Remember when venture capital was all about apps that helped you order lunch or share photos? Those days feel like ancient history now. Eclipse Ventures just closed $1.3 billion across two funds, and they’re not interested in your next social media platform. They’re backing AI that moves, builds, and operates in the physical world.
This marks the firm’s largest raise to date, and the timing tells us something important about where smart money thinks AI is heading. Founded by Lior Susan, Eclipse has been quietly building a portfolio that includes names like Wayve and Cerebras—companies working on autonomous vehicles and AI computing hardware, respectively. These aren’t chatbots. These are systems that need to work in factories, on roads, and in environments where failure means more than a bad user experience.
Why Physical AI Matters for Agent Builders
If you’re building AI agents, you might wonder why a massive fund focused on robotics and hardware should matter to you. Here’s the connection: the infrastructure these physical AI companies are creating will become the foundation your agents run on. Cerebras isn’t just making chips for fun—they’re solving the computational bottlenecks that make real-time agent decision-making possible at scale.
Wayve’s work on autonomous driving is essentially solving the same problem you face when building agents: how do you make decisions in unpredictable environments with incomplete information? The difference is that their agents are navigating London traffic instead of navigating API calls, but the underlying challenges around perception, planning, and execution overlap significantly.
The Software-Hardware Convergence Nobody Talks About
We’ve spent the last decade treating AI as a purely software problem. Train a model, deploy it to the cloud, call it a day. But Eclipse’s $1.3 billion bet suggests that era is ending. The next generation of AI applications will need purpose-built hardware, edge computing capabilities, and systems designed to operate in the messy reality of the physical world.
For agent developers, this means your deployment targets are about to expand dramatically. Today, you might be building agents that live entirely in Slack or email. Tomorrow, those same agents might need to coordinate with robotic systems, manage physical inventory, or make decisions that affect real-world logistics chains.
What This Funding Round Signals
When a firm doubles down with its largest raise ever, it’s making a statement about market timing. Eclipse clearly believes we’re at an inflection point where AI is ready to leave the screen and enter the physical world at scale. The companies in their portfolio—Wayve, Cerebras, and Redwood Materials among them—represent different pieces of this puzzle.
Redwood Materials, for instance, works on battery recycling. That might seem unrelated to AI agents until you consider that autonomous systems need sustainable power sources, and managing complex recycling operations requires sophisticated decision-making systems. Everything connects.
Practical Implications for Agent Development
If you’re building AI agents today, start thinking about how your work might extend into physical systems. The APIs you’re calling might soon include robotic arms, autonomous vehicles, or manufacturing equipment. The monitoring and observability tools you build need to account for real-world latency and failure modes that don’t exist in pure software systems.
The $1.3 billion Eclipse just raised will fund companies that create the infrastructure layer for physical AI. As an agent builder, you’ll eventually build on top of that infrastructure. Understanding what’s being built now helps you anticipate what’s possible next year.
This funding round isn’t just about robots and self-driving cars. It’s about the maturation of AI from a technology that lives in data centers to one that operates in the world we inhabit. For those of us building agents, that means our scope of work is about to get a lot more interesting—and a lot more consequential.
đź•’ Published: