@AIandCloudTech@SemiAnalysis_ True for enterprises with a lot of inertia. Not for hyper scalers who will do it for better TCO. If AMD chips are say $10,000 cheaper than NVDA for same performance for example (or even slightly lower performance) that is a $1B savings on a 100,000 GPU purchase.
@HammerstoneMar3 Whether you use Microfluidics or Cold plates both involve having to pump cold liquid in and hot liquid out. $VRT is in the latter business and so will not be affected
How an Ex-Oilman Became the Face of Elon Musk’s Supercomputers
How former oilman Brent Mayo became the public face of Elon Musk’s race to build AI supercomputers for xAI.
Read more from Theo Wayt 👇
https://t.co/u9VVhNjkYn?