Eclipse raises $1.3B for physical AI

A venture firm just raised more money for robots, factories, and power systems than many software funds raise for apps: Eclipse closed $1.3 billion on April 8, with $720 million in Fund VI and $591 million in Early Growth Fund III. The split matters because one pool writes early checks and the other keeps funding companies as they get expensive to build. (thenextweb.com) Eclipse is not betting on chatbots sitting on a screen. It is betting on “physical AI,” meaning software that has to drive a vehicle, run a warehouse arm, manage a battery system, or coordinate equipment in a factory. (techcrunch.com) That kind of company burns money in a different way from a software startup. A warehouse robot needs motors, sensors, safety testing, and customers willing to change how a building works, which is why hard-tech founders usually need more capital and more time. (bloomberg.com) Eclipse has been building toward this for years. The firm said in its 2023 review that it had already closed $1.2 billion across two funds that year, so the new $1.3 billion raise is a step up rather than a sudden pivot. (eclipse.vc) The portfolio shows what Eclipse means by the physical world. Bloomberg identified Wayve in self-driving software, Redwood Materials in battery supply chains, and Cerebras in artificial intelligence chips as examples of the firm’s existing bets. (bloomberg.com) Those companies all sit on the same chain of needs. Cerebras sells the computing hardware that trains models, Wayve uses large amounts of computing for autonomous driving, and Redwood builds energy storage systems and critical minerals supply that help power data centers and the electric grid. (redwoodmaterials.com) Eclipse is also doing more than writing checks from an office. Its Venture Equity program says it works with founders and executives to explore new business ideas using Eclipse’s customer network, talent network, and early-stage funds, which is closer to company-building than passive investing. (eclipse.vc) That matters because physical-world startups often fail before the product is finished, not because the idea is bad, but because the gap between prototype and commercial rollout is brutally expensive. A dual-fund setup gives Eclipse a way to back a company at the sketch stage and still fund it when it needs a factory line, fleet deployment, or industrial sales team. (thenextweb.com) The timing also says something about where venture money thinks the next artificial intelligence wave is going. After two years of spending on models and cloud computing, investors are still willing to fund the messier layer where software has to meet steel, concrete, batteries, and labor shortages. (techcrunch.com) Eclipse said the new raise brings assets under management to about $10 billion. For founders building robots, energy systems, and industrial tools, that means one more large specialist fund is signaling that the next artificial intelligence boom may be judged not just by model benchmarks, but by whether machines actually do useful work in the real world. (thenextweb.com)