The NoCap Moment: When AI Started Writing Cheques

NoCap invests in a week. DVC replaced analysts with agents. EQT built Motherbrain.

AI native VC is here, and it is already rewriting the industry's job descriptions.

Written by: Ilya Vencjuns | Research Contributor, London Venture Capital Network

In March 2025, Artem Goldman, the founder of Wonder Family, received the kind of email every early-stage founder dreams about: a 100k dollar investment. The twist was that the cheque came from NoCap, an AI investor, and the decision was made without a single human in the loop.

This was the first widely reported example of an AI-directed investment closing in under a week, whether you think it is a stunt or the beginning of something much bigger, it marks a shift in how venture capital is starting to operate, the same AI models VCs have been funding for the past decade are now creeping into the way VCs themselves run, decide, and compete.

This piece kicks off a three-part series unpacking that shift.

Part 1: focuses on AI-powered investors, the NoCaps of the world, and the quieter but more consequential funds already pushing AI into their investment process.

Part 2: will look at relationship automation for dealmaking, including tools like Boardy, Happenstance, Folk, and Affinity, and the emerging AI social graph.

Part 3: will explore signal-hunting infrastructure such as Specter, Harmonic, and the AI-driven data stacks that surface companies before humans do.

NoCap describes itself as the world's first AI investor. Founders upload their deck and startup data, with NoCap analysing thousands of decks and founder profiles to pick the ones it considers investable. The human element is removed from the investment process - no analysts sifting inbound, no hours spent on due diligence. A network of exited founders and VCs in NoCap's community was then used to introduce the founder to potential talent and investors during a three-minute closing call. NoCap operates through voice and email, responding to and communicating with founders directly, its judging criteria were built by training the model on the views of successful exited founders, as well as public data on the metrics VCs should be evaluating. And while NoCap has been the most viral example of an AI-first VC model, funds from Europe to the US are quietly experimenting with similar approaches.

Venture Collective (DVC), a fund backed by LPs and AI researchers, announced that it had dismissed five VC analysts and replaced them with AI agent workflows. These workflows now carry out due diligence, market research, and the drafting of investor memos, the final call on whether to back a founder is still made by partners, yet the manual work of research and backend operations has shifted from analysts to AI workflows. This sped up decision cycles and helped DVC stand out from other VCs competing for deals, as well as providing partners more time that could be used to support founders.

Similarly, EQT Ventures, a Europe-based fund with 500 million euros under management, has been integrating AI models since 2016. Called Motherbrain, this platform collects publicly available data competitor analysis, market trends and integrates it with internal data from meeting notes, past founder communication, and EQT's thesis. Reporting has shown that seven out of fifty deals in the past decade were directly contributed to by Motherbrain, with the fund now using the system to evaluate personality traits and produce predictive scoring based on psychometric tests submitted by founders. As with DVC, final decisions are made by partners and investment staff, with data scientists and engineers working alongside them to build out EQT’s models.

A future where AI takes greater control over sourcing, scoring, and evaluating startups is already coming into view, and it raises important questions for LPs backing these funds and for the career of VC itself. The integration of AI models into the investment process adds a new layer of risk and opportunity to due diligence. GPs may increasingly differentiate by the internal workflows and proprietary data engines they build to speed up decisions and surface insights earlier than competitors. The key risk will be understanding how separate the human is from the AI, raising governance questions around model use, scoring, and decision transparency at a time when regulation is still in its infancy.

Understanding the AI Stages of Adoption: A framework for business leaders:

Automation of once manual, time-consuming tasks will trim headcount and create opportunities for smaller, leaner funds to spend more time picking winners and supporting founders. But a world where the work of VC analysts is outsourced to models creates ripple effects for the VC career path itself (and career opportunities more broadly). Traditionally, one entry route into VC has been as an analyst post-university or post-MBA. Yet for those whose skills are limited to analysis and research, these capabilities may become irrelevant, as those same functions will be streamlined by workflows and internal models.

At the same time, the proportion of VCs entering with MBAs has been falling: in 2000, roughly 44% of senior VCs had an MBA, By 2025, that number had dropped to 32% - in recent years, funds have started to favour domain expertise or operator and specialist backgrounds over generalist skills and the integration of AI into the investment value chain is catalysing this shift.

 Read more insights at londonvcnetwork.com/publications