Jeff Bezos is seeking $100 billion for a buyout fund targeting traditional manufacturing companies—chipmakers, defense contractors, aerospace firms, and automakers—with plans to transform them using AI. The fund would rival SoftBank’s Vision Fund as one of the world’s largest investment vehicles, though that comparison comes with a warning: Vision Fund lost $32 billion in a single fiscal year.
This ties to Project Prometheus, Bezos’ AI startup that launched in November 2025 with $6.2 billion. While the startup builds AI technology, the fund would acquire companies and deploy it at scale—vertical integration on steroids.
What Is Project Prometheus?
Project Prometheus is Bezos’ physical AI startup, co-founded with Vik Bajaj, a Google X veteran with a PhD in physical chemistry from MIT. The company focuses on AI models that understand and optimize the real world—manufacturing lines, prototypes, aerospace engineering—not just text and images like ChatGPT.
As of December 2025, Prometheus employed over 120 people, including researchers from OpenAI, Google DeepMind, and Meta. It operates out of San Francisco with offices in London and Zurich, targeting aerospace, automotive, chipmaking, and computing hardware.
Nvidia CEO Jensen Huang called 2026 the “ChatGPT moment for physical AI,” signaling the technology’s shift from R&D to commercial deployment.
The $100 Billion Fund Strategy
Bezos is meeting with asset managers in Singapore and the Middle East to raise what investor documents describe as a “manufacturing transformation vehicle.” The fund would acquire companies in chipmaking, defense, aerospace, and automotive, then modernize them using Prometheus’ AI.
The strategy mirrors Amazon’s playbook: identify inefficiencies in legacy industries, apply technology at scale, dominate. Manufacturing’s bottleneck isn’t assembly lines—it’s pre-production. Prototyping and design iteration take months. AI compresses that to days.
According to Deloitte research, 58% of manufacturers use physical AI, with 80% planning adoption within two years. Generative design reduces physical prototypes by 30-50%, digital twins cut development time by 50%, and ABB reports 80% reductions in setup times with 40% cost savings.
At $100 billion, this fund would rival SoftBank’s Vision Fund, which raised $100 billion in 2017—roughly 4-5 times larger than typical buyout funds. That scale came with problems: Vision Fund lost $32 billion in fiscal year 2023.
Physical AI Is Different
The previous AI wave disrupted software: language models, image generation, code completion. Physical AI targets harder problems—understanding physics, materials science, mechanical systems. It’s not enough to predict the next token; you have to model real-world constraints.
AI-powered generative design produces hundreds of optimized variations in minutes versus weeks manually. Digital twins simulate production lines, catching defects before they hit the factory floor. Companies like Audi and BMW are piloting humanoid robots as physical AI moves mainstream.
Prometheus isn’t building LLMs. It’s building AI that optimizes jet engines, chip fabs, and assembly lines—fundamentally different from generating text.
The Vision Fund Problem
SoftBank’s Vision Fund raised $100 billion with backing from Saudi Arabia’s Public Investment Fund ($45 billion) and SoftBank ($28 billion). The thesis: deploy unprecedented capital into tech companies and reshape industries. It lost $32 billion in one fiscal year, weighed down by WeWork and overvalued unicorns.
Bezos is betting manufacturing transformation is different. He turned Amazon from a bookstore into a $1.5 trillion company by dismantling retail and logistics inefficiencies. Bajaj brings Google X moonshot expertise. Physical AI is reaching commercialization with documented 20-40% productivity gains.
But $100 billion is $100 billion. Manufacturing transformation faces regulatory challenges in defense and aerospace, entrenched processes, and competition from China’s manufacturing AI push.
What’s Next
Watch for fund closes and acquisition announcements. If Bezos secures the $100 billion, first targets will signal whether this is disciplined transformation or another mega-fund miscalculation. The physical AI inflection point is real—2026 is when manufacturers move from pilots to production. Whether Bezos can deploy $100 billion effectively remains open.
