Dark | Light
[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.]

[@elonevilmusk](/creator/twitter/elonevilmusk)
"@PetGelsinger @Mojo_flyin @BenBajarin @intel test wafers which intel does really slow so they are at least 3-4 quarters from volume This is not factoring recent TD layoffs"  
![@elonevilmusk Avatar](https://lunarcrush.com/gi/w:16/cr:twitter::1555338088628400129.png) [@elonevilmusk](/creator/x/elonevilmusk) on [X](/post/tweet/1948166641344561643) 2025-07-23 23:41:37 UTC XXX followers, XX engagements


"@roadrunnergub @SemiAnalysis_ @elonmusk Not too far off imo around XXX GPUS per utility (not IT/net MW) MW so around 454k but also depends on PUE and the efficiency of other overhead Also I think around $4.5/hour for GB200 so closer to 18b for 1000MW but varies greatly depending on pue"  
![@elonevilmusk Avatar](https://lunarcrush.com/gi/w:16/cr:twitter::1555338088628400129.png) [@elonevilmusk](/creator/x/elonevilmusk) on [X](/post/tweet/1948594042046218538) 2025-07-25 03:59:57 UTC XXX followers, XX engagements


"@onodacapital1 $CRWV also guided to 20-23b of capex and around (900MW from 420MW exiting q1) for 2025 back loaded since only around 5b for q1 +q2 I believe 250MW is coming from $CORZ and 100MW from $APLD Quite a lot when considering Semianalysis has Oracle at 27b"  
![@elonevilmusk Avatar](https://lunarcrush.com/gi/w:16/cr:twitter::1555338088628400129.png) [@elonevilmusk](/creator/x/elonevilmusk) on [X](/post/tweet/1949198884687040844) 2025-07-26 20:03:23 UTC XXX followers, XXX engagements


"@BenBajarin @intel Intel manufacturing is not in a good spot right now on the software side TD is actually equal to or ahead of $TSM But on the equipment/operational side it is a disaster with regards to downtime and inefficiency takes 6x longer to run a wafer through vs tsmc on leading edge"  
![@elonevilmusk Avatar](https://lunarcrush.com/gi/w:16/cr:twitter::1555338088628400129.png) [@elonevilmusk](/creator/x/elonevilmusk) on [X](/post/tweet/1948046600900895200) 2025-07-23 15:44:37 UTC XXX followers, XXX engagements


"@onodacapital1 Sorry something got erased I meant guiding to 900MW EOY for MW capacity compared to 420MW exiting Q1 Should be all IT MW since $CORZ PUE is decently high like most Texas sites"  
![@elonevilmusk Avatar](https://lunarcrush.com/gi/w:16/cr:twitter::1555338088628400129.png) [@elonevilmusk](/creator/x/elonevilmusk) on [X](/post/tweet/1949200144739770482) 2025-07-26 20:08:23 UTC XXX followers, XX engagements


"@onodacapital1 @jay_21_ I think they are able to ramp up based on demand since nobody else seems comfortable with crypto sites Signing the APLD deal was a sign of demand imo given its not that great of a site and they got decent terms when everyone else (MSFT Google) backed out"  
![@elonevilmusk Avatar](https://lunarcrush.com/gi/w:16/cr:twitter::1555338088628400129.png) [@elonevilmusk](/creator/x/elonevilmusk) on [X](/post/tweet/1949173985511715094) 2025-07-26 18:24:26 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.]

@elonevilmusk "@PetGelsinger @Mojo_flyin @BenBajarin @intel test wafers which intel does really slow so they are at least 3-4 quarters from volume This is not factoring recent TD layoffs"
@elonevilmusk Avatar @elonevilmusk on X 2025-07-23 23:41:37 UTC XXX followers, XX engagements

"@roadrunnergub @SemiAnalysis_ @elonmusk Not too far off imo around XXX GPUS per utility (not IT/net MW) MW so around 454k but also depends on PUE and the efficiency of other overhead Also I think around $4.5/hour for GB200 so closer to 18b for 1000MW but varies greatly depending on pue"
@elonevilmusk Avatar @elonevilmusk on X 2025-07-25 03:59:57 UTC XXX followers, XX engagements

"@onodacapital1 $CRWV also guided to 20-23b of capex and around (900MW from 420MW exiting q1) for 2025 back loaded since only around 5b for q1 +q2 I believe 250MW is coming from $CORZ and 100MW from $APLD Quite a lot when considering Semianalysis has Oracle at 27b"
@elonevilmusk Avatar @elonevilmusk on X 2025-07-26 20:03:23 UTC XXX followers, XXX engagements

"@BenBajarin @intel Intel manufacturing is not in a good spot right now on the software side TD is actually equal to or ahead of $TSM But on the equipment/operational side it is a disaster with regards to downtime and inefficiency takes 6x longer to run a wafer through vs tsmc on leading edge"
@elonevilmusk Avatar @elonevilmusk on X 2025-07-23 15:44:37 UTC XXX followers, XXX engagements

"@onodacapital1 Sorry something got erased I meant guiding to 900MW EOY for MW capacity compared to 420MW exiting Q1 Should be all IT MW since $CORZ PUE is decently high like most Texas sites"
@elonevilmusk Avatar @elonevilmusk on X 2025-07-26 20:08:23 UTC XXX followers, XX engagements

"@onodacapital1 @jay_21_ I think they are able to ramp up based on demand since nobody else seems comfortable with crypto sites Signing the APLD deal was a sign of demand imo given its not that great of a site and they got decent terms when everyone else (MSFT Google) backed out"
@elonevilmusk Avatar @elonevilmusk on X 2025-07-26 18:24:26 UTC XXX followers, XX engagements

creator/twitter::1555338088628400129/posts
/creator/twitter::1555338088628400129/posts