Eclipse raises $1.3 billion to back and incubate physical AI startups

Venture capital firm Eclipse has raised $1.3 billion across two new funds specifically targeting "physical AI" startups—companies developing AI applications in robotics, manufacturing, energy, and related hardware-intensive fields. According to a TechCrunch report, Eclipse plans to deploy the capital not only to back existing ventures but also to incubate and build new ones from the ground up. Eclipse founder and CEO Lior Susan discussed the funds on Bloomberg Technology, emphasizing their focus on the "next era of physical industries."^[1]

This move signals a broadening of the AI investment boom beyond software into tangible, real-world hardware. Physical AI refers to AI systems integrated with physical devices, such as robots or automated factories, which require substantial capital for prototyping and scaling. As Susan explained in the Bloomberg interview, the funds aim to connect startups in robotics, energy, and transport to accelerate practical AI solutions.^[3] Eclipse's strategy stands out for its hands-on approach, including direct incubation, which could speed up development in capital-heavy sectors where traditional software AI falls short.

The announcement comes amid a surge in funding for hardware AI plays. For instance, Eclipse previously led a $125 million Series B for ForSight Robotics, an Israeli startup using robotics for eye surgery, demonstrating its track record in the space.^[4] Meanwhile, other firms are chasing similar opportunities: Munich-based Conxai recently raised €5 million to deploy agentic AI—autonomous systems that handle complex tasks—for construction workflows, backed by investors like Earlybird and Pi Labs.^[3] In defense, Hermeus secured $350 million at a $1 billion valuation to develop autonomous hypersonic fighters, with its CEO noting the need to embrace hardware failures for rapid iteration.^[4]

Infrastructure demands are also heating up, underscoring why physical AI matters. Nvidia-backed Firmus, an Australian AI data center builder, hit a $5.5 billion valuation after raising $505 million in equity plus a $10 billion Blackstone-led debt facility, eyeing a $2 billion ASX IPO soon.^[2][6] These massive builds are essential for training the AI models powering physical systems, affecting global supply chains from data centers to factories.

The trend reflects wealthy investors diving deeper into AI's riskier frontiers. Family offices, once passive through VCs, are now making direct, early-stage bets on AI startups, as discussed on TechCrunch's Equity podcast with Arena Private Wealth.^[5] Broader conversations at events like the HumanX conference highlight nonobvious AI winners in hardware, per FPV's Wesley Chan on Bloomberg Tech.^[1]

Who stands to gain or lose? Startups in robotics and manufacturing could see accelerated growth, creating jobs and innovations in energy-efficient production or advanced transport. Investors benefit from diversification beyond pure software plays, but the high failure rates in hardware—due to engineering challenges—pose risks. Regions like Australia, Israel, and Europe are emerging as hubs, potentially shifting power from Silicon Valley.

Looking ahead, Eclipse's funds could catalyze partnerships across sectors, with Firmus's IPO and Hermeus's next aircraft demonstrator as early indicators of momentum.^[2][4] As AI moves from digital to physical realms, this $1.3 billion infusion positions Eclipse at the forefront, potentially reshaping industries reliant on intelligent machines.