At the DiscoveryX conference in Toronto last week, a trio of Canadian venture capitalists (VCs) couldn’t come to a consensus on whether the AI industry is currently in a bubble. They did, however, agree that the valuation math isn’t adding up.
On a panel moderated by BetaKit reporter Josh Scott, McRock Capital partner Udit Bhatnagar, IRV Fund managing partner Neha Khera, and Wittington Ventures partner Zeeshan Ali discussed how inflating valuations have changed the types of deals being finalized.
“I think the math is not mathing at this moment,” Bhatnagar said, noting that round sizes are getting bigger despite AI’s promise of being able to build faster and cheaper. “If it were true [that AI is] more frugal, that whole narrative is not tying up at some level.”
“I think we’re in a huge bubble, and I think it’s going to burst very soon.”
Neha Khera, IRV Fund
Bubbles are characterized by growing investment in overvalued companies or sectors that eventually burst, erasing stock market gains and sometimes spilling out into greater economic destruction. Generative AI use has exploded among consumers and has become a staple of startup strategy, with some comparing the industry’s mounting investment to the dot-com crash of the early 2000s.
At the heart of this is the unprecedented amount of money spent by AI companies on infrastructure and compute. This is reflected by Google, Microsoft, and Amazon reporting big gains in their cloud-computing businesses last week.
While Ali agreed that some AI investments are priced “extremely” and lack fundamentals, he said that he doesn’t believe there’s an AI bubble.
“I think there is tremendous room to grow, and I think we’re at the start of something [with true] potential in many different industries,” Ali said. Bhatnagar agreed, saying that while there is “overload,” AI is still very early in demonstrating its value.
RELATED: A “terrifying” time to build: Canadian VCs debate AI bubble at RevStar Summit
Khera, meanwhile, took the exact opposite position.
“I think we’re in a huge bubble, and I think it’s going to burst very soon,” she said.
Khera’s justification is that roughly 45 percent of the S&P 500 is reliant on AI-linked stocks, while even giants like OpenAI have missed metrics targets for new users and revenue, and are struggling with profitability. Khera asked what the solution for the company was other than raising another $100 billion.
“AI [and] LLMs have to become profitable business models, and that is not happening today,” Khera said. “I don’t want the bubble to pop, but I think this is not …sustainable.”
Those large rounds from players like OpenAI are “skewing the benchmarks” for early-stage AI startups, even though they’re not proving whether the problems they solve are big enough or if they have product-market fit, Bhatnagar said. Ali explained that this gravitational pull is changing who gets to play in what rounds, with traditional late-stage investors prowling in the early stages to get better value.
Zeeshan Ali,
“I think there is tremendous room to grow, and I think we’re at the start of something [with true] potential in many different industries.”
Wittington Ventures
“The multiples are getting pretty high, even at early stages,” Bhatnagar said. “You are taking way more risk as a VC, because now you’re only betting on the scaling of the company … we are just betting on a single variable.”
Khera said that to invest in these early-stage companies with high valuations, they have to believe the company will maintain its value and surpass the average exit values of companies in other sectors.
“If we think there’s going to be friction in the market [like a bubble burst], then the math just doesn’t add up,” Khera said.
She added that this has led to her experiencing “FOMO,” or fear of missing out, as she’s been forced to pass on multiple companies she sees as inflated in value.
“It’s hard when we walk away from companies over and over and over again because of that valuation being so high,” Khera said. “You’re like, ‘Darn I don’t want to miss it,’ but then your pragmatic side kicks in, and you’re like, I can’t do it.”
Feature image courtesy Ontario Centre of Innovation.