- ■
Ricursive Intelligence raised $300M Series A at $4B valuation led by Lightspeed, just two months after launching with a Sequoia-led seed round
- ■
The startup has raised $335M total and is building AI systems that design and improve AI chips autonomously using technology derived from Google’s AlphaChip
- ■
Ricursive joins similarly named Recursive (Richard Socher) and Unconventional AI (Naveen Rao) in the billion-dollar AI chip startup cohort, all raising at $4B+ valuations
- ■
The funding frenzy signals massive investor confidence in self-improving AI hardware as the next frontier beyond current GPU architectures
Ricursive Intelligence just pulled off one of the fastest climbs to unicorn status in tech history. The AI chip startup raised $300 million at a $4 billion valuation just two months after formally launching, according to an announcement Monday. The company, founded by former Google researchers Anna Goldie and Azalia Mirhoseini, is building AI systems that design and automatically improve their own chips – and investors are betting billions that this recursive approach could accelerate the path to AGI.
Ricursive Intelligence just became the latest AI chip startup to hit unicorn status at warp speed. The company announced Monday it raised $300 million in Series A funding at a $4 billion valuation, led by Lightspeed Venture Partners. The round comes just two months after the startup formally launched with a seed investment led by Sequoia Capital, bringing total funding to $335 million according to the New York Times.
The breakneck fundraising pace reflects growing investor conviction that AI-designed chips represent the next leap in computing infrastructure. While Nvidia continues to dominate AI hardware with its GPUs, a new wave of startups is betting that chips designed by AI systems – rather than human engineers – can break through current performance bottlenecks.
Ricursive’s technology builds on breakthrough research from Google. Founders Anna Goldie (CEO) and Azalia Mirhoseini (CTO) pioneered a reinforcement learning method for chip design called AlphaChip during their time at Google Research. The approach has already been deployed in four generations of Google’s Tensor Processing Unit chips, the company says. Now they’re taking that foundation further – building systems that don’t just design chips, but create their own silicon substrate layers and continuously improve performance in a recursive loop.
The investor lineup reads like a who’s who of top-tier venture capital. Beyond lead investor Lightspeed, the round includes DST Global, venture arm NVentures, Felicis Ventures, 49 Palms Ventures, and Radical AI. That Nvidia is backing a potential competitor speaks volumes about the strategic importance of next-generation chip architectures.
