Pete Martin recalls raising a $5 million seed round at a $25 million valuation for his AI-enabled cybersecurity company Realm back in 2024, like the last “AI” millennium.
He remembered that the amount seemed high for that amount at the time. But today, it’s “normal” to see a $10 million interest in the $40 million to $45 million range, especially if you’re an AI company.
Actually, that kind of thing only happens if you are an AI company, as investors show little interest in anything else.
At the recent Y Combinator Demo Day held in March, everyone was talking about how highly valued the companies are, said Ashley Smith, general partner at Vermilion Fund. Many startups already had six to seven customer contracts, including a company that was only eight weeks old, he said, so there are companies asking for $5 million for $40 million in postage.
In this case, it was more than the so-called “YC tax,” which is how much an investor is willing to pay just because a startup goes through YC, he believes. Even with those early revenue numbers, Smith said investors in this market are calling for price rounds “years before the drawdown.”
Big business firms, flush with cash, are also entering early rounds, raising startup prices and valuations in the hopes of getting big money if these companies exit or IPO one day. Small VC firms have an insatiable appetite for AI companies, too. As an investor focused on AI infrastructure, Smith said he can easily find out-of-cycle pricing, especially if a large company enters. This is one of the reasons why the number of seed deals is low but valuation is high, both founders and VCs say, and data from Carta shows.
Shanea Leven, founder of the AI application platform Impromptu, blames Cursor, which, in early 2025, hit $100 million in revenue in just 12 months. It was one of the first high-profile AI companies to raise the bar on how fast these startups can get, though it wasn’t the only one. Others include Lovable, Bolt, OpenEvidence, ElevenLabs, all of which boast their fast downloads. Although these are outliers, it is difficult for others to feel the heat again.
Techcrunch event
San Francisco, CA
|
October 13-15, 2026
“Investors expect that now,” he said. “The pressure is very high, not to be a billion-dollar company, but to be a $50 billion company.”
Fast moving, great prices
VCs are quick to defend the reason for the rise in seed prices. For example, Marlon Nichols, a general partner at MaC Ventures, said proofs are a way to get traction out of the gate, driving seed prices. When he launched his company in 2019, he said his paycheck was $2.5 million. Today, it’s $5 million.
“The best quality companies no longer look like seed companies,” he said. Advances in AI tools mean that innovators can get to small viable products and find early customers faster than ever, even among large enterprises, which are eagerly looking for ways to employ AI.
Nichols’ last seed investment already generated more than $2 million, with “paid pilots from big businesses” and “a clear line of full commercial deals.” He cut checks between $3 million and $4 million, and agreed to value the startup at $25 million and $30 billion post-money, respectively, which is a lot of money compared to a few years ago.
The backgrounds of the founders also played a role in his term paper submission. “They had the right information” and a “history of execution,” he said, “which reduced a lot of the risk of the first phase.”
Also, investors are willing to pay astronomical premiums for proven AI talent, favoring second-generation founders or those with a relevant pedigree from a relevant previous employer (such as OpenAI). This, in turn, brings the expected values across the board.
“There’s a battle of the big researchers right now, and I don’t think it’s good or bad; it’s the current state of the market,” says Amber Atherton, partner at early stage consumer fund Patron.
That’s what’s driving the most extreme seed valuations, like OpenAI Mira Murati’s previous $2 billion seed for Machine Thinking Labs at a $12 billion valuation.
Leven, who is also the second founder, said that the valuation of his startup in this category is twice that of the first in the same category. Not only is his latest company AI, but it also has more impact than his previous startups at this time, showing how quickly new companies like his can grow.
“Right now I have a lot of contracts for six people, I’m closing seven people right now. You have to have that to raise,” said Leven. “My friend suggests the same round, not AI, and it took him two years compared to my three weeks, to get half of what I got.”
Pre-seed is the new seed
Seed VCs like Vermilion’s Smith are dealing with rising seed valuations by doing more pre-seed deals. Pre-seed startups are the type of startups that seed companies used years ago: very early, early revenue.
Jonathan Lehr, a general partner at Work-Bench, invests in a $160 million fund focused primarily on seed acquisitions, though he said the firm has become “comfortable” with entering the market as companies grow more rapidly.
It’s more common to see early stage investors, as increased exposure is the price of “accessible companies that can grow quickly and become category leaders,” Lehr explains.
Atherton, on the other hand, said that to get a share of these promising startups, the average check size for his company’s $100 million Fund II is now from $4 million to $5 million, from $1 million to $2 million for its $90 million Fund I.
“AI has raised the bar so much that creators can have a live product with users and revenue right out of the gate,” he said. “Investors should move quickly and document real-world developments early because the best innovators are shipping products with users and revenue quickly.”
So seed VCs are no longer “supporting ideas”, they are “supporting initial evidence of real demand for the product,” he explained. Seed VCs are also moving quickly, “from less active to more confident decisions about distribution, retention, and founder taste.”
But there is a catch
As the numbers have grown, so have investors’ expectations.
It’s no longer enough, Atherton says, for a company to simply build and ship a product. Anyone can do that these days. It’s not about traction, although that helps a lot. It is about the future, the inventors can say how they will be able to do better than everyone else and beat everyone else in the market. That’s what these seed VCs believe will propel these startups into long-lasting, $50 billion+ companies, or at least some kind of profitable exit.
“People are just trying to survive the pressure,” Leven said. Otherwise, you won’t have enough money to grow, to really compete.
The good part about raising a lot of money early on as a founder is that it helps the company move quickly and hire valuable talent. VCs know, as they price their term sheets, that talent in the age of AI is expensive, as they run the AI models that support these startups, and compete with other powerful competitors, sometimes large SaaS competitors that already cost billions.
Everyone, Leven said, is trying to recreate the magic of buying Google the Wiz. But the risk is also high. Founders must grow their companies into businesses that justify high valuations before they need additional capital. Series A investors also expect bigger, faster, and more.
Nichols and his firm are now writing more young companies than ever before, with the renewed hope that they will reach their milestones within 18 months. “That behavior is as important as supporting the winners,” he said.
Higher seed rates mean less margin for error, Lehr said, adding: “Less testing ground, less tolerance for pivots, and more scrutiny if progress doesn’t match the money raised.”
Martin, a cybersecurity founder, successfully raised his Series A late last year, saying the benchmark was no problem for his company to remove. But he also had a warning for the founders.
“You can end up stuck in the middle,” Martin said. “It’s very expensive for new investors, but without the traction to justify the next round.”
