The co-founders of startup Ricursive Intelligence seemed destined to become co-founders.
CEO Anna Goldie and CTO Azalia Mirhoseini are very well-known in the AI community, and they were among the AI engineers who “received a weird email from Zuckerberg with a crazy offer,” Goldie told TechCrunch with a laugh. (They didn’t accept the offer.) The two worked together at Google Brain and were early employees at Anthropic.
They gained recognition at Google for creating Alpha Chip, an AI tool that can generate solid chip layouts in hours. This process typically takes a year or more for human designers. This tool helped design three generations of Google’s Tensor Processing Units.
This pedigree explains why Ricursive announced a $300 million Series A round at a $4 billion valuation led by Lightspeed last month, just four months after launching and just two months after raising a $35 million seed round led by Sequoia.
Ricursive is building the AI tools that design the chips, not the chips themselves. That makes the company fundamentally different from almost every other AI chip startup. So they’re not trying to be a competitor to Nvidia. In fact, Nvidia is also an investor. The GPU giant is the startup’s target customer, as are AMD, Intel, and every other chipmaker.
“We want to be able to build all kinds of chips, including custom chips and more traditional chips, in an automated and very fast way, and we’re using AI to do that,” Mirhosseini told TechCrunch.
Their paths first crossed at Stanford University, where Mirhoseini taught a computer science class and Goldie earned her Ph.D. Since then, their career has been on a roll. “We joined Google Brain on the same day. We left Google Brain on the same day. We joined Anthropic on the same day. We left Anthropic on the same day. We rejoined Google on the same day. We left Google again on the same day. And we founded this company together on the same day,” says Goldie.
tech crunch event
boston, massachusetts
|
June 23, 2026
During my time at Google, my colleagues were close enough to work out together, and we both enjoyed circuit training. The pun also made its way to their collaborator, famed Google engineer Jeff Dean. He named their Alpha Chip project “Chip Circuit Training.” This is a play on the workout routines they share. Internally, the pair were also given the nickname “A&A.”
The Alpha chip attracted industry attention, but it also sparked controversy. Wired reported that in 2022, one of his colleagues at Google was fired after spending years trying to discredit A&A and its chip work, even though that work was used to produce AI chips, Google’s most important and business bet.
The Alpha Chip project at Google Brain proved the concept that would become Ricursive, using AI to dramatically accelerate chip design.
Chip design is difficult
The problem is that computer chips have millions to billions of logic gate components integrated on a silicon wafer. Human designers can spend more than a year placing these components on a chip to ensure performance, good power utilization, and other design needs. As you can imagine, accurately determining the placement of such tiny components digitally is difficult.
Alpha Chip “can produce very high-quality layouts in about six hours, and the great thing about this approach is that you actually learn from experience,” Goldie said.
The premise of their AI chip design work is the use of “reward signals” to evaluate the goodness of a design. The agent then uses that assessment to “update and improve the parameters of the deep neural network,” Goldie said. After completing thousands of designs, the agent has become very good at it. And as it learned, it got faster, the founders said.
Ricursive’s platform takes this concept even further. The AI chip designers they are developing “will be learning different chips,” Goldie said. So each chip you design should help you become a better designer for every chip that follows.
Ricursive’s platform also utilizes LLM, which handles everything from component placement to design validation. Any company that manufactures electronic equipment and requires chips is a target customer.
If Ricursive’s platform proves out, as it likely will, it could play a role in the moonshot goal of achieving artificial general intelligence (AGI). In fact, their ultimate vision is to design AI chips, which means that AI essentially designs its own computer brain.
“Chips are the fuel for AI,” Goldie says. “I think building more powerful chips is the best way to move that frontier forward.”
Mirhosseini added that the speed of AI progress is limited by the long chip design process. “We think we can also enable this rapid co-evolution of models and the chips that essentially drive them,” she said. So AI can grow faster and smarter.
If the idea of AI designing its own brain at increasingly faster speeds conjures up visions of Skynet or the Terminator, the founders point to the benefits of what they see as more aggressive, immediate, and more likely. It’s about hardware efficiency.
If AI Labs can design much more efficient chips (and ultimately all the underlying hardware), their growth won’t require as much of the world’s resources.
“If you can design a computer architecture that is uniquely suited to that model, you can achieve nearly a 10x improvement in performance per total cost of ownership,” Goldie said.
The young startup won’t reveal the names of its early customers, but its founders say they’ve heard from every major chipmaker imaginable. Naturally, they also choose their own initial development partners.
