Ricursive Intelligence, a startup founded by Anna Goldie and Azalia Mirhoseini, raised $335 million in just four months, achieving a $4 billion valuation. The company’s rapid ascent is rooted in the founders’ deep expertise and a proven track record in AI-driven chip design. Goldie, the CEO, and Mirhoseini, the CTO, first met at Stanford, where Goldie earned her PhD and Mirhoseini taught computer science. Their careers have run in parallel ever since—starting at Google Brain on the same day, leaving together, joining Anthropic at the same time, and returning to Google before launching Ricursive. They even worked out together, a habit that earned them the internal nickname A&A and inspired Jeff Dean, a top Google engineer, to dub their groundbreaking project “chip circuit training.” At Google Brain, they developed Alpha Chip, an AI system capable of generating high-quality chip layouts in just six hours—something that typically takes human designers a year or more. The tool played a key role in designing three generations of Google’s Tensor Processing Units, critical hardware for the company’s AI infrastructure. Their work earned widespread recognition but also sparked controversy, including the firing of a colleague who opposed their approach. Ricursive is building AI tools that automate and accelerate chip design, not manufacturing. Unlike most AI chip startups aiming to rival Nvidia, Ricursive targets chipmakers themselves—companies like Nvidia, AMD, and Intel—by offering a platform that uses AI to optimize every stage of chip development, from component placement to design verification. The platform leverages large language models and learns from each design, improving over time. The founders believe their AI can enable a faster co-evolution of AI models and the hardware that powers them. “Chips are the fuel for AI,” Goldie said. “By building more powerful chips, we can advance the frontier of artificial general intelligence.” Mirhoseini added that current chip design timelines are a bottleneck for AI progress, and Ricursive’s platform could help break that barrier. The ultimate vision is for AI to design its own chips—essentially creating the hardware that runs future AI systems. This could lead to massive efficiency gains, with potential improvements of up to 10x in performance per dollar. The founders stress that while the idea of self-improving AI systems may evoke sci-fi scenarios, the immediate benefit is more practical: drastically reduced energy and resource consumption in AI development. Ricursive’s rapid fundraising—$35 million in seed funding followed by a $300 million Series A led by Lightspeed—reflects strong investor confidence. The company has already attracted interest from every major chipmaker, giving it a wide choice of early partners. Despite its ambitious goals, Ricursive remains focused on delivering tangible improvements in speed, efficiency, and scalability for the next generation of AI hardware.