A new venture fund, Zero Shot, has secured its first close toward a $100 million target. Its founding team includes former OpenAI engineers Evan Morikawa, Andrew Mayne, and Shawn Jain, who bring direct experience scaling foundational products like ChatGPT and DALL·E. They’re joined by venture capitalist Kelly Kovacs and operator Brett Rounsaville.

The fund’s thesis is built on a specific engineering perspective. After leaving OpenAI, the founders were inundated with requests for technical due diligence from investors and founders. They observed a market misalignment: many funded startups were building on shaky technical premises or chasing problems that core model improvements would soon solve.

Their initial investments reflect a bias toward applied systems engineering over pure model work. They’ve backed Worktrace AI, which focuses on enterprise task discovery for automation—a data pipeline and integration challenge. Another bet is on Foundry Robotics, applying AI to factory robotics, a domain requiring robust physical-world data loops. A third, stealth investment is already in place.

Their technical background informs what they avoid. Mayne is skeptical of most ‘vibe coding’ platforms, anticipating that model providers will directly absorb their functionality. Morikawa questions startups collecting video data for robot embodiment, citing a fundamental ‘embodiment gap’ that current research cannot bridge. Mayne also finds ‘digital twin’ startups often offer little over a well-prompted LLM, a conclusion reached after building reasoning models to test them.

The fund positions itself as a connector between deep technical insight and practical deployment, with advisors including former OpenAI heads of people and communications. For data engineers, the story is about practitioners applying first-hand scaling experience to identify startups solving hard infrastructure problems, while filtering out those likely to be obviated by the next model release.

Source: TechCrunch