Agentrys raises $24.5 million to automate chip design with AI agents, giving the startup fresh resources to expand its Agentic Design Automation platform.
Seed and pre‑seed financing details
The round combined a $19.1 million oversubscribed seed round with a $5.4 million pre‑seed tranche. Etna Labs led the seed financing, while MediaTek headed the earlier pre‑seed investment.
According to the filing, the capital will fund talent recruitment, development of agent‑native tooling, and broader customer engagements across verification and physical design.
Agentrys’ founder and CEO, Mark Ren, noted that the money will help the company scale engineering and research capabilities.
Agentic Design Automation and its workflow
Traditional electronic design automation (EDA) tools automate discrete tasks. Agentrys’ approach, called Agentic Design Automation (ADA), lets semiconductor teams build AI workforces that oversee entire design workflows.
The platform plugs into commercial EDA suites, internal software, and infrastructure, allowing autonomous agents to act within existing environments.
Its open architecture lets customers construct and own AI workforces tailored to their processes. A design intelligence layer learns from workflow activity and evaluation signals.
Related: Hike Medical Raises $22.5M for Device-Based Care
In a demonstration, the system took a 32‑bit CPU from specification through sign‑off‑clean GDS layout without human intervention. The company also reported over 90 % accuracy on NVIDIA’s public CVDP verification benchmark.
Agentrys is already collaborating with major fabless semiconductor firms, a global foundry, and emerging chip startups. Current engagements span digital and analogue design flows, while system‑level design remains on the roadmap.
The model could help firms retain specialized knowledge while reducing reliance on scarce engineering talent. It also allows design teams to customize AI agents around proprietary processes and infrastructure.
Market reaction and future outlook
Etna Labs highlighted that chip design lends itself to recursive AI improvement because results can be evaluated objectively. MediaTek pointed to the chance to capture semiconductor expertise and reuse it systematically across projects.
Ren, who previously worked on AI and EDA research at NVIDIA Research and IBM Research, said building production‑grade agents that reliably automate real engineering work is far from easy.
With the new funding, Agentrys plans to roll out agent‑native tools for additional semiconductor workflows and deepen engagements in verification and physical design.
Leave a Reply