DeepSeek’s Wenfeng says 2026 compute spend may stay under RMB 20B
teortaxesTex · x · 2026-07-23
- The post says DeepSeek founder Wenfeng expects the company will not be able to spend more than 20 billion RMB on compute in 2026.
- It also quotes him saying the current largest model activates about 800B parameters, while a comparable train run would require roughly 50,000 GB300s or 200,000 Huawei 950s.
- The broader point is that DeepSeek sees the main gap with the U.S. as compute resources, not talent, and is willing to spend virtually all available capital on GPUs, even at a premium.
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