Ricursive Intelligence Bets on AI-Designed Chips to Help Solve Compute Crisis
Ricursive Intelligence is betting on AI-designed chips to ease a looming compute crisis, its founders said at TechCrunch Disrupt.
Goldie said on stage that there is an understanding that in the next five years demand for compute will be at least 10,000 times greater, and that estimate may be conservative. She said the aim is not to build 10,000 times more data centers. Instead, she argued, the industry needs help making chips more efficient, using bandwidth in data centers more efficiently and using wafers and fabs more efficiently. She added that even if data centers were built in space, the sun would no longer be visible.
Goldie addressed concerns that AI might design its own chips, arguing that letting AI handle chip design would allow humans to focus on developing applications and exploring new possibilities that expand what computing can accomplish. She said the shift is toward humans bringing great ideas and creative thoughts while AI handles execution, and that this is happening in chip design, software development and other areas. Ricursive does this for chip design, she said.
The founders said AI is challenging traditional chip design methods by discovering unconventional layouts that can outperform traditional ones. AlphaChip, the AI system the founders previously developed, produced a circular, “donut-shaped” arrangement of chip components that initially surprised engineers but ultimately proved successful, they said.
They explained that they are optimizing entire chips rather than following the traditional approach of partitioning them into dozens of smaller blocks. This approach allows AI to create intricate patterns, including wave-like designs, and the pair said they plan to share more about that in the future.
The founders also argued that AI-designed chips will make it possible to create all sorts of chips that might not have been possible before. Goldie said if designing a new chip takes six months, a year or two years, companies have to be extremely selective about the chips they design. But if a chip can be designed in about a week, then for manufacturing it becomes financially sensible to design custom chips for applications including space, AR, VR and biological computing, while also being more energy efficient, she said.