DeepSeek Reportedly Adapting to Huawei Chips, Reducing CUDA Dependence with Proprietary Compilers
xiaosun86 · x · 2026-07-24
Reports indicate DeepSeek is working closely with Huawei to secure around 16,000 AI chips. By using its own compiler and TileLang environment, DeepSeek has significantly reduced its software dependence on NVIDIA’s CUDA ecosystem. If ported to Huawei chips, this could resolve Chinese hardware ecosystem bottlenecks within a year. Separately, AMD plans to use Claude Code to accelerate its ROCm ecosystem.
More from Infra
- Vercel Eve × agentOS pitches a WebAssembly runtime with 4.8ms cold starts — cramforce · 2026-07-24
- CoreWeave leads MiniMax M3 serving benchmark with 357 tokens/sec and $0.22/M — wandb · 2026-07-24
- A DIY AI server gets a homemade PWM fan controller — MackThax · 2026-07-24
- Diamandis says U.S. golf courses use 30 times more water than data centers — PeterDiamandis · 2026-07-24
- AMD and Cerebras Partner on Disaggregated AI Inference System — Sethwinterroth · 2026-07-24
- Echo routes open-weight models to match Fable at one-third the inference cost — adam_rida · 2026-07-24