OpenAI alums have been quietly investing from a new, potentially $100M fund

A new venture capital fund staffed by alumni from OpenAI has made its first close on a $100 million target and has already begun writing checks. The fund, named Zero Shot, brings together former key personnel from the AI lab, including the head of applied engineering for DALL·E and ChatGPT, an original prompt engineer, and a former researcher. Their emergence as capital allocators is significant because it introduces a team with direct, operational experience from the epicenter of the generative AI boom, offering a different perspective than that of traditional financial VCs.

Zero Shot was co-founded by former OpenAI engineers Evan Morikawa and Shawn Jain, original prompt engineer Andrew Mayne, venture partner Kelly Kovacs, and operator Brett Rounsaville. The partners explained that after leaving OpenAI, they were frequently consulted by VCs and founders, which revealed gaps between the technology being funded and market needs. After raising an initial $20 million from institutions and family offices, they have backed Worktrace AI, a management software platform, and Foundry Robotics, an AI-enhanced factory robotics company, with a third investment in stealth.

The fund's core thesis relies on its founders' ability to predict the trajectory of AI models and sidestep overhyped investment areas. For example, the partners are bearish on most 'vibe coding' platforms, believing model makers will absorb that functionality. They are also skeptical of robotics startups focused heavily on certain types of video training data and most 'digital twin' applications, arguing that the underlying technology is either premature or that standard LLMs can already accomplish the task. This insider knowledge could prove to be a critical filter in a market saturated with AI startups.

The emergence of operator-led funds like Zero Shot signals a maturation in the AI investment landscape. Their value proposition is not just capital or network access, but a grounded, predictive understanding of the core technology's roadmap, allowing them to bypass hype cycles and invest in companies with genuine, defensible moats.