5 VCs sounds off on the AI question du jour

Are we in an AI bubble or not?

The views seem to range from bubble-wary to bubble-dismissive. We hashed it all out over eggs and sausages at Fortune’s IRL Term Sheet Breakfast at Brainstorm AI in San Francisco yesterday. This is Amanda Gerut, Fortune’s West Coast news editor, pinch-hitting for my colleague Allie Garfinkle.

Allie hosted five VCs with funds ranging in size from $5 million to $25 billion and views varied across the panel. This group alone is collectively going to deploy anywhere from tens to hundreds of millions over the next decade into companies with AI as a backdrop and these investments will either prove spectacularly right or wrong.

Here’s a roll call:

Jenny Xiao, partner at Leonsis Capital and former researcher at OpenAI, came in with a nuanced take. There’s something of a bubble, but it’s “relatively contained” in the infrastructure layer with overinvestment primarily in data centers, GPUs and in large language model companies. But right now, there’s actually underinvestment in the application layer because there are so many ways AI can make an impact in various enterprises, Xiao said. 

Vanessa Larco, former partner at New Enterprise Associates (NEA) and co-founder of new venture firm Premise, has a contrarian view. “Everyone thinks enterprise is safer,” Larco said. “But I actually think the consumer might, this time around in the current environment, be what survives.” Larco’s reasoning is that if a consumer adopts your AI product, it’s because you’re giving them something faster, “radically cheaper, or much easier to use.” Once you’ve done that and built a brand, it’s very hard for people to quit you. 

Rob Biederman, managing partner at Asymmetric Capital Partners and chairman of Catalant Technologies, had a sobering view. “In every boom, 99% or 99.9% of companies fail, and one or two of them become Amazon or Google,” said Biederman, who had to dash off to catch a flight. Only companies that can systematically create value for customers, which most of them aren’t doing right now, will survive. 

Aaron Jacobson, partner at NEA, said the history of technological innovation “is always overhyped in the near term and underhyped in the long term, and that will be true of AI.” So at some point there will be a correction and there will be cycles of pain around valuation and funding, “but ultimately, in 10 years, we’re going to have a lot of really big, impactful companies.”

Daniel Dart, founder and general partner of Rock Yard Ventures, had the boldest counter to fears about a bubble. He sees a total addressable market we can’t yet imagine. People think self-driving Waymos will replace Ubers, but Dart sees elementary schools and elderly care centers with Waymos waiting out front and that proves to him we’re still in the early innings. 

“You’re really going to tell me there aren’t going to be any trillion-dollar companies in 2030 or 2034? No one here is going to take that bet,” said Dart. “There is going to be so much value creation that it’s like the birth of fire.”

See you tomorrow,

Amanda Gerut
Email: Amanda.gerut@fortune.com
Submit a deal for the Term Sheet newsletter here.

Joey Abrams curated the deals section of today’s newsletter. Subscribe here.

This story was originally featured on Fortune.com