An AI agent designed a RISC-V CPU core in about 12 hours. Now an AI agent is writing and testing its firmware on real silicon.
Embedder, the AI coding agent for firmware and embedded systems, and Verkor, the company behind the autonomous chip design system Conductor, today announced a partnership linking AI-driven silicon design to AI-driven firmware development. Together, the two companies cover the full stack — from a written spec to verified embedded software.
New silicon has always waited on software. Every chip needs bring-up code, drivers, and reference firmware before anyone can build on it — work that normally leans on vendor ecosystems built up over years. VerCore — the RISC-V core Conductor just designed — has none of that. Embedder builds that context from scratch, grounding it in the reference documentation. Then it drives the test equipment to prove the firmware actually works on real silicon.
“Chip design used to take months. Verkor just cut it down to a day,” said Ethan Gibbs, CEO of Embedder. “That makes firmware the next bottleneck, and firmware is the problem we built Embedder to solve. A brand-new core nobody has ever written code for is normally the hardest week of a firmware engineer’s career. For Embedder it’s a normal Tuesday.”
Verkor’s Conductor produced VerCore, a complete RISC-V CPU core, from a 219-word prompt in roughly 12 hours. In simulation the core runs at 1.48 GHz and posts a CoreMark score of 3,261. An FPGA implementation of Vercore is fully operational.
“A system is only proven when software runs successfully on its hardware. AI-based software and hardware co-development speeds up building and validating a fully working system, taking it from years to days. This partnership gives AI-designed silicon a software story on day one and pushes fully integrated system development and deployment to the next level for the industry,” said Suresh Krishna, CEO of Verkor.
The post Embedder and Verkor to Take Agentic AI From Chip Design to Working Firmware first appeared on AI-Tech Park.
