- Prometheus AI has raised $12 billion at a $41 billion valuation, backed by JPMorgan Chase, Goldman Sachs, and BlackRock.
- Prometheus AI is building an ‘artificial general engineer’ — software designed to automate the design and manufacture of complex physical systems.
- Co-founded by Jeff Bezos and former Verily co-founder Vik Bajaj, the company has just 150 employees across three cities.
- Bezos argues AI will create ‘labor scarcity,’ not mass unemployment — a view that puts him at odds with many in the industry.
- Prometheus AI has raised $12 billion at a $41 billion valuation, backed by JPMorgan Chase, Goldman Sachs, and BlackRock.
- Prometheus AI is building an ‘artificial general engineer’ — software designed to automate the design and manufacture of complex physical systems.
- Co-founded by Jeff Bezos and former Verily co-founder Vik Bajaj, the company has just 150 employees across three cities.
- Bezos argues AI will create ‘labor scarcity,’ not mass unemployment — a view that puts him at odds with many in the industry.
Table of Contents
Prometheus AI Just Raised More Money Than Most Countries’ Space Budgets
Prometheus AI, the physical AI startup co-founded by Jeff Bezos and Vik Bajaj, has closed a $12 billion funding round at a staggering $41 billion valuation — making it one of the most richly valued AI startups ever to come out of stealth. The round was backed by some of the heaviest names in global finance: JPMorgan Chase, Goldman Sachs, and BlackRock, alongside Bezos himself. Combined with the company’s initial raise of $6.2 billion when it launched late last year, Prometheus has now pulled in more than $18 billion before shipping a single public product.
To put that in perspective: OpenAI, which had a working product and millions of users before it raised at similar valuations, took years to get there. Prometheus AI has done it in months, with 150 employees and offices in San Francisco, London, and Zurich. The sheer scale of investor confidence here is either a sign of extraordinary technology or extraordinary hype — and right now, we can’t fully tell which.
What Is an ‘Artificial General Engineer,’ Exactly?
The company’s stated mission is to build what it calls an ‘artificial general engineer’ — software that can automate the design and manufacturing of complex physical systems. Think jet engines. Drug compounds. Aerospace components. The kinds of things that take teams of highly trained engineers years to design, test, and iterate on. Prometheus AI wants to compress that process with AI.
It’s an audacious framing. The term deliberately echoes ‘artificial general intelligence,’ which has become the holy grail — and the most contested phrase — in AI research. By positioning its goal as ‘AGE’ rather than ‘AGI,’ Prometheus AI is making a pointed claim: that the harder, more defensible frontier isn’t in language models or reasoning engines, but in systems that can interact with and design for the physical world.
The physical AI category has been gaining serious momentum. Unlike pure software, physical systems have to contend with real-world constraints — tolerances, materials, regulatory requirements, supply chains. The argument from investors is that this creates moats that code alone can’t replicate. A language model can be fine-tuned and replicated relatively quickly. A system that genuinely understands how to design a turbine blade or optimize a drug delivery mechanism is much harder to commoditize.
That said, Prometheus AI hasn’t shown its cards yet. The company is keeping the specifics of what it’s already built tightly under wraps, which is unusual for a startup raising at this scale. Most companies at the $41 billion mark have at least a flagship product, a published research paper, or a set of early enterprise customers to point to. Prometheus AI has none of that publicly. What it does have is Bezos’s credibility, Bajaj’s deep background in life sciences AI at Verily, and apparently the confidence of some of the world’s largest financial institutions.
Bezos’s Bet on ‘Labor Scarcity’
Perhaps the most interesting thing about Prometheus AI isn’t the valuation or the funding — it’s what Bezos said about what happens when this technology works.
While much of the AI industry is having an uncomfortable conversation about automation and job displacement, Bezos is making a contrarian call. He told CNBC that AI-driven productivity will lead to what he calls ‘labor scarcity’ — a world where the demand for human workers actually outpaces supply. In his framing, this is a good thing.
‘Significant productivity in the economy is going to raise the standard of living,’ Bezos said. ‘People who today have two-earner households, they’ll become one-earner households. Maybe some people who are working overtime will stop working overtime.’
It’s an optimistic read — and one that’s worth scrutinizing, especially from someone in his position. Amazon, where Bezos serves as executive chairman and remains the largest individual shareholder, employs more than 1.5 million people worldwide. Under CEO Andy Jassy, the company has laid off tens of thousands over the past year while aggressively accelerating automation across its warehouse and logistics operations. The gap between Bezos’s macro-level optimism and Amazon’s operational reality is hard to ignore.
That doesn’t mean he’s wrong. Economists have long debated the net employment effects of automation, and history does offer some support for the view that productivity gains create new categories of work. But the transition costs are real, they fall unevenly, and they don’t always resolve on a comfortable timeline. Bezos’s ‘labor scarcity’ thesis assumes a relatively smooth reallocation of human effort — which is a significant assumption when you’re building software designed to replace large swaths of engineering work. Whether Prometheus AI can deliver on that promise remains the central question.
Why Physical AI Is the Sector Everyone Wants a Piece Of
Prometheus AI isn’t alone in attracting massive capital. Physical AI — broadly defined as AI systems that interact with, design for, or operate in the physical world — has become one of the hottest investment categories of 2025 and 2026. From robotics companies like Figure and Physical Intelligence to industrial AI platforms targeting manufacturing and energy, venture capital is flowing in at a pace that would have seemed implausible two years ago.
The thesis is consistent across most of these investments: software AI has become crowded and commoditized faster than anyone expected. The big foundation model race is largely being run by well-capitalized incumbents — OpenAI, Google DeepMind, Anthropic, Meta. Startups looking to carve out defensible territory are increasingly turning to domains where the data is harder to acquire, the domain expertise is harder to replicate, and the deployment environment is fundamentally more complex.
Engineering design fits that description almost perfectly. The training data for a system that can design jet engines or pharmaceutical compounds isn’t scraped from the internet — it lives inside proprietary CAD files, simulation outputs, manufacturing records, and decades of institutional knowledge. Getting access to that data requires partnerships, trust, and deep domain expertise. That’s a meaningful barrier to entry, and it helps explain why investors are willing to bet on Prometheus AI at this scale before seeing a finished product.
What $12 Billion Actually Buys You
Bezos indicated that a significant portion of the new capital will go toward compute — which, for a company building AI systems that model and simulate physical systems at this level of complexity, is entirely plausible. Physics-based simulation is computationally brutal. Training models that understand the mechanical, chemical, and thermodynamic properties of real-world materials requires orders of magnitude more compute than training a chatbot.
The staffing picture is also worth noting. At 150 employees, Prometheus AI is extraordinarily lean for a company at this valuation. That either means it’s planning to scale headcount aggressively with the new funds, or it’s building in a deliberately capital-efficient, small-team model — which would be an unusual choice given the infrastructure ambitions. Compute, talent, and proprietary data partnerships are all expensive, and $12 billion gives the company room to pursue all three without compromise.
What Prometheus AI does next will be closely watched. The company can’t stay in stealth indefinitely at this scale — at some point, the product has to emerge. When it does, it’ll either validate one of the boldest bets in recent tech history, or force a very public reckoning with what $18 billion in private capital actually built.
Source: TechCrunch
Frequently Asked Questions
What exactly is Prometheus AI building?
Prometheus AI is developing what it calls an ‘artificial general engineer’ — software capable of automating the design and manufacturing of complex physical systems, ranging from jet engines to pharmaceutical compounds. The company hasn’t disclosed specific products yet, keeping most of its technical work under wraps.
Who are the investors behind Prometheus AI?
The $12 billion round was backed by Jeff Bezos himself, alongside major financial institutions including JPMorgan Chase, Goldman Sachs, and BlackRock. This follows an initial raise of $6.2 billion when the company launched late last year.
Does Jeff Bezos think Prometheus AI will cause job losses?
Bezos argues the opposite — he believes AI-driven productivity will create ‘labor scarcity,’ a world where demand for human workers outpaces supply. He suggests people may shift to single-income households or work less overtime, rather than face mass unemployment.
Why are investors so interested in physical AI right now?
Investors argue physical AI is more defensible than pure software because the physical world creates competitive moats that code alone can’t replicate. Venture capital has been pouring into the sector, with Prometheus AI representing one of the largest single bets on the space to date.
