The co-founders of startup Ricursive Intelligence seem destined to be co-founders.
Anna Goldie, the CEO, and Azalea Mirhosseini, the CTO, are so well-known in the AI community that they were among the AI engineers who “received these weird emails from Zuckerberg making us crazy offers,” Goldie told TechCrunch with a laugh. (They did not accept offers.) The two worked together at Google Brain and were among the first employees at Anthropic.
They won Google’s favor by creating Alpha Chip — an artificial intelligence tool that can create solid layouts for chips in hours — a process that typically takes human designers a year or more. The tool helped design three generations of Google’s Tensor Processing Units.
This pedigree explains why last month, just four months after launching Ricursive, they announced a $300 million Series A round at a $4 billion valuation led by Lightspeed, just two months after raising a $35 million seed round led by Sequoia.
Ricursive builds the AI tools that design the chips, not the chips themselves. This makes it fundamentally different from almost every other AI chip startup: it’s not a wannabe Nvidia competitor. In fact, Nvidia is an investor. The GPU giant, along with AMD, Intel and every other chip maker, are the startup’s target customers.
“We want to enable any chip, like a custom chip or a more traditional chip, or any type of chip, to be built in a very automated and fast way,” Mirhosseini told TechCrunch. “We use artificial intelligence to do that.”
Their paths first crossed at Stanford University, where Goldie earned her doctorate while Mirhosseini was teaching computer science classes. Since then, their career has been on a roll. “We started 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, and then we left Google again on the same day. Then we started this company together on the same day,” Goldie said.
TechCrunch event
Boston, MA
|
June 23, 2026
During their time at Google, the colleagues were so close that they would exercise together, and they both enjoyed circuit training. The pun was not lost on Jeff Dean, the famous Google engineer who was their collaborator. He called the Alpha Chip project “Chip Circuit Training” — a play on the common exercise routine. Internally, the duo also had the nickname: A&A.
The Alpha chip has earned them fame in the industry, but it has also generated controversy. In 2022, one of their colleagues at Google was fired, Wired reportedhaving spent years trying to discredit A&A and its chip work, even though that work has been used to help produce some of Google’s most important AI chips.
Their Alpha Chip project at Google Brain demonstrated the concept that would become Ricursive, which is using artificial intelligence to dramatically speed up chip design.
Designing chips is difficult
The problem is that computer chips have millions to billions of logic gate components built into their silicon chip. Human designers can spend a year or more putting these components on the chip to ensure performance, good power usage, and any other design needs. Numerically determining the exact position of these tiny components is difficult, as you might expect.
The Alpha Chip “can create a high-quality layout in about six hours,” Goldie said. The great thing about this approach is that he actually learns from experience.”
The basic premise of their work in designing AI chips is to use a “reward signal” that evaluates how good a design is. The agent then takes that classification to “update the parameters of its deep neural network to improve,” Goldie said. After completing thousands of designs, the agent became really good. The founders say it also became faster as it learned.
The Ricursive platform will take the concept even further. The AI chip designer they are building “will learn across different chips,” Goldie said. So every slide she designs should help her become a better designer for every next slide.
The Ricursive platform also takes advantage of LLMs and will handle everything from component placement to design verification. Any company that makes electronics and needs chips is its target customer.
If their platform proves itself, as it seems likely to do, Ricursive could play a role in achieving the ultimate goal of achieving artificial general intelligence (AGI). In fact, their ultimate vision is to design AI chips, which means AI will essentially design its own computer brains.
“Chips are the fuel of artificial intelligence,” Goldie said. “I think by building more powerful chips, this is the best way to advance those limits.”
Mirhosseini adds that the long chip design process limits the speed of AI progress. “We think we can also enable this rapid co-evolution of the models and chips that essentially power them,” she said. So AI can get smarter faster.
If the idea of AI designing its own brains at ever-increasing speeds brings to mind visions of Skynet and the Terminator, the founders point out that there is a more positive and immediate benefit that they believe is more likely: hardware efficiency.
When AI labs can design more efficient chips (and eventually all underlying hardware), their growth won’t have to consume so many of the world’s resources.
“We can design a computer architecture that uniquely fits this model, and we can achieve approximately a 10x performance improvement per total cost of ownership,” Goldie said.
While the fledgling startup hasn’t named its first customers, the founders say they’ve heard from every big name you could imagine. Unsurprisingly, they choose their first development partners as well.
