Meta Muse AI agent designs RISC-V crypto SoC for $2,000
A semiconductor expert known as @bubbleboi, who serves as a mascot for the Thru blockchain, reported using Meta's Muse AI agent to design a specialized System-on-Chip (SoC) based on the RISC-V architecture. The AI agent completed the place-and-route process using OpenROAD software, generated the final GDSII blueprint file for manufacturing, and integrated support for standard cryptographic…
Key points
- Meta's Muse AI agent designed a RISC-V crypto SoC with 61,590 logic cells for about $2,000.
- The chip supports SHA-256, Keccak, Ed25519, and secp256k1 algorithms and passed DRC and RISC-V tests.
- NVIDIA and ChipAgents report the best AI agent scored only 16.6% on their new chip design benchmark.
The resulting chip design contains approximately 61,590 logic cells, operates at 100 million cycles per second with a 4.5 nanosecond slack, and consumes only 20 milliwatts of power. The design passed Design Rule Check (DRC) without electrical issues and validated against RISC-V rules using ACT suites and 1.5M Berkeley TestFloat vectors. The entire process consumed roughly 100 million tokens, costing approximately $2,000.
The expert plans to add vector instructions and port the Thru blockchain natively to the chip, aiming to make it the fastest blockchain on earth. They are currently seeking multi-Project Wafer runs to begin the tape-out process. This achievement contrasts with recent findings from NVIDIA and ChipAgents, whose Build-A-Chip-Bench showed the best AI agent, powered by Anthropic's Claude Fable 5.1, scored only 16.6 percent, indicating significant gaps remain in AI-driven chip design reliability.
Model pages: Muse → · Claude Fable 5.1 →
The story so far
7 episodes →- Meta Muse AI agent designs RISC-V crypto SoC for $2,000this story
Meta’s Muse AI Agent Just Designed A Crypto-Specific SoC On The RISC-V Architecture For $2,000 To Transform The Thru Blockchain Into “The Fastest On Earth”
wccftech.com · 14 October 2026
Loading the full article…
This text was published by wccftech.com and written by Rohail Saleem. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Hardware & Compute
All →- Microsoft unveils Azure Maia 300 chip for 40% more tokens per megawatt in 2027 deployments · 12 src
- Apple delays Apple Intelligence launch in China to December 2026 · 1 src
- NVIDIA adds cuPotential library to CUDA-X for AI chip simulations · 3 src
- Chinese firms buy 1M Nvidia GPUs yearly amid chip shortage, report says · 3 src
- Nvidia reportedly invests in AI inference chip rival d-Matrix · 5 src
Comments
via GitHub Discussions