Analysis-Tech spending plans will test stock market's AI trade

By Lewis Krauskopf

NEW YORK (Reuters) -Spending updates this week from U.S. megacap companies that are investing massively in artificial intelligence could jolt the AI trade that has been a primary driver of the record-breaking U.S. stocks rally.

Capital spending is a crucial component of the AI trade that has propelled equities following the launch of AI assistant ChatGPT in November 2022. With the benchmark S&P 500 up about 17% year-to-date and 90% since the bull market began just over three years ago, investors are on guard for any weakness while wary of parallels with the dot-com market bubble 25 years ago. 

The rally has been led by a select group, including chipmaker Nvidia, which has symbolized the enthusiasm with its market-leading AI chips and mammoth share-price gains that have made it the world's largest company by market value, on the cusp of an unprecedented $5 trillion.

Investment by the biggest companies, including Microsoft, Alphabet, Meta Platforms, and Amazon, serves to build out the infrastructure that supports the burgeoning industry while signaling confidence in the theme, investors said.

"You want to still see that they are investing in these AI data centers and still continuing to build," said Don Nesbitt, senior portfolio manager at F/m Investments. "If you start to see them falter, then that is kind of a bellwether for the whole AI theme."

Annual capex from the four giants plus Oracle doubled between 2022 and 2024 to over $200 billion and is expected to soar to about $500 billion by 2027, according to Barclays equity strategists.

Wall Street will parse the companies' commentary in their quarterly reports, including the extent of their confidence in spending trends and any details about expected returns on the investments. Microsoft, Alphabet, and Meta report on Wednesday,  while Amazon reports on Thursday. Nvidia reports on November 19.

Still, investors are mindful of past capex cycles going bust and the possibility that the spending could result in overcapacity eventually.

"The key to determining whether this might be a bubble probably sits with the same sign we got in the late '90s, which is wasteful capital spending," said Jeff Buchbinder, chief equity strategist for LPL Financial.

SENDING AN AI SIGNAL

For now, investors said, Wall Street appears to want to see more AI spending, not less.

Investors see it as a signal of how much the companies are willing to commit and a reflection of the "ultimate market opportunity," said Nick Giorgi, chief equity strategist at Alpine Macro.

"If the hyperscalers pared back their spending, it's then more or less signaling that ... they may not believe in the actual monetizable potential of these technologies that they are investing heavily on building out," Giorgi said.

The surge in spending means that the five hyperscalers are spending 60% of operating cash flow on capex, a record amount, according to Torsten Slok, chief economist at Apollo Global Management.

Investors noted that many companies in the AI industry have stronger financial positions than the players caught in the dot-com bubble.

Free cash flow margins for the AI hyperscalers have averaged 15% in recent years, compared to 3.5% margins in the 1990s for telecom companies that built out fiber-optic network infrastructure "in anticipation of explosive internet demand that failed to materialize," investment firm Glenmede said in a recent report.

In much of the ensuing years until recently, the market had rewarded "asset-light" companies that did not need to implement major investments, said Michael Reynolds, vice president of investment strategy at Glenmede.

Now, investors are seeing "the script flip on that," Reynolds said. "This year is very much an outlier where those companies that are deploying capex at large scale are outperforming to a significant degree."

"It's getting to a scale where you have to make sure that the thesis on demand is coming through," he added.

RIPPLE EFFECTS FROM SPENDING

Microsoft, Google-parent Alphabet, Amazon, and Meta can sway equity indexes because of their huge market values.

Those companies, along with Apple - which also reports results on Thursday - Nvidia, and Tesla comprise the "Magnificent Seven," which collectively account for 35% of the weight of the S&P 500.

Changes to the hyperscalers' capex plans could reverberate widely. Citigroup strategists estimate that over 80 S&P 500 companies, comprising nearly half of the index's market capitalization, have significant exposure to AI.

Bespoke Investment Group recently found that 28 AI-related stocks accounted for about one-third of the $48 trillion increase in global market cap since the launch of ChatGPT.

AI capex spending is "driving the data-center build-out, which in turn drives a host of companies that are benefiting from the data-center infrastructure build out," such as industrial, construction and utilities companies, said Chuck Carlson, chief executive officer at Horizon Investment Services.

"The AI trade has so many tentacles."

(Reporting by Lewis Krauskopf; Editing by Alden Bentley and Rod Nickel)