- Morgan Stanley’s Lisa Shalett says that debt-funded AI projects, like Meta’s $30 billion Louisiana data center, are a signal that the narrative driving the bull market in tech stocks is becoming more complicated. She says the new reliance on private credit, the rising complexity in AI vendor ties, and the uncertain profitability of AI startups, increases the pressure from investors to see actual returns. “What was a very simple story is suddenly getting a lot more complex,” she says.
Meta’s $30 billion debt-financed deal for a gigantic AI data center in Louisiana is an example of how “the landscape has suddenly gotten a lot, lot, lot more complicated” for tech stocks going forward, Morgan Stanley Wealth Management chief investment officer Lisa Shalett told Fortune.
The Meta agreement is the largest private debt deal ever, according to the Wall Street Journal. Eighty percent of the Hyperion data center in Richland Parish, La., will be owned by Blue Owl Capital, with Meta retaining only a 20% stake. The site will technically be owned by a special-purpose vehicle, and thus will not appear on Meta’s balance sheet. Morgan Stanley was the bookrunner that put the deal together, according to Bloomberg.
The deal is a departure from the way AI has been funded in the past, Shalett said. Previously, investment came directly from the cash on big tech companies’ balance sheets. Now, with off-balance-sheet debt in the picture, companies will come under increased pressure to show a return on their investments, she said.
“In the first phase, in the first three years, Zuckerberg was building everything with cash on his balance sheet, with free cash flow,” Shalett said, referring to Meta founder Mark Zuckerberg. “When you start using debt, and you use debt in the shadow banking market, meaning you’re partnering with private-credit players where you’re using private wealth money, etc., etc., it just gets harder to track, and it gets more and more unwieldy, and that creates pressure to actually earn that return on investment.”
Shalett also expressed some skepticism about the high level of interconnectedness between AI companies and their vendors. Earlier this month, Morgan Stanley analyst Todd Castagno and his team produced a diagram of what he called the “increasingly circular” AI ecosystem:
“In the last 30, 60, 90 days, the landscape has suddenly gotten a lot, lot, lot, more complicated,” Shalett said. “The deals and the cross-dealing have gotten more and more and more complicated, where some of this starts to feel and look and smell like circular dealing, like vendor financing. And I say this not because I think anyone is doing anything nefarious. I don’t. But what I’m seeing is what was a very simple story is suddenly getting a lot more complex.”
Nonetheless, Shalett believes stocks will “grind higher … but we don’t think it’s going to be this reaccelerating boom.”
Along the way she warned traders to watch out for an “accident” in AI. For instance, “the accident could be that somehow OpenAI doesn’t actually develop a true revenue model to pay for all this capacity that they’ve committed to buy from everybody,” she said.
In that case, a 10% to 20% correction in the S&P 500 might be on the cards, she said.
“Is [generative] AI not ultimately going to pay off? It probably will ultimately pay off. But the path will not be a straight line,” she said. “We’re not trying to beat up on the story. We’re not trying to say, ‘We’re entering a bear market,’ any of that. We think that this has legs, but we think that the legs are getting weaker and weaker and weaker as the days go by.”
Meta and OpenAI were both contacted for comment.
Here’s a snapshot of the markets ahead of the opening bell in New York this morning:
- S&P 500 futures were up 0.62% this morning. The last session closed up 0.58%.
- The STOXX Europe 600 was flat in early trading.
- The U.K.’s FTSE 100 was flat in early trading.
- Japan’s Nikkei 225 was up 1.35%.
- China’s CSI 300 was up 1.13%.
- The South Korea KOSPI was up 2.5%.
- India’s NIFTY 50 was up 0.08%.
- Bitcoin is up at $111K.
