It’s Not Your Imagination: AI Startups Have Higher Valuations

Pete Martin recalls raising a $5 million seed round at a post-cash valuation of $25 million for AI cybersecurity company Realm in 2024, also known as a thousand “years of AI” ago.

He noted that this valuation seemed high for that amount at the time. But today, it’s “pretty common” to see a $10 million seed round at a post-money valuation of $40 million to $45 million, he said, especially if you’re an AI company.

In fact, this kind of thing only happens if you’re an AI company, as investors show little interest in anything else.

At Y Combinator’s last demo day held in March, everyone was talking about how expensive companies were, said Ashley Smith, general partner at early-stage fund Vermilion. She said several startups had already secured six- to seven-figure client contracts, including one that was only eight weeks old, so there were companies asking $5 million for $40 million.

This time, it was more than just the so-called “YC tax,” meaning how much an investor is willing to pay just because the startup went through YC, I thought. Even with these early revenue numbers, Smith said investors in this market are making pricing rounds “years before there is traction.”

Larger venture firms, flush with cash, also move into early rounds, driving up startup prices and valuations in hopes of making big gains if these companies exit or go public one day. Small venture capital firms have an insatiable appetite for AI companies, too. As an investor focused on AI infrastructure, Smith said she could easily find herself left out of a round, especially when a larger company moves on. That’s one reason the number of seed deals is low but valuations are high, both founders and venture capitalists said, data from Carta shows.

Shania Levine, founder of enterprise AI application platform Empromptu, puts the blame at the feet of Cursor, which in early 2025 generated $100 million in revenue in just 12 months. It was one of the first high-profile AI companies to raise the bar at how quickly these startups could gain traction, though it certainly wasn’t the only one. Others include Lovable, Bolt, OpenEvidence, and ElevenLabs, all of which boast fast traction. Although these values ​​are outliers, it is difficult for some not to feel the hesitant heat.

“Investors are anticipating this now,” she said. “The pressure is at an all-time high, not to become a billion-dollar company, but a $50 billion company.”

Faster traction, bigger ratings

Venture capitalists are quick to defend the case for high seed valuations. For example, Marlon Nichols, general managing partner at MaC Ventures, said the proof is in the form of traction right out of the gate, driving seed pricing. When he launched his company in 2019, he said his average entry check was $2.5 million. Today it is worth $5 million.

“The best seed-stage companies don’t look like traditional seed-stage companies anymore,” he said. The advancement of AI tools means that founders can have a minimum viable product and acquire their first customers faster than ever before, even among larger companies, which are eagerly looking for ways to employ AI.

Nichols’s last two seed investments were already generating more than $2 million in revenue, with “paid pilots from large institutions” and a “clear line of sight to full commercial agreements.” He cut checks between $3 million and $4 million, and agreed to value the startups at $25 million and $30 million after payment, respectively, a significant sum compared to what it was a few years ago.

The backgrounds of the founders also played a role in its term sheet offerings. “They had the relevant experience” and “a proven track record of execution, which reduced a lot of that early-stage risk,” he said.

Additionally, investors are willing to pay astronomical premiums for proven AI talent, favoring second-time founders or those with the right pedigree from the right previous employer (e.g. OpenAI). This also shows expected ratings across the board.

“There’s a war between the great researchers right now, and I don’t think it’s good or bad; it’s just the current state of the market,” said Amber Atherton, partner at early-stage consumer fund Patron.

This is what drives the more extreme initial valuations, such as OpenAI Mira Murati’s $2 billion initial valuation and Thinking Machine Labs’ $12 billion initial valuation.

Levin, a second-time founder, said that the valuation of her startup at this stage is double its first valuation at a similar stage. Not only is her latest company AI, but it also has much greater traction than her previous startup had achieved at this time, showing how quickly new companies like hers can grow.

“I currently have multiple six-figure contracts, and I’m currently closing on a seven-figure contract. You have to have that amount to raise,” Levine said. “A friend of mine is offering a similar round, not the AI, and it took her two years versus three weeks to get half of what she got.”

The pre-seed is the new seed

Venture capital firms like Vermilion’s Smith are dealing with the rise in seed valuations by doing more pre-seed deals. Pre-seed startups are the kind of startups startups used to be years ago: very early, before revenue was generated.

Jonathan Lear, general partner at Work-Bench, is investing from a $160 million fund that focuses primarily on seed funding rounds, though he said the company has become “increasingly comfortable” entering the pre-seed funding stage as companies scale much faster.

It’s more common to see investors pour capital into startups early, because increased exposure is just the price of “access to companies that have the potential to scale faster and become leaders in their category,” as Lehr described.

Meanwhile, Atherton said that to get a share of these promising early-stage startups, the average check size for her company’s $100 million fund now ranges from $4 million to $5 million, up from $1 million to $2 million for her $90 million debut fund.

“AI has raised the bar much higher for founders to have a live product with users and revenue right out of the gate,” she said.  “Investors have to move faster and ensure real-world traction very early because the best founders ship products to users and generate revenue almost immediately.”

So VCs are not “backing ideas” anymore, but rather “backing early evidence of real demand for consumer products,” as she described it. Seed VC firms also move faster, “from slow diligence to high-conviction decisions about distribution, retention, and founder taste.”

But there is a catch

As risks rose, so did investor expectations.

It’s no longer enough for a company to simply build a product and ship it, Atherton said. Anyone can do it these days. It’s not even about traction, although that helps a lot. It’s about the future, where story founders can tell how they will be able to execute better than anyone else and beat everyone else in the market. This is what mainstream venture capitalists believe will propel these startups into solid $50 billion-plus companies, or at least to some sort of profitable exit.

“People are just trying to survive the stress,” Levine said. “Otherwise you won’t have enough money to grow and actually compete.”

The good part about raising a lot of money in the early stages as a founder is that it helps the company move quickly and hire expensive talent. As they price their term sheets, VC firms realize that talent in the AI ​​era is expensive, as is running the AI ​​models that power these startups, competing with other well-capitalized competitors, sometimes large SaaS competitors that are already worth billions.

Everyone is trying to recreate the magic of Google buying Wiz, Levin said. But the risk is also higher. Founders must grow their companies into companies that justify high early valuations before they need more money. Series A investors also expect bigger, faster and more.

Nichols and his company are now securing more startups than ever before, with new expectations that they will reach their milestones in about 18 months. “This discipline is as important as supporting the winners,” he said.

Higher initial valuations mean less margin for error, Lear said, adding: “Less room for experimentation, less tolerance for pivots, and more scrutiny if progress doesn’t match the capital raised.”

Martin, the cybersecurity founder, successfully raised a Series A late last year, saying the standard was a no-brainer for his company to pass. But he also had a warning for the founders.

“You might end up stuck in between,” Martin said. “Very expensive for new investors, but doesn’t have the traction to justify the next round.”