XXX followers, XX engagements
"@christophauto @Silicon_Fly Pat was the one put the projects in Germany Israel & Malaysia on hold. So what you are saying is not definitely not true. He built X fab in Ireland for Intel 3/4 X fabs in Arizona for 18A. Israel Fab XX you can say was the one is the overdone. Ohio one was meant to be for 14A" @D_K_Rajasekar on X 2025-07-25 22:37:55 UTC XXX followers, XXX engagements
"So Intel lost more than what AMD gained. There is also ARM servers but those are still very less in number per Mercury & does not explain 12Mu unit difference from '22-'23 for Intel. imo ARM as a visible % is only showing up because total units of x86 CPU are down in '23.(4/n)" @D_K_Rajasekar on X 2025-07-24 17:57:00 UTC XXX followers, XXX engagements
"@christophauto @Silicon_Fly I am basing my guess on the TSM USA fabs are 20k wspm capacity per some analyst notes I have seen. Now add another XX Mu of desktop processors which will be slightly bigger die. They can easily fill Fab XX & XX with 18A in next X years" @D_K_Rajasekar on X 2025-07-25 22:49:15 UTC XXX followers, XX engagements
"So demand destruction for DC CPUs due to rise of AI in Q4'22 is the real culprit for Intel's predicament in DC. Intel's server CPU shipment volume dropped nearly XX% in '23 vs AMD's volume drop of XX% in '23. AMD shipment grew @15% in '24 but Intel lost -X% volume in '24. (5/n)" @D_K_Rajasekar on X 2025-07-24 17:57:01 UTC XXX followers, XXX engagements
"@XYang2023 @johnrheritage @Srasgon The idea of Foundry breaking even is relying on Intel's outsourced products coming back from TSMC to Intel. If you reverse that the foundry will turn to loss making business again and will drag P&L for the business. They will have significantly shed assets to achieve this" @D_K_Rajasekar on X 2025-07-25 11:35:45 UTC XXX followers, XX engagements
"@christophauto @Silicon_Fly Now Intel ships 150Mu units of notebooks in a given year. The faster they switch their entire product to 18A the better for them in cost. So that translates to 31250 wspm. Fab XX & XX are 20-25k wspm capacity (my guess)" @D_K_Rajasekar on X 2025-07-25 22:46:41 UTC XXX followers, XXX engagements
"Intel 18A is critical node for Intel products for the next X generation of products in client and server market - LBT. Reuters authors 🤡" @D_K_Rajasekar on X 2025-07-24 21:11:37 UTC XXX followers, 5531 engagements
"It is also fully abandoning the Germany and Poland expansion plans (was put on hold last year). Also winddown down the Assembly & Test plant in Costa Rica (consolidate into Vietnam & Malaysia sites). Intel Ohio construction is officially being slowed down to match market demand" @D_K_Rajasekar on X 2025-07-24 20:21:11 UTC XXX followers, XXX engagements
"@johnrheritage @XYang2023 @Srasgon How is that different from today Intel foundry couldn't handle XX% wafer outsourcing with very mature depreciated fabs I can imagine it will be a way worse scenario of outsourcing XX% wafers & depreciating costlier leading edge fabs at the same time" @D_K_Rajasekar on X 2025-07-25 13:36:42 UTC XXX followers, XXX engagements
"@christophauto @Silicon_Fly Now lets assume Intel ships 40-60 million units of PTL in 2026. Considering what we know about PTL compute die 114sq-mm lets say XXX good dies per wafer. That leads to 100k to 150k total wafers in 2026 for PTL alone. That translates to 8333 to 12500 wspm" @D_K_Rajasekar on X 2025-07-25 22:43:54 UTC XXX followers, XX engagements