Jeff Bezos launches AI start-up “Project Prometheus,” will serve as co-chief executive

Amazon founder Jeff Bezos is stepping back into an operational leadership role for the first time since 2021, unveiling an ambitious new artificial intelligence venture called Project Prometheus. Backed with a staggering $6.2 billion in initial funding, the start-up enters the world as one of the most heavily financed early-stage AI companies globally.

According to individuals familiar with the project, the company is developing advanced AI systems designed to enhance engineering and manufacturing processes across a wide range of industries—including computing hardware, aerospace, automotive production, and spacecraft engineering.

Bezos reenters the spotlight with a hands-on leadership role

Since stepping down as Amazon’s CEO, Bezos has devoted substantial energy to his private space company Blue Origin, while also drawing public attention for his personal life and high-profile events. However, Project Prometheus marks a significant shift: it is his first formal leadership role at a new company since leaving Amazon.

The initiative also signals Bezos’ deepening interest in the global race to build next-generation AI—an increasingly competitive field dominated by giants like Google, Meta, Microsoft, and leaders such as OpenAI and Anthropic.

An AI Vision Aligned with Space and Advanced Manufacturing

Though many details about the company remain under wraps—including its launch date and physical headquarters—Project Prometheus is designed to advance technologies that align with Bezos’ long-standing fascination with space exploration.

The start-up’s AI research will focus on enhancing real-world engineering tasks, improving manufacturing systems, and accelerating development cycles across physical industries.

A high-profile leadership team

Bezos will co-lead Project Prometheus alongside Vik Bajaj, a physicist and chemist with an extensive track record in deep-tech innovation. Bajaj previously worked at Google X, the company’s famed “Moonshot Factory,” where he collaborated closely with Google co-founder Sergey Brin.

His career includes co-founding:

• Verily (2015), Alphabet’s life-sciences research organization
  
• Foresite Labs (2018), an incubator for AI-driven and data-centric start-ups
  

Bajaj recently left his role at Foresite Labs to focus entirely on Project Prometheus.

Part of a broader wave of AI-for-science start-ups

Project Prometheus joins a growing movement of companies developing artificial intelligence tailored to physical tasks—fields where AI is expected to revolutionize productivity. Start-ups such as Periodic Labs are exploring AI for scientific discovery, physics research, and drug design, drawing talent from major AI research labs.

Interest in this space is accelerating. Bezos himself invested last year in Physical Intelligence, a company focused on AI-powered robotics. Other entrants, like Thinking Machines Lab—founded by former OpenAI researchers—have raised billions to pursue similar goals.

A high-powered team and a competitive edge

Sources say Project Prometheus has already hired nearly 100 employees, including researchers recruited from leading AI organizations such as OpenAI, Google DeepMind, and Meta. This rapid expansion suggests the company is poised to compete aggressively in the race to build the next generation of AI systems.

Established labs are also pushing hard into AI-for-science initiatives. Google DeepMind researchers recently earned a Nobel Prize in Chemistry for their contributions to AlphaFold, a transformative tool for drug discovery. Meanwhile, OpenAI and Meta continue to push the boundaries of what large language models can achieve in mathematics and theoretical science.

Going beyond text: AI that learns from the physical world

While large language models learn by analyzing enormous quantities of digital text, Project Prometheus aims to build AI systems that can learn through experimentation, simulation, and interaction with the physical world.

Companies like Periodic Labs plan to use robotic laboratories to run experiments autonomously at massive scale, enabling AI to acquire hands-on scientific knowledge—a step beyond today’s text-driven models.

Project Prometheus is expected to explore similar approaches, developing new ways for AI to accelerate breakthroughs in engineering, manufacturing, and scientific research.