In the AI arms race, quick gains could just as easily lead to quick losses.
“Great fortunes will be made. Great fortunes may be lost,” Apollo Global Management CEO Marc Rowan told Yahoo Finance. (Disclosure: Yahoo Finance is owned by Apollo Global Management.)
Rowan warned of intermediary companies, such as specialized real estate firms that build and own the data centers that are used by tech giants. There’s a distinction between lending money to data centers and owning a stake in them, he said. For credit investors like Apollo, the current demand for data is a safety buffer.
“As a credit investor, you want to get paid back before you get to renewal, because it is very clear today and for the next few years that the demand for data is unbelievably strong,” he said, referring to the non-negotiable leases signed with “Magnificent Seven” companies.
“Therefore, for the next period of time, based on contracts going in today, we should have very safe cash flows in which to amortize credit.”
But while lenders are focused on the leases — typically lasting 10 to 15 years — equity investors are making a longer-term bet. The current demand landscape has to hold up for the leases to be renewed next time.
Read more: How to protect your portfolio from an AI bubble
If a tech giant decides not to renew, owners could be scrambling for new deals or be left with buildings with no client.
“Ultimately, the return on equity, though, is a function of renewal,” he said. Rowan also noted there are “a whole host of factors” that can change the future economics of data centers, like chipsets, energy uses, and quantum computing.
So while the AI boom is real, the investment in data center companies can be speculative, especially since they don’t have diversified businesses like the hyperscalers.
“I don’t think it is a foregone conclusion that because you own a data center and you own equity in it, and data is valuable, that you will make money in data center equity,” he said.
Though Wall Street has also raised concerns over hyperscalers like Amazon (AMZN), Microsoft (MSFT), and Google (GOOG) using debt to fund AI infrastructure expansion, Rowan said he’s “not worried” about the tech giants.
They are the “largest and most successful companies that exist … [with] massive amounts of cash flow,” he noted.
Rowan argues the danger lies with the highly leveraged companies that serve the hyperscalers, such as specialized cloud providers, data center real estate investment trusts, and niche hardware firms. These AI infrastructure companies rely on tech giants’ funding, making them beholden to contracts and financing.
