Wiwynn source says AMD MI455 ramp is hard while Nvidia Vera Rubin is smoother
Midnight_Captl · x · 2026-07-23
A post citing an ODM, Wiwynn, claims AMD’s MI455 is facing major ramp challenges because it is AMD’s first rack-scale system and also has Broadcom-based scale-up networking complexity.
The same source says customers are optimistic and would buy it if AMD gets the rack working well, but adds that Nvidia’s Vera Rubin ramp looks smoother because Nvidia has more experience from GB200 and has shifted to a more modular design.
This is a supply-chain-style anecdote rather than an official statement, but it points to the execution gap between the two vendors in rack-scale AI hardware.
Related event: AMD MI455 System Faces Production Scaling Challenges(2 posts)→
More from Infra
- AWS on agent prompt caching: explicit + implicit patterns to slash token costs — AWS ML Blog · 2026-10-27
- Thieves spoofing GPS steal $100M+ in data center cargo since June — Ars Technica AI · 2026-10-27
- Lam Research: weekly token usage up ~250x in 18 months — Beth_Kindig · 2026-10-27
- AWS shows how to build enterprise agentic retrieval on Bedrock with CDK — AWS ML Blog · 2026-10-27
- Encrypted GPT-2 speedrun challenges FHE inference: best is 3.031 s/token, goal sub-1s — tensorqt · 2026-10-27
- Open-source Latch compiles PyTorch into training megakernels, 2.75x faster on H100 — sguada · 2026-10-27