Many VCs have been talking about "Physical AI" lately, but for most it remains a PowerPoint-level concept — they mention robotics and autonomous driving, then go back to funding software companies and call it a day.
Eclipse Capital has done something different. In early April, the Palo Alto-based VC announced it had closed $1.3 billion in fundraising: a $591 million early-stage incubation fund and a growth-stage fund.
The amount isn't the largest in the industry, but the investment thesis is worth a close look.
Not Just Investing, But Incubating
Partner Jiten Behl described the fundraising round bluntly:
"Over the past two decades we've watched many waves of technological innovation… This is the first time things are moving from the screen into the physical world."
Eclipse's approach differs from typical VCs — it not only invests in external companies but also incubates its own startups. Several projects are already "cooking" internally, which Behl described as "a few really cool ideas," primarily enterprise-facing products that can span multiple industries.
More distinctive is the data strategy. Eclipse deliberately structures its portfolio as an ecosystem, pairing autonomous driving companies with energy firms, and industrial robotics companies with construction firms to share data, training better AI models with cross-industry data:
"The key is how to connect across industries, how to build scale across industries, and then use cross-industry data to build a moat."
Existing Portfolio Companies
Eclipse's portfolio spans several areas:
- Arc: Electric boats
- Redwood Materials: Battery recycling and materials
- Bedrock Robotics: Autonomous construction machinery
- Wayve: Autonomous driving technology
- Mind Robotics: Industrial AI robotics
What these companies have in common: they all operate on physical objects in the real world, all rely on massive sensor data, and all have opportunities to complement each other on the data dimension.
Why Now
The software side of AI has already reached saturation. A few major players dominate large models, code tools are consolidating, and big tech has entered the infrastructure layer.
The physical side is different: challenges are greater, barriers are higher, data is scarcer, and big tech cannot replicate it quickly. True competitive advantage comes from accumulated sensor data, engineering experience, and industry relationships — not from compute power alone.
Eclipse's logic: use $1.3 billion to build data moats at an early stage, so that by the time others realize the Physical AI opportunity, the trenches are already dug.
An interesting observation is that the red ocean of software AI is driving smart money toward the physical world. This inflection point is becoming increasingly clear in 2026.
Source: CocoLoop, VC Eclipse has a new $1.3B to back — and build — 'physical AI' startups (TechCrunch)