Investors’ excitement for the AI boom increasingly comes with a caveat: the risk to their portfolios—and the global economy—if the bet doesn’t pay off. Questions of an AI bubble were somewhat inevitable (just look to the dotcom era) but have ramped up in recent months as new data captures the eye-watering capital expenditure in the sector.
Valuations are being pulled into the limelight as a result (the S&P 500 currently sits at a P/E ratio of 31.50) with bearish analysts worried share prices have spiraled beyond their true potential.
At the Fortune Global Forum in Riyadh on Monday, Mary Callahan Erdoes, CEO of JPMorgan’s asset and wealth management business, argued some AI stocks have “a little too much concentration.”
But that doesn’t detract from the potential of AI technology itself, she said. “Everyone says: ‘Is there an AI bubble?’ That’s like saying: ‘Is there a computer bubble?’ That’s a crazy question.”
She explained: “AI has not even been deployed anywhere to the extent that it will be. Less than 10% of companies actually say that it’s embedded in the services and the products that they deliver today. There’s an enormous amount of opportunity.
“That’s why you’re seeing the multiples are the way they are,” Callahan Erdoes added. “And the question is, How fast will we grow into those multiples? It’s not that the multiples are wrong, they will eventually be right; they may not be right for every company.”
The true benefit of AI is still some way off, when businesses reimagine their ways of working and the efficiencies available to them. Data from Morgan Stanley supports the idea: A September report found full adoption across S&P 500 companies could add up to an annual net benefit of $920 billion.
“In the long term, this potentially translates to a market cap increase of $13 trillion to $16 trillion for the S&P 500 alone, or 24% to 29% from the current level,” the report adds.
Much of the consternation around an AI stock bubble is directed at the U.S. market, said Tan Su Shan, director and CEO of Singapore-based DBS Group. She added that multiples for tech stocks in Asia tend to be lower—Japan’s are at 12 to 14, for example—so “there is an AI arbitrage, I guess, between Asia and the U.S.”
AI use cases in Asia also differ from the priorities of the U.S., she highlighted, focusing on small language models and a blend of hardware and software as opposed to large language models (LLMs). She added: “Thank God the U.S. market is open, so global clients can participate in the U.S. market and take advantage of the momentum there.”
