AI designs chip from spec to hardware in 2 weeks
rohanpaul_ai · x · 2026-08-30
A new paper demonstrates an AI system converting chip specifications into working hardware, marking a shift in chip design speed.
- Case Study (Redwood): A small, low-power inference chip. 2 human architects wrote the spec; AI generated the design, tests, firmware, and kernels.
- Efficiency: Spec-to-hardware loop took under 48 hours. The first FPGA design had 0 bugs with 95% test coverage.
- Performance: Runs Qwen3-0.6B at 12.1 t/s on AMD FPGA. Projected 49 t/s @ 1.335W on Samsung 8nm, offering 3.4x better tokens-per-watt than NVIDIA Jetson Orin Nano.
Hardware development could soon match the pace of model releases.
More from Infra
- Optimal Settings for Llama.cpp + Qwen 3.8: n-max 4 Fastest — GodComplecs · 2026-08-31
- PyTorchCon to Feature vLLM Sessions on Serving Stack — PyTorch · 2026-08-31
- Oracle nears $1T market cap as AI infrastructure captures top value — thedealdirector · 2026-08-30
- Grok Bot Provides Cloud Desktop Environment for Bots — tristanbob · 2026-08-30
- Token Economics: How Cache & Latency Impact Pricing — AccBalanced · 2026-08-30
- NVIDIA Studio Driver Optimizes ComfyUI and LTX-2.5 Performance — PixWizardry · 2026-08-30