From banana farm to $12.5B AUM — Lior Susan on why Eclipse builds what it backs
Key points
Key takeaways from a Sourcery with Molly O'Shea interview with Eclipse Ventures Founder & CEO Lior Susan (September 2026):
Operators with capital. Susan describes Eclipse as "operators with capital" rather than a classic venture firm: he left Flex and founded Eclipse in 2015, and the firm now runs about $12.5B in assets under management — a figure he expects to have moved up since recording. Recent public-market wins in physical industries, including Cerebras Systems and SpaceX, are doing extremely well.
The co-founder who changed everything. Eclipse was born out of a 2015 meeting between Susan, then 31, and Pierre Lamond, then 85 — a veteran of Fairchild, National Semiconductor and 35 years at Sequoia. Four months later the firm existed, Lamond worked full-time in the office for seven years, and at 96 he still mentors the team; Susan calls him father, partner, mentor and friend rolled into one.
Building what it backs. Eclipse counts roughly 90 portfolio companies, about 30 of them incubated and built inside the firm — including Bright Machines, Bedrock and Maitra, with more unannounced. Susan says founding companies keeps the team's operating skills fresh rather than going stale a decade into investing.
Mind: a Rivian carve-out. After three years of conversations with RJ Scaringe about general-purpose robotics for manufacturing, Eclipse convinced him to spin the effort out of Rivian: a mobile, high-dexterity robot deliberately not shaped like a human, able to attract external capital and work with anyone. The first product is due this year.
The physical-world thesis. Susan argues 85% of global GDP — roughly $100T — sits in the physical world, and that China built a ~$20T economy on it by aligning five forces: talent, policy, capital, technology and customer demand. With deglobalisation, the West can no longer rely on imported energy, manufacturing and defence; metal manufacturing alone is a $3T industry that drew almost no venture capital when Eclipse started.
Second acts. Physical companies keep spawning new businesses: Cerebras went from chip to system to running data centres, while Redwood Materials moved from battery recycling into energy storage for data centres — Eclipse led a very large round and Deepak Ahuja, Tesla's former CFO, joined as CFO. SpaceX went from launch to Starlink, and True Anomaly from space-domain awareness to Golden Dome.
The Eclipse economy. Portfolio companies signed roughly $40–50B in commercial deals last year, and one large deal served by three or four holdings turns a dollar into four or five. Eclipse helped shape US AI policy with the administration, avoids foundational models entirely in favour of AI picks-and-shovels and physical AI, and built a new gas turbine business after seeing five-year lead times at incumbent OEMs.
Gross margin is not the point. Susan argues that free cash flow, not gross margin, is what public markets price — and that inflated SaaS gross margins are partly an accounting trick moving engineering spend from COGS into R&D. Cerebras trades at a 40–50% gross margin better than most SaaS companies, and Tesla at mid-teens multiples like the best software businesses, because they solve hard problems in gigantic markets.
Read more: Sourcery with Molly O'Shea
