So TechCrunch dropped [a pretty wild story yesterday](https://techcrunch.com/2026/02/16/how-ricursive-intelligence-raised-335m-at-a-4b-valuation-in-4-months/) about [Ricursive Intelligence](https://www.ricursive.com/), and honestly, I’ve been thinking about it since. A $4 billion valuation four months after founding? That’s not just fast — that’s absurd, even by Silicon Valley standards.
Here’s the backstory. Dr. Anna Goldie and Dr. Azalia Mirhoseini spent years at Google DeepMind building [AlphaChip](https://deepmind.google/blog/how-alphachip-transformed-computer-chip-design/), an AI system that could design chip layouts faster and better than human engineers. We’re talking about technology that shipped in four generations of Google’s TPU chips and got adopted by external semiconductor companies. It was a big deal when the research was published in Nature. Now both of them have left Google to commercialize that work — and the money followed immediately.
The numbers are honestly staggering. They pulled in a [$35 million seed round from Sequoia Capital](https://www.prnewswire.com/news-releases/ricursive-intelligence-launches-frontier-ai-lab-to-transform-semiconductor-design-and-accelerate-path-toward-artificial-superintelligence-302630776.html), then just two months later closed a [$300 million Series A led by Lightspeed Venture Partners](https://news.crunchbase.com/venture/startup-ai-lab-ricursive-seriesa-unicorn/). The investor list reads like a who’s who — DST Global, Felicis, Radical Ventures, and notably NVentures, NVIDIA’s own venture arm. Yes, NVIDIA is literally investing in a company that might compete with them. Make of that what you will.
What makes Ricursive interesting to me is the core thesis: AI designs better chips, better chips run better AI, which then designs even better chips. It’s a feedback loop — recursive self-improvement, hence the name. Mirhoseini told TechCrunch they want to “enable any chip to be built in an automated and very accelerated way.” If they pull that off, the implications for the semiconductor industry are massive. Traditional chip design takes months or years. Ricursive is betting they can collapse that timeline dramatically.
One more thing I find fascinating about the founders — Goldie and Mirhoseini have been practically inseparable career-wise. They started at Google Brain on the same day, left on the same day, joined Anthropic on the same day, left on the same day, went back to Google on the same day, and then left again to start Ricursive — on the same day. That’s either the greatest co-founder chemistry ever or the most coordinated LinkedIn update strategy in history.
Whether Ricursive can actually deliver on the hype is still an open question. But with this team, this much capital, and [backing from investors like Sequoia](https://sequoiacap.com/article/partnering-with-ricursive-intelligence-a-premier-frontier-lab-pioneering-ai-for-chip-design/), they’ve got a real shot at shaking up how chips get made. I’ll be watching this one closely.
Leave a comment