US start-ups have raised $42 billion (£33.1 billion) in Series A funding so far this year, a 60% increase on 2025, but the number of companies securing investment has barely changed.
Data shows that larger cheques, particularly for artificial intelligence businesses, are driving the surge. The annual number of Series A rounds worth at least $1 million has remained broadly stable at about 1,200 since 2023, with 2026 expected to be within roughly 100 deals of that average.
Meanwhile, the total raised has almost doubled compared with 2023. Mirendil, an AI research laboratory, secured $1 billion last week, while AI infrastructure company TypeSafe AI raised $870 million in October.
Bigger Series A rounds and higher expectations
Roseanne Wincek, co-founder of early-stage investment firm Renegade Partners, said rounds of between $25 million and $50 million were no longer unusual for a Series A.
“Today, a $25 to $50 million Series A is not that high. When I started in venture, that was a large C or D,” she said.
Wincek said start-ups were more expensive to operate, while investors were demanding greater progress before committing funds. The supply of capital has also increased, including from multi-stage firms moving into earlier funding rounds.
“The supply of capital has never been larger and especially when multi-stage firms come and do earlier rounds. The difference between them writing a $10 million and a $50 million check is actually pretty de minimis,” she said.
However, the additional money does not necessarily go as far. Wincek said competition for talent had pushed up employment costs in San Francisco, while the cost of computing had risen as AI adoption increased.
Data centre demand has more than doubled the price of memory over the past 12 months, according to technology research firm Gartner.
Start-ups are also expected to demonstrate more traction at an earlier stage. Wincek said the level of annual recurring revenue that once supported a Series A could now be enough only for a strong seed round.
“I think that everything has just shifted up, but the bars are higher,” she said. “It used to be that if you got to $1 million of ARR, you could raise a Series A and today you probably need $1 million of ARR to raise a decent seed.”
AI founders attract outsized backing
One exception is a small group of highly credentialled founders whose previous experience can help them raise substantial sums before building a track record.
“There’s this phenomenon around ultra-pedigree founders coming out of the gate and raising very large rounds with a deck and nothing else,” Wincek said.
Mirendil’s co-founders, Behnam Neyshabur and Harsh Mehta, previously worked at Google and Anthropic. Its wider founding team includes researchers and engineers from OpenAI, Google DeepMind and xAI.
TypeSafe AI was founded by former OpenAI researcher Diogo Almeida and former Meta research engineer Sasha Sheng.
The scale of the AI funding boom will face a test as Anthropic goes public, bringing private-market enthusiasm into sharper focus against the realities of the public markets.
