US start-ups are raising more cash than at any point since 2021 thanks to investor bullishness about artificial intelligence, but the venture capital market has tilted sharply towards funding a handful of huge private tech companies.
More than $30bn has been invested into fledgling groups already this quarter, according to PitchBook data. A further $50bn of fundraising is also in train, as venture capitalists work on a series of major deals involving OpenAI, Safe Superintelligence and defence tech start-up Anduril.
The fervour over AI has led investors to spend at their fastest rate since the market’s peak in 2021, a period in which $358bn flooded into tech groups, saddling many with unrealistic valuations.
But VC groups believe this investment cycle will be different. “AI is a transformative force that makes these companies better,” said Hemant Taneja, chief executive of General Catalyst, one of Silicon Valley’s largest venture firms.
“The way to think about it is ‘can these businesses reasonably grow 10x from where they are?’ The answer with all of these is yes, so they are reasonably priced,” he added.
After a two-year slump, US fundraising leapt to about $80bn in the last quarter of 2024, according to data from PitchBook. That represented the best fourth quarter since 2021. But just six large deals — involving OpenAI, xAI, Databricks and others — accounted for 40 per cent of that total, said Kyle Stanford, director of research at PitchBook
“It’s a very elite group of companies that are commanding the VC investment,” he added.
On the basis of deals already closed and those expected to do so in the coming weeks, the first quarter of this year is set to see similar levels of investment — which would make it the best first quarter for fundraising since 2022.
In the past two weeks alone, fintech companies Stripe and Ramp have announced funding rounds at valuations of $91.5bn and $13bn, respectively, and AI start-ups Anthropic and Shield AI have inked deals at $61.5bn and $5.3bn, respectively.
VCs are also working on a series of massive investments. OpenAI is in talks with SoftBank to raise $40bn at a $260bn valuation, which would be the largest funding round ever, surpassing the $10bn investment into Databricks late last year.
Anduril, founded by Palmer Luckey, is in discussions to raise at least $2bn at a $30bn-plus valuation, more than doubling the valuation it achieved in a funding round last summer, according to two people with knowledge of the matter. Anduril declined to comment.
These more established companies have annual revenues in the hundreds of millions or billions of dollars and are growing fast. That makes them relatively safe bets, according to General Catalyst’s Taneja, who has backed Anduril, Anthropic, Ramp and Stripe.
“It’s so ambiguous where money will be made in AI, that lots of capital ends up concentrating into those companies that are category leaders with a customer base and large markets,” he said.
But the excitement over AI has also boosted younger companies with no revenues and, in some cases, no product.
Safe Superintelligence, launched last year by Ilya Sutskever, co-founder and former chief scientist at OpenAI, raised $1bn at a $5bn valuation in 2024 and is in talks to raise new capital at a valuation of $30bn or more, according to two people with direct knowledge of the deal. It has not yet announced a product. SSI declined to comment.
The huge funding rounds being carried out mark a significant departure from traditional venture capitalism, which targets nascent companies and is governed by the “power law” that states the best start-up in a portfolio will more than pay off the losses from the remainder that fail.
“We’ve always thought [a venture fund’s] 50x return will come from a seed investment which they exit at IPO,” said PitchBook’s Stanford.
In a largely untested experiment, that logic is now being applied to companies that are orders of magnitude larger and more developed by a new breed of what Stanford calls “pseudo-VCs”.
Those include Josh Kushner’s Thrive Capital, General Catalyst and Lightspeed Venture Partners, all of which have invested in more than one of the large rounds in recent weeks. All three firms are registered investment advisers, allowing them to invest in a wider range of asset classes and to hold companies after they go public.
Each of the three groups have also raised $5bn-plus funds, giving them “scale enough to invest in start-ups at a $1bn valuation and hold on for 15 years until its worth $50bn, investing in multiple ways along the way”, said Stanford.
According to Sebastian Mallaby, author of The Power Law, the conviction that even the priciest start-ups can still scale by 10 times is what “enables fund managers to rush in with high enthusiasm to marquee names and say ‘who cares what I’m paying? I’m a genius getting into this name.’”
While the chances of an established company failing are slimmer, Mallaby cautioned, so too are the odds its valuation will increase ten- or a hundred-fold. “The habits that worked really well in early-stage investing need to be adapted when you move to much bigger rounds.”
The large funding rounds in discussion today represented “a completely different style of venture than I’ve ever experienced”, said Stanford.
VC’s peak in 2021 was characterised by a rising tide of round sizes and valuations: there were about 854 deals of $100mn or larger that year, according to PitchBook. This year, total investment is tracking close to 2021 levels, but the market has become increasingly lopsided.
“If you’re OpenAI or Anduril — a high-growth, named brand — you are very well positioned. The money is there for you . . . If you’re on the other side, as most companies are, the money is not there,” said Stanford.
“Maybe it ends up at $80bn [raised this quarter], but $40bn of that is just one round . . . even the outliers in 2021 were minuscule compared to that.”
Additional reporting by Cristina Criddle in San Francisco
