A Wall Street Journal Report on $3 Trillion in Off-Balance-Sheet AI Commitments Recently Tanked Vertiv and GE Vernova. Here's What Actually Changed.

Most investors understand that technology giants like Alphabet (NASDAQ: GOOG) (NASDAQ: GOOGL), Microsoft (NASDAQ: MSFT), and Facebook parent Meta Platforms (NASDAQ: META) are spending a fortune on artificial intelligence infrastructure. What they may not fully appreciate is just how much money these companies have earmarked for AI infrastructure investments.

That's the big takeaway from recent reporting from The Wall Street Journal. Digging deeper into all of the industry titans' disclosure documents, reporters Peter Rudegeair and Peter Santilli found that artificial intelligence powerhouses like Amazon (NASDAQ: AMZN) and the aforementioned Alphabet collectively have an additional $3 trillion in AI-related liabilities -- like data center leases and technology purchase commitments -- that aren't reflected on their balance sheets.

For perspective on that number, the biggest names in the business are jointly budgeting on the order of $750 billion worth of infrastructure this year alone, and that's been viewed by investors as a jaw-dropping figure.

Sheer shock rattled shares of the companies implicated by the WSJ's reporting -- but not just those companies' stocks. Companies like GE Vernova (NYSE: GEV) and Vertiv (NYSE: VRT) that benefit directly from the massive AI build-out saw their stocks stumble in response to the news as well, and understandably so.

However, maybe shares of these ancillary outfits didn't actually deserve their knee-jerk punishment.

Investors know companies can't spend money they don't have

Off-balance-sheet obligations and liabilities are neither illegal nor immoral. These planned commitments shouldn't yet be on these companies' balance sheets, in fact, according to GAAP (generally accepted accounting principles). For the sake of complete transparency, these companies disclosed these additional future obligations in their most recent quarterly Securities and Exchange Commission (SEC) filings anyway.

What exactly are these liabilities that will eventually be moved to actual balance sheets in the future?

Some of them are commitments to future purchases of technology like AI-capable processing chips, data center networking solutions, or even the electricity that power-hungry data centers require.

Image source: Getty Images.

Another chunk of this $3 trillion worth of off-balance-sheet liabilities represents future leases of these data centers themselves. Many of these companies would rather rent access to them and walk away from a lease if need be -- even with a penalty for doing so -- than commit to the cost of outright ownership of a massive technology facility they may not want to actually own in the long run.

Some of the facilities that could potentially be leased in the future have yet to even be built.

That's where and why Vertiv and GE Vernova enter the picture. The former makes cooling solutions and power-management equipment for data centers. The latter makes onsite electricity-production solutions, including, most notably, natural gas power turbines. GE Vernova's orders soared 88% last quarter, largely due to AI data center-driven demand for power-production equipment. Vertiv's second-quarter sales grew 24% year over year, largely for the same reason. Both companies and their investors are looking for more of the same for the foreseeable future.

However, if that $3 trillion worth of off-balance-sheet planned spending never makes it to an actual balance sheet because it's canceled before being deployed, demand for Vertiv's and GE Vernova's wares could be upended in an instant.

What actually changed for Vertiv and GE Vernova?

Those are the dots investors are connecting, and to be fair, it's not an unreasonable concern.

For a handful of reasons, however, The Wall Street Journal's suggested number doesn't necessarily expose a new, potentially bearish problem for AI infrastructure players like GE Vernova and Vertiv.

One of these reasons is simply that -- while $3 trillion worth off-balance-sheet commitments is an admittedly huge figure -- it's not actually a shocking one.

Most investors understand that Big Tech's collective capital expenditure budget of $750 billion for 2026 is only the beginning of a multiyear spending spree of comparable annual amounts. And prior to the WSJ's reporting, a similar assessment published by Nikkei in late July put the artificial intelligence industry's off-balance-sheet liabilities in the same ballpark, at $1.65 trillion. Whether they readily realize it or not, The Wall Street Journal's calculation is within the scope of the amount that most investors have tacitly understood for some time now was going to be committed to investments in AI infrastructure. We now just have another specific working number, which initially jarred the market, but arguably didn't actually surprise it.

Another reason Vertiv and GE Vernova shares were arguably unduly punished by the WSJ's report is the argument that the earmarked $3 trillion is still very likely to be spent exactly how the artificial intelligence industry's top dogs say they're planning on spending it, for a couple of reasons.

One of them AI's newly proven value.

Despite its rocky start and revenue growth that's yet to keep up with its cost growth, there's a proverbial light at the end of the tunnel for the customers that "big tech" has been building AI platforms to serve. In its recently published "The State of AI in 2026" report, consulting firm McKinsey explains that enterprises' investments in artificial intelligence solutions are finally "on the road to ROI [return on investment]."

That doesn't mean all of it is paying off well enough yet. However, it does highlight that the latest iterations of AI tech and institutions' understanding of how to best use it are finally what was hoped for in artificial intelligence's infancy. Now that it's (reasonably) well-proven to add value, look for demand for AI solutions to pull that $3 trillion in off-balance-sheet commitments onto balance sheets with actual investments in actual artificial intelligence infrastructure.

The other reason this earmarked money is going to be spent regardless? While no agreement is entirely unbreakable, many of these off-balance-sheet commitments are indeed contracts that must be honored, or be resolved by sizable penalties or potential litigation, which can still result in high costs. Affordability or reason aren't really factors in the matter.

Perhaps more important to interested investors, although it's a dynamic that will take years to fully play out, their recent setbacks are all the more reason to step into GEV and VRT. Both are currently trading below analysts' current consensus price targets, by the way, and both are currently considered strong buys by the analyst community as well.

James Brumley has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet, Amazon, GE Vernova, Meta Platforms, Microsoft, and Vertiv. The Motley Fool has a disclosure policy.