The Definitive Edition:
Analysis of Every Single Intel Fab Across the US, Ireland, Israel, and Malaysia
The Lord of the Fabs and the Return of the King
https://t.co/bAgBDcABcv
@chipflare@Intellionaire Hyperscalers doing more COT means more supplier diversity
Especially cuz hyperscalers can get Intel to design and modify IP/libraries for them
Nvidia/Broadcom stick heavy to TSMC. While hyperscaler CEOs don't like 1 relationship, and fear Taiwan risk more
@evrgn11112231@TBU12345678 They're raising 75 billion
Tesla will do a raise sometime this year. Maybe another 50
Then they combine and you have like $150B. Who knows they do raise then too
@Eric26573545571@Midnight_Captl I've said/thought Nvidia is making a mistake by not taking all their capital and becoming a frontier lab (not the half measure currently)
All their customers are competing with them. Why not
There's about a 2 year window before this opportunity is shut down
TechInsights has an amazing breakdown of Intel's 18A technology node
Alot of new materials in metal layers and device, new patterning, etch, deposition
(many are firsts in the industry)
Alot goes over my head, but if you enjoy SEM pictures. It's great!
https://t.co/dIDVjTSG0r
@Midnight_Captl If u look at revenue for labs
Labs will be hundreds of billions to trillions. They'll get leverage in the relationship with Nvidia and fabs
(Not great for Amazon and Microsoft if they don't have a model long run)
@Midnight_Captl You can't be AGI pilled (trillions in revenue)
Then say oh ya AI can't break my software moat and chip design won't be easier
Ultimately it's going to be those with hard assests (fabs) and companies with a lab selling to enterprise customers
Most optimal for lowering cost
@speenhamlander Intel doesn't want Elon to be a foundry (that's Intels business)
Elon has said directly he wants to make chips for himself, this also him to iterate more quickly since being a foundry means you have to do everything vs subset
He's gonna buy same equipment Intel uses
Lmao they pitch TeraFab at 1 million WSPM in the late 2030s period
So if Intels license fee is 15% of let's say a 8A/10A node at 75K per wafer
That's $135B per year in pure profit before any other business income
Let's uhhh not consider that for now
To Boldly Go: The Case for Space Datacenters
Space DC Total Cost of Ownership Explained. Unpacking constraints from Terrestrial DCs
and Chip Production.
Space-Earth Parity in the late 2030s,
Space DCs could start to be viable even sooner.
https://t.co/HYkBaQdh2G
Clearly, Intel 3 is going to be the bottleneck for intel’s future growth: Panther, Nova, Razer Lake need it; Granite, Clearwater, Diamond Rapids need it. Fab 34 alone (plus F24) is not enough to fulfill those explosive server and client demands.
So I believe Fab 38 in Israel will be used for additional intel 3 ramp in the near future. F38.1 space is now ready for tool move-in.