Giant AI funding rounds hide how hard it is to raise a second round

Headlines about giant AI funding rounds suggest startup cash is everywhere. For most young companies, it is not. US startups closed 272 Series A rounds, the first big round after an initial seed round, in the third quarter. That is about 28% fewer than the 377 a year earlier, according to Crunchbase data.

What the numbers show

Crunchbase reported in late September that startups worldwide had closed at least 114 Series A rounds of $100 million or more this year. More than 70% went to AI companies. Those deals pull the averages up and hide how few second rounds are happening overall.

Seed rounds, the earliest outside money, have held up better. There were 984 in the US last quarter, compared with 1,059 a year earlier. The squeeze comes after that first check.

Bigger seed checks, fewer follow-ups

The median US seed round was about $3 million last year, three times the 2018 level. Yet only 24% of companies that raised a seed round of $1 million or more in 2023 have moved past seed. For the 2024 group the figure is 16%. Through 2020, it was 55% or higher.

The bar has also moved. Andy McLoughlin of Uncork Capital told Crunchbase News that Series A investors now expect $2 million to $3 million in ARR, sometimes $4 million. ARR means annual recurring revenue, the yearly income a company can count on from customers who pay on a regular basis. The time from seed to Series A has stretched past two years.

What founders can do

The advice comes from Marc Schröder, co-founder and managing partner of the venture firm MGV, in a piece for Crunchbase News. His firm sells sales coaching to startups, so he has a stake in the topic. His core point is to size the seed round around what the next round demands.

• Raise enough to reach the Series A revenue bar with six months to spare. If a seed will not do that, a smaller round is the better deal.
• Find out which funds lead Series A rounds at your stage. Funds of $100 million to $500 million typically do.
• Check when each fund's current fund closes, then send the right partners updates for a year before asking for money.
• Build sales that other people can repeat, not only deals the founders close themselves.

Schröder argues that $2 million in ARR with a sales process a new hire can run beats $3 million built on founder-led deals nobody else can repeat.

A example that worked

Atomic, a New York company founded in 2023 that makes supply chain planning software, raised a $3 million seed in April 2025. That is right at the median. Madrona and DVx Ventures led it. About 17 months later, Atomic closed a $12.5 million Series A led by Madrona and Klass Capital. Atomic is not an MGV portfolio company.

Who it affects

Founders, early employees and anyone who backs young companies feel this most. A large seed round no longer signals an easy road to the next one. The article does not say how many companies will fail to raise again. It shows that many are now stuck between rounds.