Ricursive Intelligence raises funding to automate chip design with AI

Ricursive Intelligence, a stealth startup, describes its core technology as AI systems capable of designing chips, learning from each iteration, and improving subsequent designs. The founders, Anna Goldie and Azalia Mirhoseini, presented the approach at TechCrunch Disrupt 2026. They argue that automating chip design could help address an anticipated surge in compute demand—estimating at least a 10,000-fold increase within five years.

Goldie and Mirhoseini assert that traditional chip design methods are too slow and selective to meet future demand. They claim AI can reduce design cycles from years to weeks, enabling custom chips for applications such as space computing, augmented and virtual reality, and biological computing. They also emphasize energy efficiency as a key benefit.

The founders highlight that their prior AI system, AlphaChip, produced unconventional layouts—including a circular, 'donut-shaped' arrangement—that initially surprised engineers but ultimately proved successful. They describe optimizing entire chips rather than partitioning them into smaller blocks, allowing AI to generate intricate patterns such as wave-like designs.

Goldie and Mirhoseini position AI as a force multiplier: humans contribute ideas and creative direction, while AI handles execution in chip design, software development, and other domains. They suggest this division of labor could unlock new possibilities in computing by accelerating innovation cycles.