News & Analysis

driven by the PitchBook Platform
AI_Bubble_Header Image.png

Jenna O’Malley/PitchBook News

Weekend Analysis

The AI bubble was leaking air at Milken in Beverly Hills

If the chatter at the financial world’s flagship conference is any indication, AI couldn’t be any frothier right now.

AI was the obsession du jour at the 2024 Milken Institute Global Conference in Beverly Hills this week. Investors—fed up with the extortionate going rate for pre-revenue and even pre-product startups—grumbled about a bubble.

“It’s like people didn’t learn anything from the last cycle,” an investor at a fund-of-funds told me.

“Valuations have just gotten insane,” vented a growth-stage investor over lunch.

Money managers sang a different tune on stage. Even as GPs sounded off on the sidelines, investors and allocators universally touted the technology’s transformative potential across three days of panels.

An annual gathering of some 5,000 investors, allocators, bankers, celebrities and journalists, the Milken Conference is a stage for the world’s financial and political elite to pontificate on the state of the markets—and behind closed doors, for GPs to court allocators.

This article appeared as part of The Weekend Pitch newsletter. Subscribe to the newsletter here

The unofficial theme of this year’s conference was artificial intelligence. Ten panels had AI listed in the event title, and many more veered into a discussion of it.

“Like the internet becoming commercializable with the browser, these discrete things spawn a whole new way that the world will be. We think this is one of those points,” said Ravi Mhatre, managing partner at Lightspeed, referring to OpenAI’s launch of ChatGPT-4.

Lightspeed is one of the most active early-stage AI investors, closing over 40 deals in the space in the past year, according to PitchBook data. “Any business that has a technology element is going to be disrupted, and quickly,” Mhatre said.

Median pre-money valuations for early-stage AI startups jumped 51% from $47 million in 2023 to $71 million in Q1 2024, compared to a 29% increase across all verticals, according to PitchBook’s Q1 2024 US VC Valuations Report. Late-stage AI valuations appear to be accelerating even more rapidly, and 35% of unicorn rounds are now AI companies.

Even startups that have little or no revenue are courting deals at significant valuations.

What’s wrong with it?

The AI hype and corresponding flood of dumb money is complicating the relationships between GPs and the founders they’re advising and prospecting. Many investors have voiced the same conundrum to me: They’ll advise founders to take capital at the right price, not the highest price—but for founders, that 9- or 10-figure valuation is intoxicating and difficult to resist.

So now when you, as a GP, actually see a startup that’s priced correctly, your immediate reaction is: What’s wrong with it?

How much value creation the generative AI boom brings venture GPs—and the allocators who commit their capital—is still an open question. “What we have to watch out for more than anything else, as early-stage investors, is: What is OpenAI just going to do themselves?” said Rebecca Lynn, co-founder of Canvas Ventures. “How well will they do it and where will the white space be? We’re really not sure.”

OpenAI itself made a Milken appearance: COO Brad Lightcap was predictably bullish on the technological advancement still to come. “I think we will look back in a year and realize how laughably bad [AI systems] were,” he quipped.

OpenAI is reportedly preparing to launch a search function to compete against Google Gemini and Perplexity AI’s search product. Perplexity, an AI chatbot developer backed by Daniel Gross, Nvidia and IVP, nearly doubled its post-money valuation from $520 million to $1 billion between November 2023 and March of this year.

The collective ability of large enterprises to make strides in generative AI and outcompete startups looms large in GPs’ minds. “I think you have to pay very close attention when you’re investing to what’s already out there,” Lynn said.

The overwhelming majority of AI startups on the market will eventually stall or collapse. As Elad Gil put it: “99% of the companies are going to be awful, overvalued and in many cases, fail. Everybody, including myself, we’re all investing in the 1% and we’re probably wrong.”

Time will tell which models stand up to commercial pressures and which use cases stick, as large enterprises simultaneously pour in research and development dollars.

In the meantime, there’s plenty of money going around to burn—especially if the Bombardier private jets pop-up at Milken is anything to go by.

Jenna O’Malley/PitchBook News

  • rosie-headshot.jpg
    Rosie Bradbury is a reporter covering startups and venture capital for PitchBook News. Based in New York, she previously reported for the Bureau of Investigative Journalism, Business Insider and Wired. Rosie studied history and politics at the University of Cambridge.
Join the more than 1.5 million industry professionals who get our daily newsletter!

I agree to PitchBook’s privacy policy