How Generative AI Is Reshaping Venture Capital

The venture capital world has witnessed countless technological disruptions over the decades, but few have arrived with the speed and scope of generative artificial intelligence. In the three years since ChatGPT’s launch, gen AI has fundamentally altered how startups are conceived, pitched, built, and funded. According to a report from Stanford, global private AI investment reached a record high of $252.3 billion in 2024, with generative AI funding soaring to $33.9 billion—more than eight times higher than 2022 levels. VC firms are well positioned to help founders make sense of this new world. Our recent interviews with seven prominent venture capitalists reveal an evolving landscape for both startups and VC firms that could eventually reshape the industry. (We selected these firms to interview based on their funding of generative AI startups and variance in size and industry focus.) These venture capitalists told us that AI is now involved—as a core capability or a feature—in virtually every tech investment they consider. But it’s also changing the way VC firms do their work. Today, it’s shaping startup pitches and required funding levels; it is also beginning to alter the very structure of the firms themselves. Here’s what’s changing, as well as what’s differentiating startups in this hot market right now. More Pitches, Some of Them Better The VCs we interviewed report seeing direct impacts from gen AI on both the number and quality of pitches that they receive. Entrepreneurs now leverage AI to create product and market descriptions, go-to-market plans, branding materials, and even software demos. That’s a double-edged sword in terms of the quality of investment proposals. Vivjan Myrto of Hyperplane Ventures reported that, given the ability for everyone to use the tools to expedite the creation of pitches, “noise is at an all-time high.” Dan Von Kohorn of Broom Ventures sees some benefits, though: “The general quality has ‘barbelled,’” he explained, “with a dramatic increase in low quality pitches, and also a sizable increase in the quality of the top pitches.” Elizabeth Bailey at Foreground Capital, a firm that specializes in women’s health investments, sees that opportunity as well: “In the past, we would see companies with a limited understanding of the market, and now those companies have new analysis superpowers to define and specify their opportunity.” Entrepreneurs should avail themselves of these tools, but be aware that since VC firms will have access to the same gen AI tools that founders do, surface-level research on the business issue, technology, and market for a startup’s offering won’t impress an experienced venture capitalist. Instead, founders and their teams should augment gen AI results with grounded learning from customer interviews, relevant personal experience, and network connections to expert advisors. Efficiency: Fewer Staff, Smaller Deals Founders need to be able to speak to how they operate more efficiently because of gen AI. Startup companies can launch with dramatically reduced tech staff and rescoped tech roles, typically leading to lower levels of required funding to start and scale businesses. For example, Lily Lyman, Managing Partner at Underscore, shared that she’s seeing many companies that only need a third of the number of engineers that they would have required previously; they no longer need full-time marketers; and fewer salespeople can do more. Bill Geary, a longtime venture capitalist at Flare Capital, a large firm focused on healthcare technology, observed similar changes in team composition, as core development tasks that used to require a handful of people can now be done by one talented individual. This efficiency isn’t just about headcount, but also about speed. Traditional excuses for slow progress are no longer acceptable when ventures can iterate concepts in days versus months. As a result, VCs are particularly attuned to the pace of execution. When founders can create a prototype on a vibe-coding platform in a day and immediately show it to a target customer, VCs expect a high level of digital functionality before they see a pitch. This prospect of transformed efficiency has elevated the expectations VCs place on companies’ tech leaders. They want them to think like CTOs—not just developing a solution, but focusing early on issues like integration with the customer’s tech stack, scaling the solution, and incorporating agentic AI capabilities. Virtually all of the VCs we interviewed mentioned that early-stage startups are able to do more with less capital because of gen AI. The combination of AI-enabled efficiency and reduced capital requirements has led to increased seed-stage funding, and less emphasis on later stages. Startups can often combine seed funding with bootstrapping, raising a relatively large seed round ($5–$10 million instead of $1–$5 million) and relying on that capital to get to $50 million or so in revenue, without needing to take further investments. This efficiency trend is supported by industry data showing remarkable examples of lean growth. Notable examples include companies like Cursor achieving $100 million in revenue in its first year, and Gamma reaching $50 million in revenue having raised less than $25 million. Increasing numbers of startups are able to bootstrap their businesses without any venture capital at all, at least for the early stages of the business. The ultimate version of this efficiency trend is the “one-person unicorn” (a startup valued at $1 billion or more), fully enabled by AI. Broom’s Von Kohorn predicted that this animal is coming soon: “I’ll be bold and call it 2026.” Most other VCs we interviewed did not believe that one lone founder could generate a billion dollars in valuation. However, there are already “tiny team” unicorns, like Anysphere and X, achieving $100+ million in revenues with fewer than 50 people (see the chart “Faster Growth, Fewer Employees”). See more HBR charts in Data & Visuals This reality shifts expectations for founders pitching VCs dramatically. Given that VCs expect that founders will be doing more work with less capital because of gen AI, founders and their teams should moderate their funding asks accordingly. And at their first meeting with a VC, they should be prepared to pitch not a concept, but rather a well-defined and in-motion business with functioning software, marketing materials, customers, and perhaps even positive cash flow. What’s Not Changing: The Human Element Despite the widespread adoption of generative AI, many VCs told us that key human aspects of investing remain irreplaceable, such as building relationships, establishing trust, and assessing the character and vision of founders and their teams. These interpersonal judgments still lie at the heart of venture investing and, for now, exceed AI’s capabilities. Broom Ventures’s Von Kohorn said that at least a third of their due diligence is focused on trying to assess the psychology of the founders: Do they have the disposition to learn and act rapidly? Can they lead a larger team and scale this venture to something great? They find that these questions are difficult for AI to answer well because it neither knows the details of founders’ backgrounds nor knows whether to trust their statements. Similarly, Flare Capital’s Geary views the firm’s work as “people mining,” characterizing a human-first approach that contrasts with large-scale data-driven investing. The firm often seeks out promising individuals before they even formalize a startup, emphasizing relationships and a deep network rather than relying solely on inbound pitches that AI could analyze. If anything, the human component seems to be becoming even more important in the age of AI. Underscore’s Lyman said that while they have always been very people-driven investors, now this is even more important because the markets and the technology are both changing so fast. Since VC firms devote considerable attention to the human dimension of founders and teams, it’s important for founders and their teams to emphasize the attributes of entrepreneurial leadership in pitches. It’s difficult to display these traits in a single meeting, so founders should aim to build a relationship with venture fund partners well ahead of the crucial pitch presentation. New Differentiation Criteria As basic AI capabilities become commoditized, VCs are emphasizing factors that separate winning gen AI startups from the swelling ranks of new ventures. Firms attracting capital are focused on several key differentiators: Proprietary data access. VCs are placing added weight on access to unique data and the defensibility of intellectual property. Companies are more likely to receive funding if they develop proprietary datasets that are difficult to imitate. Broom’s Von Kohorn said that his firm was investing in the suppliers of the data—those that own the sensors themselves. The firm expects that the intermediate processing of data to become more and more commoditized over time. AI-native architecture. Rudina Seseri of Glasswing Ventures, which focuses on AI and cybersecurity startups, distinguished between companies that are “AI native”—built with AI from the beginning—versus those simply adding AI features or large language model “wrappers.” AI-native companies also tend to leverage “ensembles” of various AI techniques, indicating a more sophisticated and robust approach to problem-solving. Seseri believes that this kind of deep integration of AI is crucial for building sustainable advantage. Learning agility. Several VCs identified the “rate of learning” as the most critical founder attribute given the rapid pace of change in AI and related technologies. One commented that what you know today is not going to be true in six months, so leaders of startups have to be “maniacal” learners. When pitching they should be prepared to discuss how they will scale the company’s capabilities to the next level and learn and adapt rapidly as gen AI continues to evolve. Founder-market fit. This attribute, a common focus among VC firms, becomes even more critical in the AI era. Hyperplane’s Myrto explained, “We love founders that are absolutely obsessed with the problem they’re trying to solve. They come right from the world where they’re looking to innovate—they have deep knowledge about existing solutions and what new opportunities are.” Several commented that deep empathy for customers and understanding of the target market are key. While gen AI models can output target market descriptions, actual experience with customers in the targeted market space is what really differentiates one founder from another. Deep execution ability. Given that gen AI now helps with the creation and researching of ideas, demonstrated ability to execute on those ideas becomes a primary differentiator for startups. Previous experience at successfully developing and scaling digital products was mentioned by several of the VCs we interviewed. The Future of Venture Capital Some of the VCs we interviewed predict that the venture capital industry itself may be undergoing significant structural changes driven by gen AI. AI appears to be accelerating existing trends toward either massive scale or deep specialization. Underscore’s Lyman explained: “You may ultimately have the multibillion-dollar, multi-stage mega-funds…And then there will be smaller, niche specialized players. I think it will be increasingly difficult for firms to live in the middle.” In that scenario, large funds could leverage AI for broad market coverage and extensive data analysis, while specialized funds could deepen their domain expertise using AI tools for hyper-focused research and analysis. Broom’s Von Kohorn envisioned radical changes in how capital allocation works. He speculated about a future where “there are ‘VC AIs,’ and there will be large, popular AI VCs entering the landscape. Every startup could have a conversation with these VC AIs…which would allow for a theoretically much more efficient allocation of capital.” While this might sound like science fiction, this VC visionary believes that future will arrive quickly. However, not all VCs share this vision of AI dominance. Hyperplane Ventures’s Myrta believes that early-stage investing will remain “much more boutique-feeling” due to the inherently human nature of assessing entrepreneurial potential. And Senofer Mendoza of Mendoza Ventures introduced a crucial ethical dimension to the future of AI in VC, warning that if AI is trained on historical data, it risks perpetuating existing biases in funding decisions. If the models are trained only upon externally visible attributes, the risk would be that “It will tell you, find a traditional founder that went to Stanford and then went into Y Combinator and has these five friends and invest in him.” This underscores the need for a diverse set of entrepreneurs considered by AI to ensure equitable and innovative outcomes. More broadly, even if the idea of an AI VC appears increasingly realistic, there are many challenges to the concept becoming a reality. With no human relationships, for example, it might be difficult for an AI VC to raise the funds to invest. While founders today won’t encounter an AI VC, recognizing this future may be around the corner can prepare them for a hybrid of high-touch and increasingly automated screening. The venture capital industry’s relationship with generative AI is still rapidly evolving, but one thing is certain: The firms and founders who understand how to balance technological capability with the critical human dimensions of innovation and entrepreneurial leadership will define the next generation of investment success.

How Generative AI Is Reshaping Venture Capital

Summary.

Partner Center