AI spending, inflation, and 3 more reasons why tech is hurting

AI companies—surprise, surprise—are just like us. Even deep-pocketed Silicon Valley players can have money problems.

Tech has been dropping for over a month under the weight of a host of worries. And given that AI is the engine driving the market—for the past few years, in fact—tech’s troubles are everybody’s troubles.

Luckily, this too shall pass.

To be clear, there hasn’t been anything new that is sabotaging the AI trade. No, this agonizingly slow, tortured drawdown is because of five issues—none that haven’t been talked about plenty but laid out well by Sevens Report’s Tom Essaye.

1. AI spending. Tech has done a 180 and become a capital intensive sector, spending hundreds of billions this year alone. The amounts are so massive that even the biggest companies, from Oracle to Meta Platforms, are financing their capex plans with debt.

Just three years ago, when Silicon Valley Bank’s problems kicked off a market selloff, tech companies were nearly immune—in part because of their huge war chests. That can’t be said of all those same players today.

2. Inflation. The gold rush has led to a problem that all American families are all too familiar with. Plenty of companies are scrambling to provide the “picks and shovels,” but there is so much demand that they can’t always keep up. In response, “the richest companies in the world (the major tech companies) have started a ‘spending war’ with each other to secure components,” writes Essaye. “That spending war has caused the prices of previously commoditized goods such as semiconductors and memory to skyrocket, which is feeding back into the market’s first concern, which is all of this is costing too much.”

Just as a little inflation is a good thing—the Fed’s annual target rate is 2%, after all—this metric is a must to thread a tight needle.

As BCA Research’s Peter Berezin writes: “If AI spending were to slow, margins would decline for many companies currently benefiting from the AI boom. Conversely, if the AI boom intensified, the economy could start to overheat.” And given that “economywide margins generally decline in the lead-up to recessions,” again, that’s an everyone problem.

3. Share supply. From the earliest debt issuances, worries have surfaced about demand for that debt, and now there’s a glut of equity offerings as well, Essaye notes. ChatGPT parent OpenAI and Anthropic are expected to go public later this year, on the heels of the volatile debut of SK Hynix’s American depositary receipts and SpaceX’s record-breaking initial public offering.

4. Competition. It’s definitely increasing. The launch of Kimi-K3, the latest AI model from the Chinese company Moonshot AI, was one of the few pieces of concrete tech news that sparked selling, reigniting fears kicked off almost 18 months ago by DeepSeek. Will cheap Chinese AI really undercut American companies—and their stock prices?

5. Capacity. Despite growing opposition from the public, data centers are popping up faster than mushrooms after a rainstorm. Some on Wall Street even wonder out loud whether they’re needed at all. Any number of headwinds could emerge, from companies turning to cheaper Chinese models, or not adopting AI as quickly as hoped overall.

Yet consensus estimates depend on swift AI uptake from big U.S. providers. Free cash flow for hyperscalers is expected to more than double over the coming years, as Apollo chief economist Torsten Sløk has pointed out, and if the AI payoff comes more slowly than expected, there are three major knock-on effects: “Cash flows and earnings disappoint…Balance sheets stretch and credit risk rises,” and a Mag7 selloff brings the rest of the market down with it, he writes. Again, recession could soon follow.

Of course, tech has weathered worse, and betting against the AI trade has never been a good long-term move in this bull market.

The sector was up on Tuesday, and though time will tell if it’s a blip or an all-clear signal, the reality is that normal markets need a shake-up once in awhile. They don’t go straight up all the time.

“[T]here is some validity to the ‘it’s all too much’ concerns, and while that doesn’t mean AI is in a bubble, it does mean a period of digestion and consolidation (and restraint in AI spending) may be needed for the longer term health of the rally,” Essaye concludes.

Tech bull Jessica Rabe agrees the party isn’t over, but might start to have different decorations.

“U. S. Big Tech’s performance edge over the S&P 500 has averaged 30.5 percentage points in any given rolling 1-year holding period since 2015,” writes Rabe, co-founder of DataTrek Research.

“However, each peak has been smaller than the last in recent years as the group shifts to funding AI capex with debt on top of free cash flow. With the group’s relative returns near the zero line again and well below average, we expect Big Tech to perform better from here, but more modestly than in the past.”

Hmmm. Will investors be happy with those kinds of gains?

Write to Teresa Rivas at t[email protected]