Huawei 950 supernode is described as a GB200/GB300 substitute, but capacity is still tight
teortaxesTex · x · 2026-07-23
A post quoting an industry discussion says DeepSeek-style training has already moved away from NVIDIA’s ecosystem and could be ported to Huawei GPUs next.
Key points
- Huawei’s 950 supernode is described as able to fully substitute for NVIDIA GB200 and GB300 in both performance and price.
- The quoted view claims the main remaining gap is chip-level, not ecosystem-level: Huawei is said to be about 4× behind plus roughly two years.
- Huawei’s current capacity is described as limited: about 16,000 GPUs for some partners, while internet giants may want 100K+.
- The conclusion is that the ecosystem problem is largely solved, but capacity still constrains training ever larger models.
Related event: DeepSeek Founder's Investor Call Reveals AGI Roadmap and Compute Plans(37 posts)→
More from Infra
- Engram's random reads don't suit SSDs; CPU-memory over NVLink could serve all 72 GPUs — bookwormengr · 2026-09-11
- 80% of the DIY LLM inference hype posters have already quit — it's brutally hard systems work — abhijithneil · 2026-09-11
- Hugging Face's Ultra Scale Playbook: a free book on training LLMs on GPU clusters — mdancho84 · 2026-09-11
- Is inference latency becoming the biggest bottleneck for production AI agents? — Euphoric_Sea632 · 2026-09-11
- LLM Serving Metrics Thread: Why TPOT and Uptime Make or Break User Experience — abhijithneil · 2026-09-11
- PlanetScale launches sharded Postgres: 768 servers acting as one, 1PB scale — dhruv2038 · 2026-09-11