Google DeepMind Alumni’s AI Chip Design Startup Ricursive Intelligence Reaches $4B in Under 60 Days

PALO ALTO: A semiconductor design startup founded by former Google researchers secured three hundred million dollars at a four billion dollar valuation just eight weeks after launching, achieving one of the fastest billion-dollar milestones in venture capital history.

Lightspeed Venture Partners led the Series A round for Ricursive Intelligence, joined by DST Global, Nvidia’s venture arm NVentures, Felicis Ventures, and Sequoia Capital. The company debuted in early December with a thirty-five million dollar seed round at a seven hundred fifty million dollar valuation.

Anna Goldie and Azalia Mirhoseini, who created Google’s AlphaChip technology for automated processor design, launched Ricursive to solve what they identify as artificial intelligence’s critical constraint: the multi-year timeline required to develop advanced semiconductors. Traditional chip design cycles can span three to five years from concept to production.

The platform uses machine learning to automate chip architecture development, creating what the founders describe as a continuous improvement loop where AI designs the silicon that powers subsequent AI generations. This recursive approach aims to collapse development timelines and reduce capital requirements for custom processor creation.

AlphaChip, developed during the founders’ tenure at DeepMind, helped Google accelerate four successive generations of tensor processing units and has been adopted by external chip manufacturers. The technology demonstrated that neural networks could optimize complex physical design challenges traditionally requiring extensive human engineering.

Guru Chahal from Lightspeed characterized the hardware development bottleneck as the most significant constraint facing AI advancement today. The investment reflects confidence that automating semiconductor design could unlock faster iteration cycles across the entire AI stack.

Since its December launch, Ricursive recruited engineers from Google DeepMind, Anthropic, Apple, and Cadence Design Systems. The funding will expand the research team and build additional compute infrastructure needed for full-stack chip design automation.

Goldie emphasized that hardware determines AI progress velocity, making accelerated chip development central to advancing artificial intelligence capabilities. Mirhoseini added that the company targets maximum computational efficiency per watt, essential as energy consumption becomes a limiting factor for large-scale AI deployment.

The rapid valuation increase reflects intensifying investor focus on AI infrastructure as cloud providers and tech giants commit hundreds of billions toward data center expansion and custom silicon development.