Jeff Bezos may have missed the first wave of A.I., dominated by large language models. But with his startup Prometheus reaching a $41 billion valuation this month, he is betting on the frontier concept of physical A.I.—by building what he calls an “artificial general engineer” (AGE).

“This is an age-old dream, the idea that you might build a set of tools that can actually do engineering,” Bezos told CNBC earlier this month, adding that his team has been working on the concept since late 2024. Prometheus, officially founded in November 2025, is led by Bezos as co-CEO alongside Vikram Bajaj, a former director of Google X.

Details about the company remain scarce. Prometheus operates offices in London, Zurich and San Francisco and has a team of roughly 150 employees, according to Bezos. While he has claimed the startup is not being “secretive,” it has shared little publicly about its work. Observer reached out to multiple employees but did not receive a response.

What is known is that Prometheus is developing A.I.-driven tools to accelerate manufacturing and building models designed to perform physical tasks across engineering workflows. This emphasis on physical A.I. could differentiate the company in a crowded market.

“Everyone is debating which A.I. model wins. Bezos is asking a different question: what happens when A.I. can compress the time between an idea and a manufacturable product?” Shashi Bellamkonda, principal research director at Info-Tech Research Group, told Observer.

“The ‘artificial general engineer’ framing is ambitious, but the underlying bet that physical world data creates a moat that internet-trained models cannot replicate is the most strategically sound position I have seen staked out in this funding cycle,” Bellamkonda added.

Prometheus has assembled a high-profile team to pursue that vision. In addition to Bajaj, who co-founded Google Life Sciences (now Verily), the company has recruited Kyle Kosic, an xAI co-founder and former OpenAI employee who led infrastructure for xAI’s Colossus supercomputer and now works on A.I. infrastructure at Prometheus.

The startup also acquired General Agents last year, bringing in founders Sherjil Ozair and William Guss, with prior experience at DeepMind, Tesla and OpenAI. Additional hires have come from Microsoft, Meta, Anthropic, Nvidia and Grammarly.

Recent LinkedIn activity suggests continued hiring momentum. Drew Jaegle, a former DeepMind researcher and Head of A.I. at Mirage, joined Prometheus this year as a member of technical staff, noting on his website that he works on A.I. for the physical world. In another post, founding member Cristian Bodnar, formerly of Microsoft and co-lead of the Aurora Earth System Forecasting Model, wrote, “We are putting together a world-class team at the intersection of A.I. and materials science.”

Still, the company’s rapid rise raises questions. A $41 billion valuation is striking for a startup that has yet to ship a product. While PWC estimates the global physical A.I. market could reach nearly $500 billion, translating that potential into real-world impact remains a challenge.

“They’re skipping a step,” Nicholas Nadeau, co-founder and CTO of Onix, told Observer. “Calling it an artificial general engineer dodges the real question: are they building a world model that actually encodes the physics, or an engineer-agent that reasons through the design loop using the solvers we already have? Different bets, different data needs, and AGE lets him avoid committing to either.”

Nadeau added that Prometheus is “worth watching, mostly for who’s involved,” noting that few founders can assemble manufacturing supply chains at the scale Bezos can, given his ties to Amazon and Blue Origin.

For now, the startup joins a growing group of high-valuation, early-stage A.I. ventures, including Mira Murati’s Thinking Machines Lab ($12 billion), Yann LeCun’s AMI Labs ($3.5 billion) and Fei-Fei Li’s World Labs ($5 billion).

• SEE ALSO: Why Cutting Humans Too Fast Could Backfire in the A.I. Era