Prometheus, the industrial AI startup co-led by Jeff Bezos, closed a $12 billion Series B in June at a $41 billion valuation, according to reporting confirmed by Bezos himself, less than eight months after the company launched with a $6.2 billion Series A. Combined, Prometheus has now raised more than $18 billion in under eight months, a pace of capital-raising that puts it in the same conversation as the frontier AI labs even though it has nothing to do with chatbots.
Bezos, who stepped down as Amazon CEO in 2021, serves as co-CEO alongside Vik Bajaj, a scientist whose career has run almost entirely through Alphabet’s life-sciences arm rather than through robotics or industrial engineering. Bajaj holds a combined bachelor’s and master’s degree in biochemistry from the University of Pennsylvania and a PhD in physical chemistry from MIT, with postdoctoral work at Berkeley. He co-founded Google Life Sciences within Google X in 2013, serving as its chief scientific officer while it incubated projects including the Wing drone-delivery service, the self-driving car program that became Waymo, and the Project Baseline health study, before the unit was rebranded Verily in 2015. He later became chief scientific officer at Grail, the Illumina-spun-out cancer-detection company that Illumina reacquired for $8 billion, and spent several years as a managing director at the investment firm Foresite Capital before joining Bezos at Prometheus in November 2025. Under the arrangement, Bezos is expected to shape business direction and growth strategy while Bajaj drives scientific vision and large-model engineering, an unusual pairing of a retail-and-logistics billionaire with a life-sciences research executive to build what the company calls an “artificial general engineer.”
The company describes what it’s building as AI tooling meant to compress the entire pipeline from physical design to manufacturing, for objects ranging from bridges to computer chips. Bezos has been explicit about what Prometheus is not: “nothing to do with robotics,” he’s said, describing the system instead as “a very, very modern version” of computer-aided design software.
The investor list reads more like an infrastructure financing than a typical AI funding round. JPMorgan, BlackRock, Goldman Sachs, DST Global, and Arch Venture Partners all participated, institutional money that tends to price deals as long-duration infrastructure bets rather than speculative software plays. Bezos was the largest individual backer of the Series A and put in again for the Series B, a level of personal financial commitment that goes well beyond a typical founder-investor’s involvement in a company he also runs day to day.
Prometheus’s valuation jump lands inside a broader physical AI funding boom that’s accelerated sharply through 2026. Physical AI startups collectively raised roughly $6.4 billion across 27 deals of at least $50 million each in the first quarter of 2026 alone, with robotics companies absorbing about $4 billion of that and AI semiconductor and hardware firms taking the rest. Figure AI, the humanoid-robotics company, has raised roughly $1.9 billion total and was valued at $39 billion after a September 2025 Series C, with its Figure 03 robot named a TIME Best Invention of 2025. Physical Intelligence, a robot-foundation-model startup, has raised $1.1 billion at a $5.6 billion valuation and was, as of early 2026, in talks to raise close to $1 billion more at a valuation above $11 billion. Skild AI raised $1.4 billion in January 2026 alone, one of the largest single robotics funding rounds ever recorded, pushing its total past $2 billion. Even national governments are now competing for a piece of the category: South Korea announced a $10.3 billion national growth-fund push into physical AI in July. Against that backdrop, Prometheus’s $18 billion raised in under eight months isn’t just large in absolute terms, it’s larger than any single competitor’s entire funding history to date, in a category multiple serious capital pools are now racing into simultaneously.
Prometheus’s rise is part of a broader capital rotation industry watchers have flagged through 2026: money moving beyond foundation-model chatbots and toward AI applied directly to the physical world, engineering, materials, manufacturing, and other high-stakes, regulated workflows where a wrong output has real-world consequences beyond a bad chat response. It’s a different bet than the one Anthropic, OpenAI, and Google are making with general-purpose language models, and Prometheus’s valuation trajectory, a more than sixfold jump in roughly seven months between rounds, suggests investors are pricing physical AI as a genuinely distinct, still largely uncontested category rather than a subset of the existing frontier-model race.
Bezos has also been unusually forthcoming about the company’s ambitions relative to how guarded frontier AI labs typically are about their roadmaps. “We’re not being secretive,” he said following the raise, a deliberate contrast to the more closed posture Anthropic and OpenAI tend to take publicly, and a signal that Prometheus is treating openness about its engineering goals as part of its pitch to both investors and future industrial customers.
Manufacturing and industrial-design AI tooling is a genuinely underexplored category for Philippine startups relative to the far more crowded consumer AI and fintech AI space most local founders are building in. The Philippines has real manufacturing and engineering-services capacity, electronics assembly, shipbuilding-adjacent industries, structural and civil engineering outsourcing, that could plausibly become a downstream customer or service layer for tools in the category Prometheus, Figure, Physical Intelligence, and Skild AI are all now racing to build, the same way Philippine BPO absorbed customer-service outsourcing over the past two decades. It’s early, and none of these companies have a visible presence in the Philippines yet, but the direction of Bezos-scale and now nation-state-scale capital toward physical and industrial AI is a category shift worth Philippine engineering and manufacturing sectors actually tracking rather than dismissing as a Silicon Valley curiosity with no local relevance.
What’s notably absent from the coverage so far is any disclosed revenue figure, for Prometheus or for most of its physical AI peers. Eighteen billion dollars raised in under eight months, with no public revenue number attached, is a bet on Prometheus’s category and its founders far more than it is a bet on a proven product, and that’s worth remembering the next time a Prometheus valuation headline circulates without that context attached.
Share this article