Jeff Dean claims AI can reduce chip design time from 18–30 months to 3–6 months

Designing a cutting-edge chip currently takes somewhere between 18 and 30 months from first sketch to fabricated silicon. Jeff Dean, former Chief Scientist at Google, thinks AI can slash that timeline to as little as three months.

The claim, made during an appearance on the Dwarkesh Podcast in February 2025, centers on a straightforward observation: most of that 18-to-30-month window is consumed by human engineers making design decisions that could, in theory, be automated. The actual fabrication at a foundry like TSMC only takes about four months. The rest is people arguing over floor plans, routing wires, and running simulations. Dean’s thesis is that reinforcement learning and improved electronic design automation tools can compress the human-intensive design phase from over a year down to a few months, with a smaller team.

From AlphaChip to production silicon

Dean didn’t pull this idea from thin air. His argument rests on years of research at Google, most notably the AlphaChip system detailed in a Nature paper published in June 2021. AlphaChip used reinforcement learning to automate chip layout, a task that typically keeps teams of engineers busy for weeks. The system could generate layouts in under six hours that matched or surpassed the quality of human expert work.

That wasn’t just a lab demo. AlphaChip was deployed in production designs for Google’s Tensor Processing Units, the custom AI accelerators that power much of the company’s machine learning infrastructure.

Dean argued that shorter design cycles would reduce the need for engineers to predict where AI research will be 18 months from now. Instead, teams could design chips tailored to current algorithmic needs and iterate quickly as those needs evolve.

The industry is already moving

Dean’s vision isn’t purely theoretical anymore. Several companies have begun demonstrating what AI-assisted chip design looks like in practice.

Architect Labs announced in September 2026 that it had designed an inference chip called Redwood, going from specification to proof-of-concept in under two weeks.

OpenAI has also entered the hardware game. Its Jalapeño AI accelerator went from initial architecture to first silicon in under 20 months, with the critical RTL-to-tape-out phase completed in nine months. The company used large language models to assist in the design process, achieving a 10% reduction in chip area compared to equivalent human designs.

Dean himself left Google in mid-2026 to co-found Discovery Loop alongside other researchers. The company’s focus: building systems for recursive self-improvement in hardware design automation.