Ricursive Intelligence Pitches AI-Designed Chips at TechCrunch Disrupt

Ricursive Intelligence founders Anna Goldie and Azalia Mirhoseini described their startup's approach to automated chip design on stage at TechCrunch Disrupt in San Francisco on Oct. 14, according to TechCrunch.

The company is building AI that designs chips, learns from each iteration and applies those lessons to later designs, with the aim of speeding up a process that the founders said can take six months to two years. Goldie said demand for compute could grow at least 10,000-fold over the next five years, and that making chips, data center bandwidth, wafers and fabs more efficient is preferable to building far more data centers.

The founders said their system optimizes whole chips instead of splitting them into dozens of smaller blocks, which produces unconventional layouts such as wave-like patterns. They pointed to AlphaChip, a system they previously developed, which generated a circular, "donut-shaped" arrangement of components that surprised engineers before proving effective.

Faster design cycles would make custom chips economical for applications including space, augmented and virtual reality, and biological computing, Goldie said. She added that humans would supply ideas while AI handles execution in chip design and software development.

TechCrunch Disrupt ran from Oct. 13 to Oct. 15, 2026.

This story was first reported by TechCrunch.