- Jeff Bezos is closing a $10 billion funding round for Project Prometheus, valuing the physical AI lab at $38 billion just five months after launch.
- BlackRock and JPMorgan are leading the round, bringing total funding past $16 billion since November 2025.
- Prometheus trains AI models on real-world experimental data, robotics interactions, and engineering workflows rather than text alone.
- Bezos is reportedly seeking up to $100 billion for a holding company to acquire industrial businesses whose operational data would feed Prometheus’s models.
When Jeff Bezos stepped down as Amazon’s CEO in 2021, most assumed he was heading into semi-retirement. Five years later, the world’s second-richest person is running his most ambitious venture yet. Project Prometheus, the physical AI lab he co-founded in San Francisco last November, is on the verge of closing a $10 billion funding round that would peg the company’s value at $38 billion. This is not another chatbot play. This is a bet that AI’s real future lies in understanding the physical world.
From Text to Touch: What Prometheus Actually Does
The Core Thesis
While most AI labs are racing to build better language models, Prometheus is pursuing something fundamentally different. The company trains AI systems on real-world experimental data, robotics interactions, and engineering workflows. Its models are designed to understand the laws of physics for use in industry, engineering, and manufacturing. Think of it as moving AI from the world of words into the world of atoms.
Target Industries
Prometheus is going after some of the largest sectors of the global economy: aerospace, automotive, advanced manufacturing, and drug discovery. These are industries where AI has long promised transformation but rarely delivered because the data is messy, physical, and expensive to collect. Bezos appears to believe that whoever solves the data problem in these verticals will own the next platform shift.
Business Insight — Physical AI represents a fundamentally different market from the LLM wars. While OpenAI, Anthropic, and Google compete on text and reasoning benchmarks, Prometheus is carving out a lane where the moat is not compute but proprietary physical-world data. Companies with deep manufacturing or engineering data should be watching this closely.
The Money Behind the Mission
$16 Billion in Five Months
The numbers are staggering even by AI-era standards. Prometheus launched in November 2025 with $6.2 billion in initial funding. The current $10 billion round, which is nearing completion, would push total capital raised past $16 billion. For context, that is more than Anthropic had raised in its first three years of existence. BlackRock and JPMorgan are confirmed as lead investors in the latest round, signaling that this is not just a tech play but a bet backed by the world’s largest asset managers.
The $100 Billion Holding Company Plan
Perhaps the most provocative detail in the Prometheus story is Bezos’s reported ambition to raise up to $100 billion to create a holding company. The plan: acquire industrial businesses affected by AI, then use their operational data to feed Prometheus’s models. It is a flywheel strategy that mirrors what Bezos built at Amazon, where customer data fed better recommendations, which drove more purchases, which generated more data. Except this time, the data is not shopping habits. It is how factories, laboratories, and supply chains actually work.
Business Insight — The holding company model could reshape M&A in traditional industries. If Bezos executes on a $100B acquisition strategy, mid-market manufacturers, pharma labs, and engineering firms become targets not for their revenue but for their operational data. Enterprise leaders should begin auditing the strategic value of their proprietary process data.
The Dream Team Behind Prometheus
Leadership and Talent
Bezos shares the co-CEO title with Vikram “Vik” Bajaj, a physical chemist with a PhD from MIT who previously led early-stage projects at Google X, including foundational work on Wing (drone delivery) and Waymo (autonomous vehicles). The combination is deliberate: Bezos brings capital, operational scale, and brand gravity, while Bajaj brings deep technical expertise in bridging scientific research with real-world deployment.
Aggressive Recruiting and Acquisitions
Prometheus has been aggressively poaching talent from the industry’s biggest names. The company has recruited researchers from OpenAI, xAI, Meta, and DeepMind. It also brought on Kyle Kosic, co-founder of xAI, as a key executive. On the acquisition front, Prometheus acquired General Agents, an agentic AI startup co-founded by former DeepMind researcher Sherjil Ozair. With offices in San Francisco, London, and Zurich, the company is building a global footprint from day one.
Business Insight — Prometheus’s talent strategy reveals where the AI industry’s center of gravity is shifting. By hiring researchers from OpenAI, xAI, and DeepMind specifically for physical-world AI work, the company is signaling that the frontier talent pool now sees industrial applications as more compelling than pure language model research. Hiring managers in enterprise AI should expect stiffer competition for robotics and physics-based ML talent.
What This Means for the AI Landscape
Project Prometheus is the clearest signal yet that the AI industry’s next chapter will not be written in chat windows. It will be written on factory floors, in pharmaceutical labs, and across global supply chains. Bezos is betting that text-based AI is becoming a commodity, and that the real value lies in understanding how the physical world works. With $16 billion in funding, a $38 billion valuation, and a plan to spend $100 billion more on acquiring industrial data, this is not a side project. It is the most capital-intensive AI startup launch in history, led by someone who has already built the world’s most successful data flywheel once before.
The competitive landscape is also notable. While Prometheus competes in the physical AI space with companies like Figure AI in robotics and Periodic Labs in materials science, no one else is combining an AI model company with an industrial holding company strategy at this scale. If Bezos succeeds, the playbook for AI dominance may shift from “build the best model” to “own the best data from the real world.”
Related
- OpenAI Bought 7 Companies in 90 Days
- $300B in One Quarter — And 80% Went to AI
- AI Spending vs. AI Results: The Value Gap Widens
- Meta Is Training AI on Its Own Employees
Sources
- TechFundingNews — BlackRock and JPMorgan back Bezos’ AI lab in a $10B raise at $38B valuation
- The Decoder — Jeff Bezos nears $10 billion funding round for AI lab Project Prometheus
- Bloomberg — Jeff Bezos Nears $10 Billion Funding for AI Lab, FT Says
AI Biz Insider — AI Business EN — aibizinsider.com
댓글 남기기