AI
AI-Designed Chips Could Shrink Development Cycle From Years to Weeks
Ricursive Intelligence co-founders are bringing their self-improving chip-design systems to TechCrunch Disrupt, promising to close the loop between artificial intelligence and hardware innovation.
TechCrunch — AI
As AI systems grow more sophisticated, the demand for faster, more powerful processors intensifies — yet the process of designing those very chips remains slow and labor-intensive. What if AI itself could accelerate that bottleneck?
That's the central mission of Ricursive Intelligence, a startup founded by Anna Goldie and Azalia Mirhoseini that is automating and speeding up chip design. The pair envision a self-reinforcing cycle: AI systems design better hardware, which in turn enables more capable AI models, which then improve the next generation of chip architecture.
Shrinking the Design Timeline#
Currently, chip development takes two to three years from conception to completion — a timeline that increasingly lags behind the rapid pace of AI advancement. Ricursive aims to compress that dramatically, reducing the cycle to just a matter of weeks by automating complex tasks like component placement and design verification.
The founders aren't starting from scratch with this vision. Before launching Ricursive in late 2025, both co-led AlphaChip at Google, an AI system that demonstrated it could generate chip layouts in hours rather than the months required by human engineers. Their work contributed to the design of multiple generations of Google's Tensor Processing Units.
From Research to Scale#
Ricursive's founding attracted swift investor confidence. Within four months of launch, the company raised $335 million at a $4 billion valuation, with Nvidia among its backers and a $300 million Series A closing relatively quickly.
Goldie, the company's CEO, holds a Ph.D. from Stanford and was named to MIT Technology Review's list of 35 Innovators Under 35. Mirhoseini, the CTO, is an assistant professor of computer science at Stanford and leads its Scaling Intelligence Lab. Both were early members at Anthropic and held senior research roles at Google DeepMind.
Why This Matters for the AI Race#
The hardware-software feedback loop matters enormously. Today's AI architectures are constrained by the capabilities and timelines of chip development — hardware innovation effectively sets the pace. By automating and learning across multiple chip designs, Ricursive's approach could unlock new architectural possibilities and allow hardware to keep stride with algorithmic breakthroughs.
Goldie and Mirhoseini will discuss this vision at TechCrunch Disrupt 2026 (October 13–15 in San Francisco), exploring how self-improving systems and tighter AI-hardware integration could reshape the trajectory of AI development.
