Ricursive Intelligence, founded by Anna Goldie and Azalia Mirhoseini, is developing artificial intelligence that designs computer chips to address the growing compute demand. The startup presented its approach at TechCrunch Disrupt on Wednesday, highlighting how its AI iteratively improves chip designs to automate and accelerate the process. Goldie noted that compute demand could increase by at least 10,000 times in the next five years, necessitating more efficient chip solutions.
The AI developed by Ricursive Intelligence learns from each chip design iteration to enhance subsequent versions, aiming to optimize the use of bandwidth, wafers, and fabrication facilities. Goldie emphasized the need to avoid building exponentially more data centers, suggesting that efficiency gains in chip design are critical. She also highlighted the shift where AI handles execution tasks, allowing humans to focus on creativity and new applications in computing.
This approach comes amid a broader industry challenge known as the compute crisis, where demand for processing power is outpacing traditional hardware development. Ricursive Intelligence’s AI-driven design contrasts with conventional chip engineering, which is slower and less adaptable. By automating design, the startup hopes to improve resource utilization in data centers and fabs, potentially reducing environmental impact and operational costs compared to scaling infrastructure alone.
Ricursive Intelligence’s founders presented their technology on October 14 at TechCrunch Disrupt, signaling a push toward AI-augmented chip development. Anna Goldie stated that the future of computing relies on AI managing execution while humans contribute innovative ideas, a model already emerging in software development and other fields.