Ricursive Intelligence raises $335M to automate chip design

Ricursive Intelligence has secured $335M in total capital just 4 months after launching. The funding includes:

• A $300 million Series A round at a $4B valuation, led by Lightspeed.
• A $35 million seed round, led by Sequoia.

The startup does not manufacture chips itself. Instead, Ricursive builds AI tools to automate the design process, positioning chipmakers like Nvidia, AMD, and Intel as target customers rather than competitors. Nvidia is also an investor in the company. According to co-founder and CTO Azalia Mirhoseini, the company wants to enable any chip, whether a custom chip or a more traditional chip, to be built in an automated and accelerated way using AI.

Co-founders Anna Goldie, who serves as CEO, and Mirhoseini previously worked at Google Brain (Google’s AI research division) and Anthropic. At Google, they created Alpha Chip, an AI tool that designed layouts for three generations of Google’s Tensor Processing Units (Google’s custom AI chips). Designing chips is difficult because millions to billions of logic gate components (the basic building blocks of a digital circuit) must be placed on a silicon wafer. While designers can spend a year or more on this process, Goldie noted that Alpha Chip was able to generate high-quality layouts in about six hours and could learn from experience. However, their tenure at Google also involved controversy: one of their colleagues at Google was fired, Wired reported, after he spent years trying to discredit A&A and their chip work.

Ricursive’s platform will expand this methodology to allow its AI designer to learn across different chips. The founders believe this approach could play a role in the goal of achieving artificial general intelligence. Mirhoseini added that the startup aims to enable the fast co-evolution of AI models and the chips that power them. By designing hardware tailored to specific models, the company hopes to achieve massive efficiency gains. “We could design a computer architecture that’s uniquely suited to that model, and we could achieve almost a 10x improvement in performance per total cost of ownership,” said Anna Goldie, CEO.

Why it matters

Ricursive is moving beyond standard chip design by using AI to automate the process. This approach could solve hardware development bottlenecks and accelerate the co-evolution of AI models and the hardware that powers them.