The co-founder of startup Ricursive Intelligence seemed destined to become a co-founder.
CEO Anna Goldie and CTO Azalia Mirhoseini are so well known in the AI community that they were among the AI engineers who “received strange emails from Zuckerberg giving us crazy suggestions.” Goldie told TechCrunch with a laugh. (They didn’t take the offer up.) The two worked together at Google Brain and were early employees of Anthropic.
They created Alpha Chip, an AI tool that can generate robust chip layouts in a matter of hours, and was well-received by Google. This process typically takes a human designer more than a year. This tool helped design Google’s third generation of Tensor processing units.
This pedigree explains why, just four months after launching Ricursive, it announced a $300 million Series A round led by Lightspeed at a $4 billion valuation, just two months after raising a $35 million seed round led by Sequoia.
Ricursive is building AI tools that design chips rather than the chips themselves. This is what makes them fundamentally different from almost every other AI chip startup: They’re not wannabe Nvidia competitors. In fact, Nvidia is an investor. The GPU giants are the startup’s target customers, along with AMD, Intel and all other chip manufacturers.
“We want to be able to build any chip, be it a custom chip or a more traditional chip, any kind of chip, in an automated and highly accelerated way. We’re using AI to do that,” Mirhoseini told TechCrunch.
Their paths first crossed at Stanford, where Goldie earned her doctorate while Mirhoseini taught computer science classes. Since then, their careers have been on track. “We started at Google Brain the same day. We left Google Brain the same day. We joined Anthropic the same day. We left Anthropic the same day. We rejoined Google the same day. We left Google again the same day. And we started this company together the same day,” Goldie said.
Tech Crunch Event
Boston, Massachusetts
|
June 23, 2026
While at Google, my colleagues were close enough to work out together and both enjoyed circuit training. The puns of their collaborator and renowned Google engineer Jeff Dean were also not forgotten. He nicknamed the Alpha Chip project “Chip Circuit Training.” This takes advantage of shared exercise routines. Internally, the two also earned the nickname A&A.
Alpha Chip has garnered industry attention, but it has also sparked controversy. Wired reported in 2022 that one of Google’s colleagues was fired after spending years trying to discredit A&A and its chip operations. Even though that work was used to produce some of Google’s most important business betting AI chips.
Google Brain’s Alpha Chip project demonstrated the Ricursive concept, which uses AI to dramatically accelerate chip design.
Designing a chip is difficult
The problem is that computer chips have millions to billions of logic gate components integrated into the silicon wafer. Human designers can spend a year or more placing those components on a chip to ensure performance, good power utilization, and other design requirements. As you might expect, it is difficult to digitally determine the placement of very small components with precision.
Alpha Chip “can produce a very high quality layout in about six hours, and the cool thing about this approach is that you actually learn from experience,” Goldie said.
The premise of AI chip design work is to use “reward signals” to evaluate how good the design is. The agent then uses that evaluation to “update the parameters of the deep neural network to get better results,” Goldie said. After completing thousands of designs, my agent got really good. Additionally, the founders say that the speed has increased as they learn.
Ricursive’s platform will take this concept further. The AI chip designers they are building will “learn from a variety of chips,” Goldie said. So each chip you design should help you become a better designer for each next chip.
Ricursive’s platform also uses LLM and handles everything from component placement to design verification. Any company that makes electronics and needs chips is its target customer.
If Ricursive’s platform proves itself, Ricursive 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 will essentially design its own computer brain.
“Chips are the fuel of AI,” Goldie said. “We believe that creating more powerful chips is the best way to advance the field.”
Mirhoseini added that long chip design processes are limiting the pace of progress in AI. “We think we can enable rapid co-evolution of the models and essentially the chips that drive them,” she said. So AI can get smarter faster.
If the idea of AI designing its own brain at increasingly faster speeds conjures up visions of Skynet and the Terminator, the founders believe the benefits are more positive, more immediate, and more numerous. That’s hardware efficiency.
If AI labs could design much more efficient chips (and ultimately all underlying hardware), their growth wouldn’t have to consume so much of the world’s resources.
“We were able to design a computer architecture that was uniquely suited to that model and deliver nearly a 10-fold improvement in performance per total cost of ownership,” Goldie said.
Although the young startup doesn’t reveal the names of its early customers, its founders say they’ve heard of every large chipmaker imaginable. Naturally, they can choose their first development partner as well.
