From Model Lab to Money Lab: What OpenAI Alumni’s Zero Shot Fund Really Signals
Headline + intro
A group of early OpenAI insiders has quietly done what many big VCs secretly feared: they’ve built their own AI‑native venture fund. Zero Shot, a new vehicle targeting $100 million, isn’t just another AI-branded pool of capital — it’s effectively a satellite orbiting one of the most powerful AI institutions on the planet.
This move matters far beyond San Francisco. When people who helped launch DALL·E, ChatGPT and OpenAI’s first prompt engineering teams start writing checks, the center of gravity in AI investing shifts. In this piece, we’ll unpack what Zero Shot says about the next phase of AI venture capital, who gains, who loses, and why European founders should pay close attention.
The news in brief
According to reporting by TechCrunch, a new venture capital fund called Zero Shot has made an initial close toward a planned $100 million first fund. The partnership consists of five founding members, including three early OpenAI alumni: former head of applied engineering Evan Morikawa, OpenAI’s original prompt engineer Andrew Mayne, and former OpenAI researcher and engineer Shawn Jain. They are joined by VC Kelly Kovacs and startup operator Brett Rounsaville.
TechCrunch notes that the partners have already raised around $20 million and begun investing. Early bets include Worktrace AI, founded by former OpenAI product manager Angela Jiang, which is building software to discover and automate enterprise tasks, and Foundry Robotics, which is developing next‑generation, AI‑enhanced factory robots. A third investment remains in stealth.
Zero Shot’s partners say their OpenAI experience gives them a view on which AI categories not to fund. They are reportedly skeptical of most “vibe coding” tools, robotics companies built mainly on video training data, and many “digital twin” startups, arguing that frontier models and standard LLMs may quickly commoditize much of that value.
Why this matters
The obvious story is “ex‑OpenAI staff launch AI fund.” The real story is that technical insiders are now formalizing their power as capital allocators, not just as advisers or angel investors.
Winners:
- AI founders gain a fund that actually understands model roadmaps, scaling limits and where capabilities are likely to land in 12–24 months. That lowers the risk of being pushed into dead‑end products by non‑technical investors chasing buzzwords.
- OpenAI’s broader ecosystem benefits from a semi‑affiliated pool of capital that can reinforce its platform position: more startups building on the same stack, more feedback into where the APIs fall short, more lock‑in for enterprises betting on that ecosystem.
Losers:
- Generalist VCs who still treat AI as a black box lose information advantage. If you’re a traditional fund partner trying to diligence an AI infra deal against a group that literally shipped DALL·E and ChatGPT, you’re starting several steps behind.
- Founders in fashionable but fragile categories — like thin wrappers around code‑generation models or generic “digital twins” — will find it harder to raise from technically rigorous AI funds. Zero Shot publicly signalling what they won’t fund is a subtle way of devaluing certain hype cycles.
The partners’ skepticism is the most important signal. They are effectively saying: the next phase of AI value won’t come from yet another front‑end to the same models, but from deep integration into workflows, physical systems and hard operational problems. That aligns with their first bets: factory robotics and enterprise automation discovery, not chatbots and slide‑deck generators.
At the same time, there is a concentration‑of‑power issue. If the people closest to frontier labs also become gatekeepers of early‑stage capital, AI innovation risks being pulled even more tightly into the orbit of a few US‑based model makers.
The bigger picture
Zero Shot fits into at least three broader trends in the AI and venture markets:
The rise of operator‑VC hybrids in frontier tech. We’ve seen this before in crypto (ex‑Coinbase and ex‑Ethereum folks spinning up funds) and in dev‑tools (ex‑Stripe and ex‑GitHub operators becoming investors). AI is now entering the same phase, where lived experience shipping at scale is more valuable than a Stanford network and a PowerPoint.
The second wave of AI specialization. The 2023–2024 boom funded hundreds of horizontal AI startups: generic copilots, content tools, “ChatGPT for X.” As models improved, many of those products became features, not companies. The next wave, which Zero Shot seems to be aiming at, is:
- robotics and embodied AI (Foundry Robotics),
- deep workflow orchestration and process mining (Worktrace AI),
- infrastructure and tooling that actually bends cost curves rather than just adding UX layers.
VCs trying to get closer to the model layer. OpenAI already has its own Startup Fund; Anthropic and other labs have strategic investment programs. The emergence of a separate, alumni‑led fund increases the density of capital around a single technical stack. That mirrors what happened around AWS and Azure in the cloud era — but the strategic dependency on a single provider is even sharper with frontier models.
Compared to big Sand Hill Road firms, Zero Shot looks small at $100 million. But its influence won’t be about check size; it’ll be about pattern recognition and signaling power. When a fund run by ex‑OpenAI talent backs a robotics or automation startup, that startup can more easily raise from larger funds that treat Zero Shot almost as outsourced technical due diligence.
In that sense, Zero Shot is less a competitor to multibillion‑dollar firms and more a filter layer: a way to pre‑sort which AI bets are tied to real technical insight versus good storytelling.
The European / regional angle
For European founders and policymakers, this development is a double‑edged sword.
On one hand, a technically sophisticated, relatively small AI fund is exactly the kind of investor European deep‑tech startups say they lack. Robotics hubs like Munich, Odense, Paris and Zurich are full of teams working on industrial automation, perception and embodied AI — precisely the areas Zero Shot seems to favour. A fund that understands both large‑scale model behavior and hard engineering constraints could become an attractive co‑investor alongside European VCs such as Point Nine, Seedcamp, or Earlybird.
On the other hand, Europe is entering the era of the EU AI Act, on top of GDPR, the Digital Services Act (DSA) and soon the AI Liability Directive. Many US funds still treat EU regulation as friction to avoid. A fund anchored in the culture of OpenAI — itself under intense regulatory and political scrutiny — may be cautious about backing startups whose business models lean heavily on sensitive data, biometric processing or high‑risk AI categories as defined by Brussels.
There is also a sovereignty question. If the most technically informed AI capital for European startups comes from US‑centric ecosystems, Europe risks deepening its dependence on foreign model providers and investors. Industrial champions in Germany or France may have to choose between AI excellence tied to US platforms and regulatory comfort with slower‑moving, local alternatives.
For now, Zero Shot has not advertised a Europe‑first strategy. But their focus areas — robotics, enterprise automation, applied AI — map well onto European strengths like manufacturing, logistics and automotive. The most likely near‑term role for the fund in Europe is as a specialist co‑investor in later seed or Series A deals sourced by local funds that need heavyweight AI diligence.
Looking ahead
Zero Shot is still small and early, but its trajectory is fairly predictable.
Fund size and deployment. Given the current appetite for AI exposure among family offices and institutions, it would be surprising if the fund did not reach its $100 million target over the next 6–12 months. The constraint is less LP demand and more the partners’ ability to maintain a tight thesis instead of becoming yet another “AI everything” fund.
Portfolio shape. Expect a relatively concentrated portfolio skewed toward:
- complex B2B workflows where AI can remove large swaths of human coordination costs,
- robotics and industrial systems that benefit from both perception and reasoning advances,
- tools that make it cheaper or safer to run large‑scale AI in production (evaluation, alignment, safety, monitoring).
Zero Shot’s public skepticism about vibe coding, pure video‑data robotics plays and shallow digital twins is a useful leading indicator for founders. If you’re building something that could be wiped out by the next frontier model release, or that relies mainly on access to proprietary data rather than real product moats, you should assume technically savvy funds will mark you down.
Regulatory and strategic questions. As capital and technical influence concentrate around a handful of labs, regulators — especially in the EU and UK — may eventually ask whether there are conflicts of interest when alumni funds invest heavily in ecosystems tied to their former employer’s APIs. That’s not today’s problem, but it will surface once such funds start to matter in Series B and later rounds.
For founders, the practical takeaway is simpler: in the next two years, technical credibility will matter more in your cap table. Having a "brand‑name" AI fund led by operators may become as important as having a Tier‑1 generalist fund.
The bottom line
Zero Shot is more than a side project for ex‑OpenAI engineers; it’s a sign that the AI era is shifting from hype‑driven capital to insider‑driven capital. That’s good news for teams tackling hard robotics and enterprise automation problems, and bad news for shallow wrappers around foundation models. For European founders in particular, the question is whether you want this kind of US‑centric technical capital on your cap table — and, if so, how you balance that with the regulatory and strategic demands of building in the EU.
