Ricursive Intelligence, a young startup building an AI system to design and automatically improve AI chips, has raised $300 million in a Series A round at a $4 billion valuation, the company said, with Lightspeed leading the financing.
The pitch is unusually ambitious even by Silicon Valley standards: Ricursive says it is developing a system that can create its own silicon substrate layer and accelerate chip improvements in a rapid, continuous loop, an approach the founders believe could, over time, push the system toward artificial general intelligence (AGI).
Ricursive Intelligence: A fast follow after launch and seed funding
The Series A closed just two months after Ricursive formally launched, after a seed investment led by Sequoia, according to the report. With this new cheque, the startup has raised $335 million in total funding.
Ricursive is led by Anna Goldie (CEO) and Azalia Mirhoseini (CTO), both described as former Google researchers.
The company is commercialising research linked to a reinforcement-learning method for designing chip layouts called AlphaChip, which the startup says has already powered four generations of Google’s TPU chips.
Who else backed the round
Along with Lightspeed, the funding round included DST Global, Nvidia’s venture capital arm NVentures, Felicis Ventures, 49 Palms Ventures, and Radical AI.
Why investors are leaning into “self-improving” chip design
The broader theme here is investor appetite for startups that blur the line between software intelligence and hardware creation, AI systems that don’t just run on chips, but help invent the next ones.
Ricursive isn’t alone in chasing that category. The report also points to Unconventional AI, an AI hardware startup founded by Naveen Rao, which is described as developing an “intelligent substrate” as well.
In December, Unconventional AI raised a $475 million seed round at a $4.5 billion valuation, with Andreessen Horowitz and Lightspeed Ventures leading, alongside Lux Capital and DCVC.
Ricursive’s quick back-to-back fundraising and its technical roadmap underline what the report calls growing conviction in AI systems that can design and iterate on, their own hardware. With multiple well-funded startups now converging on this “self-improving AI” paradigm, the fight to reinvent chip design is heating up.
Also Read: ICF Steps Up Production of Vande Bharat Sleeper Trains
