Claude helps AMD tune MI355X in a weekend as Anthropic plans 2GW of Instinct GPUs
新智元 · wechat · 2026-07-27
Anthropic says a Claude-driven engineering workflow got AMD’s new MI355X rack running over a weekend with no human code changes, and the two companies are now planning up to 2GW of Instinct GPU deployment in AMD Helios systems.
- AMD’s ROCm.AI adds AI-readable ISA docs, AMDSkills, ROCmCLI access, and tooling for Claude, Codex, and Cursor so agents can install, deploy, debug, and tune GPU workloads.
- AMD says its Hyperloom workflow can benchmark, find CPU/GPU bottlenecks, generate custom kernels, and rerun tests; a live demo reportedly improved MiniMax M3 throughput by 38%.
- The broader bet: instead of only competing on raw FLOPS and bandwidth, GPU vendors will compete on whether agents can read, operate, and optimize their software stack.
Related event: AMD Invests $5B in Anthropic for 2GW GPU Deployment(3 posts)→
More from coding & agent
- theo builds his own visualizer for today's agent models, showing how cheap Luna really is — ivan_bezdomny · 2026-09-23
- Vite+ Hits RC: One Rust-Powered CLI to Replace Your Entire Web Toolchain — cnakazawa · 2026-09-23
- Tesla's in-car Grok agent books trips across Gmail, Calendar and Notion in one command — xiaohu · 2026-09-23
- Tesla's In-Car Grok Assistant Now Executes Cross-App Tasks in One Sentence — xiaohu · 2026-09-23
- Garry Tan says Capy lets him ship PRs much faster than Codex or Claude Code — garrytan · 2026-09-23
- DeskPilot: open-source native Python desktop client for local LLMs with MCP and sandboxed tools — poofph · 2026-09-23