OpenAI Veterans Launch Zero Shot Fund, Close Initial Round Toward $100 Million Target

OpenAI alumni have quietly launched a venture fund and completed an initial close as they hunt for a larger pool of capital, the founders said. The vehicle, called Zero Shot, has already placed a handful of early bets and is aiming to raise roughly $100 million; the partners report they’ve secured about one-fifth of that target and have started writing checks. The founding team includes several former OpenAI engineers and operators who are now pooling their time, networks and sector know-how to back seed-stage artificial intelligence companies.

The group behind Zero Shot brings a mix of product, research and operating experience from the model-maker. Evan Morikawa, who led applied engineering work during the rollouts of systems including DALL·E and ChatGPT, is among the partners. Andrew Mayne — widely known for his early work on prompting and now a public commentator on AI — co-founded the fund alongside Shawn Jain, an engineer turned investor and startup founder. Joining them are Kelly Kovacs, previously a founding partner at growth-stage firm 01A, and Brett Rounsaville, a product and executive alum of Twitter and Disney who runs a consultancy tied to the team.

The partners say side work advising venture firms and founders after leaving OpenAI made clear there was an opportunity to build a specialist fund. They spent months meeting with institutions and family offices before locking down the initial capital. After seeding the vehicle with roughly $20 million in commitments, they set their sights on a full close of about $100 million and began deploying capital into companies that match their thesis. An industry observer familiar with early-stage AI investing noted that teams like this often enjoy privileged deal flow because they can evaluate technical claims more quickly than generalist firms.

Zero Shot has disclosed several early investments. The fund backed Worktrace AI, an enterprise startup founded by former OpenAI product manager Angela Jiang that aims to identify and automate routine workflows for large organizations; PitchBook data shows Worktrace raised a $10 million seed round with participation from notable AI backers. The fund also put money into Foundry Robotics, which is developing next-generation factory automation that integrates machine learning with industrial systems; that company recently closed about $13.5 million in seed financing led by Khosla Ventures. A third portfolio company remains in stealth.

The partners say their technical background also helps them rule out crowded or overhyped corners of the market. They’re wary of new coding tools that lean on user-facing subscription models, arguing major model providers may soon offer similar functionality that undercuts standalone services. Morikawa is skeptical about startups selling large collections of embodied video data for robotics, saying the science required to bridge simulated and real-world motion remains far from solved. Mayne has tested a number of so-called digital twin businesses and believes simpler reasoning models can sometimes match the pitch of pricier simulation platforms; he cautions that progress in foundational models is rarely steady or predictable.

Zero Shot has also assembled a slate of advisors who will share in the fund’s carried interest. Among them are Diane Yoon, who led people operations at OpenAI; Steve Dowling, a communications executive who worked at OpenAI and Apple; and Luke Miller, a former product leader at OpenAI. Observers say the combination of operator credibility and inside knowledge could give Zero Shot an edge in sourcing technically ambitious startups — but it also makes the fund’s activity one to watch as the market judges which AI bets actually pay off.