Simon Glairy is a visionary leader at the intersection of capital and physical innovation, currently steering a significant $1.3 billion investment initiative. As a partner at Eclipse, he specializes in identifying and building “physical AI” companies that bridge the gap between digital intelligence and real-world infrastructure. With a portfolio that spans from autonomous construction to battery recycling, his expertise lies in fostering a connected ecosystem where data and hardware converge to solve the world’s most pressing industrial challenges.
You are managing a $1.3 billion split between early-stage incubation and growth-oriented capital. What specific market shifts indicate that innovation is moving from digital screens to the physical world, and how does your investment strategy change when supporting a startup’s entire lifecycle?
The shift we are witnessing is a fundamental transition from the era of mobile cloud and social media to an age where advanced intelligence translates into tangible action. For the last two decades, innovation was largely confined to our screens, but we are now seeing a confluence of talent and policy that demands smarter solutions for infrastructure and energy. By splitting our $1.3 billion fund into a $591 million early-stage incubation vehicle and a separate growth fund, we can nurture a “cool idea” from its first prototype all the way to industrial scale. This lifecycle approach means we aren’t just writing checks; we are providing a “war chest” that ensures a company has the capital to navigate the high-stakes world of physical hardware and complex deployment. It allows us to be a consistent partner as these companies move from testing labs to making a serious dent in the global market.
Your portfolio includes electric boats, battery recycling, and autonomous construction vehicles. How do you facilitate direct partnerships between these overlapping fields to build scale, and what steps should founders take to ensure their proof points attract the next wave of demand?
Our strategy is to build a web of startups in overlapping fields—such as transportation, energy, and defense—that act as natural partners rather than isolated silos. For example, a company like Redwood Materials focusing on battery recycling can find immediate synergy with electric boat developers like Arc or autonomous vehicle firms like Wayve. We encourage founders to look beyond their specific niche and collaborate early on to build shared proof points that validate their technology in real-world environments. To attract the next wave of demand, founders must demonstrate that their solutions are not just functional in a vacuum but are capable of integrating into a larger industrial ecosystem. This collaborative scaling is what creates the “flywheel effect,” making the entire portfolio more resilient and attractive to major enterprise customers.
Physical AI is currently being propelled by a confluence of talent, technology, and policy. Which of these drivers is the most difficult to navigate right now, and what specific metrics do you use to evaluate if a startup is truly ready to solve problems in infrastructure or energy?
Policy and infrastructure readiness are often the most complex hurdles because they involve regulatory timelines and physical constraints that digital software simply doesn’t face. While talent and technology are accelerating rapidly, navigating the “actual actions” required to solve problems in the real world requires a startup to be highly adaptable to shifting government mandates. When evaluating a startup, we look at the maturity of their “physical AI” models—specifically how well their intelligence translates into mechanical precision and safety. We also measure their potential to connect different sectors, such as using transportation data to inform energy grid management. A startup is truly ready when it moves past the “cool idea” phase and shows it can execute reliably under the harsh, unpredictable conditions of physical work sites.
You are actively building new companies from within an incubation fund. Could you describe the step-by-step process of taking a “cool idea” from a concept to a functional enterprise, and how do you decide which sectors are ripe for a brand-new startup?
The incubation process starts by identifying a massive friction point in an essential industry, such as compute, defense, or infrastructure, where current technology is lagging. Once we identify a sector ripe for disruption, we leverage our $591 million early-stage fund to assemble a team of world-class talent and provide the operational support needed to build the first functional prototype. We move step-by-step from conceptual research to industrial-grade lab work, ensuring the startup has the “war chest” needed to survive the capital-intensive early phases. The decision to build a brand-new company from scratch usually happens when we see a gap in the market that existing startups aren’t addressing, particularly in how data can be used across enterprises. It is an intensive, hands-on journey that turns a theoretical breakthrough into a functional enterprise capable of handling real-world complexity.
Scale often depends on how companies use data across different sectors to build a competitive moat. How can a startup in the transportation sector effectively share or utilize data with one in the defense or compute sectors, and what are the risks of this interconnected ecosystem?
The next major insight in venture capital is understanding how to use cross-sector data to train smarter AI models that benefit a much broader group of industries. For instance, the sensory data gathered by autonomous construction vehicles can be repurposed to improve navigation algorithms for defense or enhance the spatial awareness of industrial robotics. This data sharing builds a powerful competitive moat, as the models become more sophisticated than any single-sector competitor could achieve on their own. However, the risk lies in the complexity of the interconnected ecosystem—a failure in one part of the chain, such as a battery supply bottleneck, can ripple through the transportation and energy sectors. We mitigate this by ensuring that our portfolio companies are not just partners in data, but also partners in operational resilience, building a shield against market volatility.
What is your forecast for physical AI?
My forecast is that physical AI will become the primary driver of global GDP growth over the next decade as it moves from experimental labs into the backbone of our infrastructure. We are entering a period where the intelligence we’ve refined on our screens will finally take a physical form, automating the most dangerous and labor-intensive tasks in energy, construction, and logistics. Within the next few years, the distinction between “tech” and “industry” will disappear entirely, as every physical asset will be managed by an advanced, interconnected layer of AI. This isn’t just about robots; it’s about a complete re-engineering of the physical world to be more efficient, sustainable, and responsive to human needs. It is the most significant technological era we have ever entered, and the scale of its impact will far surpass the mobile or social media revolutions.
