AI mega-rounds are changing how seed investing works, and we can see the impact at many levels.
First, this change is evident in the data. Crunchbase shows that both the number of deals and total funding in rounds from $200,000 to under $5 million in the US dropped by about 20% in 2025. The rounds from $5 million to under $10 million remained relatively flat. Only the $10 million-plus rounds grew. They accounted for 51% of all US seed funding, up from roughly a third in 2024.
At Zubr Capital, we wanted to see how far this shift had traveled down the market, so we reviewed the data and looked at how other funds were responding.
While figures like those from Crunchbase cover the entire seed market, signs suggest AI is driving the split. According to the same reporting, many larger funds that invest across different stages are now backing promising AI companies earlier and at higher valuations. As these larger investors move into seed, smaller funds are shifting their timing. They are rethinking how many companies to back and how to finance later rounds.
The battle for AI winners begins at seed
Large funds are investing in promising AI companies earlier because they want to secure meaningful ownership before competition intensifies and valuations rise. If that bet pays off, their larger capital base lets them keep investing through Series A, Series B, and beyond. A smaller seed fund usually cannot do the same.
In late July 2026, PitchBook’s 30 leading VC firms, selected based on their investment track record, participated in seed deals accounting for 34% of global seed deal value, up from 11.7% in 2023. Their median seed check was $9.1 million, compared with $3.7 million across all seed deals globally. For smaller funds, that creates pressure at two points: matching the initial check and maintaining ownership as subsequent rounds grow.
Subsequently, LPs — the institutions and individuals that invest in VC funds — are concentrating more of their capital with the largest managers. In the first half of 2026, funds of $1 billion or more captured 68.3% of the capital raised by US venture funds. By mid-April of that same year, US-based growth and late-stage venture funds had already raised a record $23.6 billion as investors pursued a narrow group of fast-growing AI companies.
How smaller funds compete when AI rounds get bigger
Some smaller funds are adapting as AI rounds grow. Katie Stanton of Moxxie Ventures calls the market bifurcated. She points to a small group of fast-growing AI teams attracting massive Series A rounds. Everyone else faces a tougher fundraising environment. Moxxie now allocates 60% to 70% of its fund to first checks in new companies, up from 50%, and meets founders earlier — sometimes before product-market fit.
At 645 Ventures, the approach is to make capital intensity part of the investment filter. Co-founder Nnamdi Okike expects investors to be more cautious about capital-intensive, “spend at all costs” strategies associated with hyperscalers and foundation models. The firm lists investment ranges of $1 million to $5 million at seed and $5 million to $10 million at Series A. Its target ownership is 8% to 10%. That model favors AI companies that can reach meaningful milestones within conventional early-stage rounds, rather than businesses whose development depends on multibillion-dollar financing.
The next challenge comes when an early bet succeeds. A smaller fund might have pro rata rights, but not the reserves for a much larger follow-on. Industry Ventures identifies deal-by-deal SPVs as one solution. Alpha Partners provides $3 million to $15 million through such vehicles to help early-stage investors fund later rounds. This mechanism is not unique to AI, but becomes more valuable when an AI winner’s capital needs to rise sharply.
Each response starts with the same question: how much capital will this company need after the first check, and who will provide it?
Smaller funds need to plan beyond the first check
As AI rounds grow, a company’s future capital needs become part of the seed decision. I have always viewed fundraising as a means rather than a measure of progress. An investor must judge a round by the result it buys. A large early round may be justified when an AI firm requires more computing power, scarce specialists, or physical infrastructure to compete. A company seeking funding to test retention, unit economics, or demand in a particular market may require less capital. In those cases, revenue, grants, debt, or strategic partnerships can sometimes finance part of that work without another equity round.
What the capital is meant to finance changes the investment decision. If the company is likely to outgrow the smaller fund’s reserves, the investor needs to assess whether the fund can assemble a syndicate, introduce later-stage capital, or accept dilution. A business that depends on multibillion-dollar follow-ons may simply require a different type of investor.
As an example, Zubr Capital led Placy’s €1 million pre-seed round. Its AI product was built to automate routine work for real estate agencies, including property valuations, market research, scheduling viewings, and drafting agreements. The relevant milestone is operational: can an agency handle more potential customers and transactions with the same-sized team?
Changes at seed
AI mega-rounds are changing venture activity at the lower end of the market. Large multistage investors are moving into promising AI companies earlier, bringing more capital and competition into the seed stage.
Smaller funds are adapting. They are entering earlier, making more initial bets, screening companies for future capital needs, and bringing outside investors into follow-ons. For a smaller fund, the first check increasingly depends on what comes next: how much capital the company will need, when, and who can fund the next stage.
Member discussion