Bezos's Prometheus Closes $10B at $38B Valuation — Physical AI Gets Its Mega-Round

$38 billion

That's the post-money valuation. For a company with 120 employees, no shipped product, and five months of existence. The market isn't pricing Prometheus — it's pricing the category.

Project Prometheus, co-led by Jeff Bezos and Vik Bajaj, closed a $10 billion round this week, bringing cumulative funding north of $16 billion. The post-money valuation: $38 billion. JPMorgan and BlackRock participated. There was no lead investor.

That last detail matters. At this scale, rounds typically have a single lead — a SoftBank, an a16z — who anchors the deal and sets terms. Prometheus pulled in institutional capital without that structure. That's partly a function of Bezos's gravitational pull, and partly a signal that no single investor wanted to underwrite the full risk of an unproven category at a $38 billion price tag.

Prometheus has assembled 120+ researchers and engineers from OpenAI, DeepMind, Meta AI, and xAI across offices in San Francisco, London, and Zurich.

What Prometheus is building

Physical AI. In one sentence: foundation models trained on real-world physics data instead of text and images.

The AI models we use today — GPT-5, Claude, Gemini — understand language, generate images, and write code. They're trained on internet-scale text and image data. But picking up an object, navigating a room, or planning a physical task? That requires a fundamentally different kind of understanding.

Prometheus is betting that the same scaling laws that turned language models from curiosities into platforms will apply to physical-world data. Feed a model enough sensor data — camera feeds, depth maps, force-torque readings, inertial measurements — and it develops a general understanding of physics that transfers across tasks.

Vik Bajaj, the co-CEO, holds a PhD in physical chemistry from MIT and ran robotics projects at Google X. His framing in an eWeek interview cuts to the core: "Software AI understands language. Physical AI understands the world."

The challenge is data. The internet gave us trillions of tokens of text for free. Physical-world data — robot interaction logs, 3D point clouds, tactile sensor streams — is expensive and slow to collect. That's a key reason the funding number is so large: building the data collection infrastructure for physical AI requires capital on a different scale than training another LLM.

Timeline: launch to close

Date	Event
November 2025	Project Prometheus launches with $6.2B initial funding
February 2026	Reports of Bezos planning $100B AI industrial holding company
April 21, 2026	Financial Times reports $10B round under negotiation
April 23, 2026	Round closes. $38B post-money confirmed

Five months from launch to $16 billion cumulative funding. OpenAI took seven years to reach comparable capitalization. The comparison isn't entirely fair — OpenAI started as a nonprofit, and Prometheus launched as a commercial venture backed by the world's second-richest person. But the velocity is unprecedented regardless of context.

The $100B holdco

The Prometheus round is one piece of a larger Bezos play. In February, Inc reported that Bezos is planning a $100 billion AI industrial holding company with Prometheus at its center — vertically integrating AI chip design, data center infrastructure, robotics hardware, and energy supply.

Think of it as a second Amazon, funded by personal wealth rather than public markets. The structural parallel to Elon Musk's Tesla-xAI-Neuralink-SpaceX portfolio is obvious. Bezos is building the full stack of physical AI — from silicon to kilowatts — under a single ownership structure.

This reframes the $38 billion valuation. You're not pricing a model company. You're pricing the anchor tenant of a vertically integrated AI industrial complex.

The competitive field

Prometheus isn't alone in physical AI, but it dwarfs the field on capital.

Company	Recent funding	Valuation	Approach
Prometheus	$10B (Apr 2026)	$38B	General physical AI foundation model
Physical Intelligence	$400M (Nov 2024)	$2.4B	Robot generalist policy model (pi0)
Skild AI	$300M (Jul 2024)	$1.5B	Scalable robot foundation model
Sereact	$110M (Mar 2026)	N/A	Industrial logistics robot AI
Boston Dynamics Atlas	Internal (Hyundai)	N/A	Humanoid robot + AI integration
Tesla Optimus	Internal	N/A	General-purpose humanoid robot

Prometheus has more funding than the rest combined. But capital doesn't equal capability. Physical Intelligence's pi0 model has public demos and papers. Skild AI's team includes CMU robotics pioneers. Boston Dynamics has three decades of hardware experience. Tesla collects real-world physics data from millions of vehicles every day.

Prometheus's advantages are twofold: unprecedented capital for data infrastructure buildout, and a talent concentration of 120+ people with foundation model scaling experience from the four labs that defined the current generation of AI.

The physical AI foundation model race: Prometheus leads on capital, but hasn't shipped a product yet.

Why JPMorgan and BlackRock matter

The participation of JPMorgan and BlackRock is the most significant signal in this round. These aren't venture capital firms. They're the world's largest bank and the world's largest asset manager.

Their presence means two things.

First, physical AI is becoming an investable asset class. Until now, AI investing was concentrated in software AI — LLMs, generative models, SaaS copilots. BlackRock putting billions into a physical AI startup is a declaration that this sector has graduated from "interesting research" to "institutional-grade investment thesis."

Second, the Bezos network is doing real work. Jamie Dimon at JPMorgan and Larry Fink at BlackRock have relationships with Bezos spanning decades. This round was enabled by relational capital as much as technical conviction. For founders without that rolodex, the implication is sobering — replicating this funding trajectory without a comparably connected backer would be nearly impossible.

Cross-references

The software AI commoditization wave amplifies the physical AI thesis. DeepSeek just cut inference prices by 75%, accelerating the compression of software AI margins. As text and code generation become commodity services, differentiation migrates toward hardware and the physical world. Prometheus's timing isn't coincidental — the cheaper software AI gets, the more valuable physical AI becomes by comparison.

Avoca's $1B valuation shows the digital-physical interface in action. Avoca's Series B funds AI voice agents that connect digital interactions (phone calls, scheduling) to physical services (HVAC repair, plumbing). That's the junction point between software and physical AI. Prometheus aims to own the "physical" side of that junction with a foundation model.

BlackRock's broader AI infrastructure strategy provides context. BlackRock has been investing billions in AI infrastructure — data centers, energy — since 2025. The Prometheus investment extends that thesis from infrastructure into the model layer itself. Traditional finance is reclassifying AI from "tech sector subcategory" to "infrastructure asset class."

Why it matters — by persona

If you're a startup founder: When physical AI foundation models ship, the vertical application layer on top of them will explode — logistics optimization, robot control, autonomous navigation. It's too late to compete on the infrastructure layer. It's too early to build on it. The optimal move is to develop domain expertise and proprietary data in a physical-world vertical so you're ready when the platform arrives.

If you're an investor: A $38B valuation on zero revenue needs a TAM in the tens of trillions to justify the math. Physical AI TAM estimates vary by 10x or more across analysts. This is a classic high-risk, high-convexity bet. JPMorgan and BlackRock passing due diligence is a signal, not a guarantee.

If you're an engineer: The physical AI talent war is on. Robotics + ML cross-functional skill sets are in acute shortage. Prometheus, Physical Intelligence, and Skild AI are all recruiting from the same pool. If you have relevant skills, your market value is spiking.

Stakes

Wins: Bezos secures first-mover positioning for the physical AI platform era. JPMorgan and BlackRock get direct exposure to the model layer of AI. Physical AI researchers and engineers see compensation packages surge.

Loses: Competing physical AI startups face a tighter funding environment as Prometheus absorbs market attention and LP capital. Software-only AI startups lose some narrative momentum — "we have a physical AI vision too" now has to compete with a $38B incumbent.

Watching: NVIDIA (potential compute partner), Hyundai/Boston Dynamics (hardware collaboration vs. competition), Google DeepMind (recalibrating physical AI strategy post-RT-2), and China's robotics ecosystem (likely to mount a state-backed response).

The skeptic's case

Daniela Rus, director of MIT CSAIL, cautioned in an AI Magazine interview: "The idea of a general physical AI model is appealing, but the physical world is far more diverse and unpredictable than text. The assumption that LLM scaling laws transfer directly to robotics is unvalidated. $16 billion doesn't make physics easier."

Marc Raibert, Boston Dynamics founder, offered indirect criticism: "After 30 years in robotics, the most dangerous belief in this field is that enough money and data will solve everything. The real world has an infinitely long tail." He didn't name Prometheus, but the target was clear.

And then there's the valuation question. Chris Metinko at Crunchbase News pointed out: "$38 billion is one of the highest valuations in history for a company with no revenue. Even OpenAI had billions in revenue when it reached this level. How much Bezos premium can the market sustain?"

Tomorrow morning

If you're in robotics or ML, check the Prometheus careers page. They're hiring aggressively in SF, London, and Zurich, with RSU packages that reportedly exceed market rates for the robotics-ML intersection.

If you're a software AI startup founder, prepare for the investor question: "What's your positioning in the physical AI era?" It's coming in your next board meeting.

If you're an investor, track the follow-on funding conditions for Physical Intelligence, Skild AI, and Sereact in the weeks ahead. How their next rounds price after the Prometheus mega-round will reveal the market's real temperature.

One-liner

Bezos didn't just fund a startup — he bankrolled the narrative that physical AI is the second act of the foundation model era.

---

References

1. Bloomberg — Bezos's Prometheus Closes $10B Round
2. TechFundingNews — Prometheus Funding Details
3. AI Magazine — Physical AI Race Heats Up
4. eWeek — Prometheus Team and Vision
5. Inc — Bezos $100B AI Holdco Plans
6. Financial Times — Prometheus Round Reports