Ex-Google AlphaChip founders raise $335M for AI that designs chips

Ricursive Intelligence, a startup founded by the two researchers who co-led Google's AlphaChip project, is building AI that designs chips, learns from each design and carries that experience into the next one. According to TechCrunch, the company's goal is to compress a chip design cycle that today takes two to three years down to weeks.

Co-founders Anna Goldie and Azalia Mirhoseini are set to lay out the approach at TechCrunch Disrupt 2026, held October 13–15 at Moscone West in San Francisco, in a session titled "When AI Starts Designing Its Own Hardware."

Automating a slow process

The motivation, as TechCrunch frames it, is a speed mismatch. AI models keep improving at a rapid clip, while the silicon underneath them advances on a much slower development schedule. Designing a chip remains slow and complex precisely at the moment when increasingly capable AI systems need increasingly powerful hardware.

Ricursive's answer is AI tooling that automates and accelerates design work. Instead of treating every chip as a standalone project, the system is meant to learn across different chips, so that what it picks up designing one informs how it approaches the next. Per TechCrunch, the company wants to automate stages from component placement through design verification.

A track record from AlphaChip

Goldie and Mirhoseini have already shown that AI can do real work here. At Google, the two co-led AlphaChip, a system that produced chip layouts in hours rather than the far longer timelines human design teams require. That work fed into multiple generations of Google's Tensor Processing Units, TechCrunch reports.

Their backgrounds stack up: the pair also co-founded Google's ML for Systems team, were early employees at Anthropic and held senior staff research scientist roles at Google DeepMind. Goldie, Ricursive's founder and CEO, holds a Ph.D. in computer science from Stanford and was named one of MIT Technology Review's 35 Innovators Under 35. Mirhoseini, founder and CTO, is an assistant professor of computer science at Stanford and founded its Scaling Intelligence Lab.

A fast climb to a $4 billion valuation

Ricursive launched in late 2025, and its funding came together quickly. Within four months, the company raised $335 million at a $4 billion valuation, including a $300 million Series A, with Nvidia among its investors, according to TechCrunch.

The loop the company is chasing

The larger bet is a feedback cycle: AI-assisted design yields better chips, better chips enable stronger models, and those stronger models help design the hardware that follows. Goldie and Mirhoseini believe that accelerating chip design could open the door to new chip architectures — and, ultimately, to more capable and efficient AI — rather than just speeding up today's workflows.

Why it matters

Compute is one of the defining constraints on AI progress, and chip design sits upstream of it. If design cycles shrink from years to weeks, the industry could iterate on hardware architectures at something closer to software pace, exploring ideas that are currently too slow or costly to attempt. The founders' AlphaChip record makes the claim credible rather than speculative, and Nvidia's presence in the round suggests established hardware players see the same opening. Much is still unproven — the startup is young, and the weeks-long design cycle remains a stated goal rather than a demonstrated result — but even a partial closing of this loop would change how quickly the hardware side of AI can move.