Jeff Bezos is about to seal one of the most eye-catching early-stage fundraising campaigns the artificial intelligence sector has yet produced: his budding physical AI lab, Project Prometheus, is reportedly closing a $10bn (£7.9bn) round that would value the company at $38bn.
The Financial Times reported on Monday, citing people familiar with the matter, that BlackRock and JPMorgan were among the institutional heavyweights that joined the round, although the deal has not yet closed. BlackRock declined to comment. If the fundraising is completed on the agreed terms, Prometheus would become one of the highest valued early-stage AI companies in the world less than six months after emerging from stealth.
Quietly founded in November 2025 with $6.2 billion in initial funding, Prometheus has a very different thesis than the generative AI giants that have dominated the investment cycle since ChatGPT launched in late 2022. Instead of training ever-larger language models on the text and images of the Internet, the company is building systems that can reason about the physical world itself, materials, tolerances, processes and the immutable laws of physics. The stated target markets are engineering, manufacturing, aerospace, robotics, drug discovery and logistics automation, sectors with which large language models have so far only made cursory contact.
Day-to-day management of the show is CEO Vikram Bajaj, a former Google X scientist and co-founder of Foresite Labs. The lab has grown to more than 120 employees, poached from companies like OpenAI, xAI, Meta and DeepMind. Bezos, who was described as one of the early backers, led fundraising alongside Bajaj and notably took on an operational role in the company. It is the first time that the Amazon founder has rolled up his sleeves at a technology company since he stepped down as CEO of the company he built in 2021.
The timing is striking. Prometheus’ raise comes just days after Amazon itself pledged up to $25 billion in new investment in Anthropic, securing a $100 billion commitment from the Claude maker for cloud spending in return. $10 billion in funding for a six-month-old lab would potentially exceed the lifetime funding of most existing AI companies.
Why are institutions the size of BlackRock and JPMorgan willing to write checks of this size to an unproven company? The answer lies in the special economics of physical AI. Unlike the vast amounts of cheap, publicly available text and code that underpin today’s language models, the data needed to teach a machine how steel wears, how a drug molecule binds, or how a robotic arm should select a part is proprietary, scarce, and fiendishly expensive to collect. This scarcity is itself a moat, and accumulating it early can give a lasting advantage to whichever lab overcomes it first.
For Britain’s small and medium-sized manufacturers, aerospace suppliers and life sciences specialists, many of which already have decades of unique operational data, the emergence of a well-capitalized, Bezos-backed lab is a development worth watching. If Prometheus lives up to its ambitions, the model for applying AI to the industrial economy will not be built on the backs of scraped web pages, but rather on partnerships with the companies that actually make, repair and move things.
Of course, that’s a significant “if.” Prometheus has not yet publicly demonstrated a product, let alone commercial deployment, and the lab is still in its early stages. Many skeptics will also point out that the broader AI market is increasingly showing stretched valuations. Peter Fedoročko, chief technology officer at analytics firm GoodData, takes a measured view. “Yes, AI has a bubble, but the technology is real,” he argues. “When dot-com crashed, the internet didn’t disappear, it became infrastructure. The same thing is happening here. It took a decade to recover financially from the dot-com crash, but the internet changed everything in that time. It didn’t kill jobs, it changed them. AI follows the same pattern. Once the hype dies down, the real developers get back to work.”
For Bezos the calculation is simpler. Having built the world’s largest logistics and cloud empire on the back of an earlier wave of technology, he is now betting personally and on scale that the next one will be written not in pixels and prose, but in physics.
