A major venture capital firm just dropped $1.3 billion on a new mission: bringing artificial intelligence out of our phones and laptops and into the real world. Eclipse, the Palo Alto-based investment powerhouse, is banking on what everyone’s calling “physical AI” , and they’re not just writing checks to existing companies. They’re actually building startups from scratch.

If you’ve been paying attention to tech investments over the past few years, you’ve probably noticed a shift. Eclipse’s portfolio tells the story perfectly. The firm has been quietly backing companies that deal with the tangible, physical world: electric boats through Arc, battery recycling via Redwood Materials, self-driving construction robots at Bedrock Robotics, and autonomous vehicle tech through Wayve. These aren’t apps or software platforms , they’re companies actually moving things around and solving problems in the real world.

According to Eclipse partner Jiten Behl, we’re entering a completely new era of innovation. For the past two decades, we’ve seen waves of internet, mobile, cloud, and social media breakthroughs. But this time is different. “We’re going to see advanced levels of intelligence, along with actual actions, in terms of solving problems in the real, physical world”, Behl explained. The convergence of available talent, technological improvements, market demand, policy support, and capital is making this moment possible.

What makes Eclipse’s strategy particularly interesting is how they’re approaching the ecosystem. Rather than just throwing money at individual startups, they’re thinking about how companies across different sectors , transportation, energy, infrastructure, computing, and defense , can work together and become each other’s customers as they scale. It’s about building interconnected partnerships that create stronger, more resilient businesses.

Eclipse is also taking an unconventional approach by actually incubating some of these companies internally. Behl confirmed they’re already working on “a couple of really cool ideas”, though he kept the details under wraps. The focus seems to be on startups that operate across multiple industries and can connect different sectors together.

The real insight, Behl suggests, is figuring out how to use data across different sectors to train smarter AI models that benefit everyone. This cross-sector data sharing could create competitive advantages , or what investors call a “moat” , that protects these companies from competition.

With $591 million dedicated to early-stage incubation and the rest for growth-stage companies, Eclipse is making a serious commitment to this vision. Whether physical AI becomes the transformative force they’re betting on remains to be seen, but with this much capital backing the movement, it’s hard to ignore.