SemiAnalysis

Most of AMD and NVIDIA’s best 10x engineers are in Shanghai. AMD’s MoRI collective team, AMD’s disaggregated applications engineering team, and other AMD teams that understand how to do first-principles-based engineering are all mostly based in Shanghai.

The most of AMD & NVIDIA's best 10x AMD engineers are in Shanghai. AMD's MoRI collective team, AMD disagg applications engineering team, & other AMD teams that understand how to do first princples based engineering are all mostly in Shanghai

ONE TRICK SAMSUNG DOESN'T WANT U TO KNOW: With Physical RAM prices increasing rapidly, it is important to take advantage of downloading RAM for free instead of buying physical ram.

This week's latest rick and morty season was DOGSHIT. D O G S H I T. bring back season 1-3 classics, stop with this C R A P

Congratulations to $CBRS on its successful IPO. Cerebras's founders can now reap the rewards of their hard work and contributions to hardware innovation. CEO Andrew Feldman who owns 10.3mn shares is now worth $24bn.* Read our latest article for more on Cerebras's innovations in silicon, servers, and marketing.  *If Andrew moves to Vietnam where cost of living is 8x lower than Sunnyvale California, resulting in an 8x increase in purchasing power of his actual net worth of $3bn.

What makes this interesting from an industrials angle is the supplier list that has emerged behind it: GE Vernova targeting roughly 24 GW per year of heavy-duty gas turbines, Siemens Energy moving from 20 to about 30 GW per year by the end of the decade, Wärtsilä shipping 800 MW into a single project for one operator, Bloom committing to 2 GW per year of fuel cells by the end of 2026, and at least twelve named suppliers now sitting on 400+ MW of AI orders each. (3/4)

The full chat with Mishek Musa on how ADI is shrinking inference down to the edge and setting up physical leaderboards for the robotics community. Chapters: 0:00 — Introduction & ADI's Emerging Tech Hub 0:56 — Inside the Multimodal Tactile Sensor 1:52 — Automating Data Center

AMD ALERT MI355 is now 40% cheaper than B200 on GLM5 architecture for Single Node serving FP8 14 weeks after the initial launch of GLM5 on both non-MTP & MTP with spec decode for SGLang v0.12 for both CUDA & ROCm.  SPEED IS THE MOAT!! Great work to @AnushElangovan ,  @roaner ,