Inside The Earliest Bets Of The AI Era

What separates a legendary seed investor from the rest isn’t just timing or returns—it’s the courage to back ideas when they’re still invisible to the market. This year’s Forbes Midas Seed List doesn’t celebrate those who rode the AI wave; it highlights the ones who saw it forming beneath the surface, years before the surge.

Gili Raanan didn’t wait for the cloud to become critical infrastructure—he bet on security only after understanding that complexity breeds vulnerability. His early stake in Wiz wasn’t lucky; it was logical, drawn from a long-held thesis that attackers evolve faster than defenses. That same foresight led him to Cyera and Fireblocks, creating a constellation of companies now central to how enterprises protect data in an AI-driven world.

Ali Partovi has always hunted for outliers. When he handed a pen-and-paper coding test to a college freshman—Michael Truell—that moment wasn’t just a curiosity, it was calibration. His seed bet on Cursor (Anysphere) now looks visionary, not because of the $60B SpaceX option, but because it revealed a new pattern: the best AI investments aren’t in tools that assist coding, but in those that redefine how software is built. And he did it again with Vanta and Kalshi, spotting behavioral shifts before they had names.

David Tisch’s approach stands out not for its brilliance in prediction, but for its consistency in participation. As the self-proclaimed 'Switzerland of VC,' BoxGroup avoids ownership battles and instead floods the zone—with stakes in Ramp, Ro, and Cursor. His strategy? Don’t force revolution; anticipate where regulation, technology, and behavior collide. Healthcare, fintech, and dev tools—all transformed by software, all now reshaped by AI. Tisch didn’t need to pick winners early; he just needed to be present where change was inevitable.

Then there’s Topher Conway, quietly assembling a mosaic across the AI stack. From Hugging Face to Cerebras, Safe Superintelligence to Harvey, his portfolio reads like a map of where AI’s value is compounding: not in one layer, but in their interplay. He didn’t just back infrastructure or models—he backed the connective tissue. That holistic view, rooted in deep technical curiosity and founder trust, let him plant flags across semiconductors, open-source, and vertical applications before each became a frenzy.

While others chased chatbots, Ross Fubini and Greg Castle were investing in rockets, robots, and radar. Fubini saw what most VC firms missed: that national security would be rebuilt by software, not bureaucrats. His 2017 bet on Anduril wasn’t speculative; it was strategic—a bet on speed, autonomy, and AI-driven decisioning in real-world systems. Castle, similarly, anchored on a belief that computing was escaping the screen. From Oculus to Rec Room to Flock Safety, his thesis centered on spatial intelligence and real-time interaction. These weren’t defense plays—they were bets on the physicalization of AI.

Sarah Guo stands in a different lane—one where frontier models meet autonomous agents. While many backed AI as an augmentation tool, she saw Sierra and Mistral AI as harbingers of a new software paradigm. Her conviction wasn’t just in stronger models, but in systems that act, not assist. That’s a critical distinction: moving from copilots to pilots. And by backing Physical Intelligence before stealth, she signaled confidence that the next wave of AI wouldn’t live in the cloud—it would walk, roll, and adapt in the real world.

What unites these investors isn’t a shared playbook, but a shared radar. They didn’t wait for product-market fit. They bet on trajectories, on inflection points, on the silent escalation of technological leverage. The Midas Seed List reminds us that seed investing at its best isn’t gambling—it’s pattern recognition on hard mode.

What’s your earliest bet in AI looked like—and where did it lead? Dive into the full list to see how foresight becomes fortune.