Eclipse Ventures has closed a $1.3 billion fund dedicated exclusively to "physical AI" — companies that combine artificial intelligence with hardware, robotics, and industrial systems. It is one of the largest single-thesis venture funds ever raised for the robotics and embodied AI sector, signalling that institutional capital is treating physical AI as a mature, investable category rather than a speculative niche.
Table of Contents
- What is Eclipse Ventures and what is Physical AI?
- How Eclipse Invests: Portfolio Companies and the Incubator Model
- Why $1.3B Is a Market Signal, Not Just a Fund Announcement
- What Physical AI Startups Could Emerge from the Eclipse Pipeline?
- What This Means for Robotics
- Frequently Asked Questions
What is Eclipse Ventures and what is Physical AI?
Eclipse Ventures is a Palo Alto-based VC firm that has spent the last decade backing companies at the intersection of software and physical systems — factories, supply chains, energy infrastructure, and robots. "Physical AI" is their term for AI that doesn't live in a data centre or chat interface, but instead acts on the physical world through sensors, actuators, and autonomous decision-making.
Think less ChatGPT, more a robotic arm that can identify a defective weld by sight and correct its own trajectory in real time. The term captures a genuinely distinct category: software intelligence fused with hardware constraints, latency demands, and the unforgiving physics of the real world.
Eclipse was founded in 2015 by former executives from Flex, one of the world's largest electronics manufacturing services companies. That operational DNA has always shaped their investment thesis — they aren't just writing cheques, they're backing founders who understand what it takes to build physical products at scale.
How Eclipse Invests: Portfolio Companies and the Incubator Model
Eclipse doesn't only write cheques — it also incubates and co-founds companies, which is what makes this fund structurally unusual among large VC raises.
A significant portion of the $1.3 billion will be directed toward building new startups from scratch, not just backing existing ones. Eclipse acts as a co-founder in these situations, providing capital, operational expertise, manufacturing relationships, and technical talent to spin up companies that might otherwise struggle to bridge the gap between software-native AI research and hardware-ready products.
Their existing portfolio already traces the arc of where this capital will likely land:
| Company | Focus Area | Relevance to Physical AI |
|---|---|---|
| Bright Machines | Factory automation software | AI-driven assembly and inspection |
| Gatik | Autonomous trucking (middle mile) | Embodied AI for logistics |
| Joby Aviation | eVTOL aircraft | Autonomous flight systems |
| Nuro | Autonomous delivery vehicles | AI + physical product delivery |
| Standard Bots | Industrial robot programming | LLM-enabled robot control |
| Machina Labs | AI-powered metal forming | Robotic manufacturing processes |
Standard Bots and Machina Labs are perhaps the clearest previews of the kind of company Eclipse wants to incubate next — businesses where the AI capability and the physical system are co-designed, rather than software bolted on top of legacy hardware. You can browse industrial robots and automation platforms similar to those these companies are disrupting.
Why $1.3B Is a Market Signal, Not Just a Fund Announcement
Fund size communicates conviction. A $1.3 billion single-thesis fund tells the market several things at once.
First, it tells LP (limited partner) investors — pension funds, endowments, family offices — that physical AI has sufficient deal flow, exit pathways, and return potential to justify concentration at this scale. LPs are notoriously conservative; getting them to commit to a focused hardware-AI fund of this size required Eclipse to demonstrate that the category has moved past early-stage risk.
Second, it signals competitive pressure on other major VC firms. Andreessen Horowitz has its American Dynamism fund. Khosla Ventures has long backed hard-tech. But a clean $1.3 billion dedicated to physical AI from a firm with Eclipse's manufacturing pedigree is a different kind of statement — it's a claim that this category is now large and mature enough to warrant a specialist investor at institutional scale.
Third, the timing matters. This raise lands as humanoid robot companies are securing nine-figure rounds (Figure AI raised $675 million, Physical Intelligence raised $400 million in late 2024), and as the US government is actively incentivising domestic manufacturing through the CHIPS Act and Inflation Reduction Act. Eclipse's fund is, in part, a bet that the policy environment will accelerate demand for physical AI in American factories.
What Physical AI Startups Could Emerge from the Eclipse Pipeline?
The incubation model gives the clearest preview of Eclipse's pipeline thesis — and it points toward several underserved problems in robotics and industrial AI.
Based on the firm's portfolio patterns and stated focus areas, the categories most likely to generate new Eclipse-incubated startups include:
Robot-native software stacks. Most industrial robots still run on proprietary, closed programming environments. Eclipse has already backed Standard Bots for its natural-language robot programming approach. Expect more startups targeting the software layer that sits between large language models and robot controllers — essentially, the middleware that translates AI intent into physical motion.
AI-enabled quality control and inspection. Computer vision applied to manufacturing defect detection is one of the highest-ROI applications of physical AI. It requires minimal mechanical hardware, has clear cost-reduction metrics, and is deployable without full factory automation. This makes it an ideal incubation target — high defensibility, short sales cycle, measurable output.
Autonomous materials handling. Machina Labs demonstrated that AI can replace traditional tooling in metal forming. The logical extension is AI systems that manage materials flow inside factories — not full humanoid robots, but purpose-built autonomous systems for specific material types and factory layouts.
Energy and grid-adjacent physical AI. Eclipse's broader portfolio has touched energy infrastructure. As AI compute demand drives power constraints, physical AI systems that optimise energy distribution, grid management, or industrial HVAC represent a natural expansion of the thesis.
If you're tracking the humanoid robot and cobot landscape these companies will eventually disrupt, the Botmarket cobot marketplace is a useful reference for current commercial deployments.
What This Means for Robotics
For the robotics industry, Eclipse's $1.3 billion fund is less a financial story than a structural one. It accelerates several trends simultaneously.
Founders get a co-builder, not just a cheque. The incubation model means robotics engineers and AI researchers who lack manufacturing or operations experience can partner with Eclipse early. This lowers the barrier to entry for deep-tech founders and should increase the volume of serious physical AI companies reaching Series A.
Corporate R&D faces tougher competition. When well-capitalised startups are co-founded by a firm with Flex-level manufacturing relationships, incumbent industrial companies — Siemens, ABB, Fanuc — face challengers that can move fast AND manufacture at scale. That's a genuinely new threat model.
Valuations in physical AI will rise. Large dedicated funds compete for the best deals. More capital chasing the same high-quality physical AI founders means valuations at seed and Series A will increase. For buyers and operators, this is largely neutral — but it signals that physical AI is entering a competitive, well-funded phase that will accelerate product development timelines.
The category name matters. "Physical AI" as a term is gaining institutional legitimacy. Eclipse using it as the explicit thesis of a $1.3 billion fund means the framing will appear in more pitch decks, job descriptions, and procurement conversations. Expect it to become the dominant label for embodied AI and robotics-adjacent AI over the next 24 months.