The episode known in market circles as the “DeepSeek shock” was an unwelcome reminder: Capex isn’t something to be pursued just for the sake it.

Implicit in the concept of “investing for growth” is the assumption that eventually, the growth materializes. Otherwise, what are you doing? (Just burning money.)

The Mag7 C-suite would doubtlessly insist that tens of billions thrown at AI development is already paying off, and they’d show you some numbers to prove it. But as we learned last week, the mere suggestion that it may be possible to create a viable OpenAI competitor with far, far less in the way of resources and cutting-edge technology was more than enough to refocus investors’ attention on what it’s fair to describe as a vast disparity between the scope of big-tech AI outlays and recognized AI-related revenue, to say nothing of profits.

We talk frequently about the extent to which “US exceptionalism” in the market context is attributable almost entirely to the Mag7 these days, but a related, underlying point is that, as Goldman’s David Kostin noted, it’s “the magnitude of corporate investment” in the US “compared with firms in other countries [which] distinguishes the US stock market.”

Kostin calculates a “Growth Investment Ratio” as follows: Capex – depreciation + R&D / cash flow from operations. Not surprisingly, that ratio reveals that US companies invest far more than their global counterparts, as illustrated on the left, below, and the disparity’s growing more pronounced over time.

The figure on the right shows that US companies also tend to get more out of their investments than management teams overseas.

For obvious reasons, this discussion is more relevant than ever. If US exceptionalism’s tied both to greater corporate investment and superior returns on those investments, and if a large share of capex and R&D is attributable to Mag7 companies’ outlays for AI, then those outlays better pay off.

“Investing money is one thing. Generating an attractive return on the growth capital is another topic altogether,” Kostin went on.

The figures above give you a sense of the “investment gap,” as Kostin called it, between the Mag7 and the so-called S&P 493. That gap — some 20ppt — is “the true source” of the mega-caps’ “magnificence,” according to Kostin, who noted that those seven stocks are responsible for half — half! — of all S&P 500 growth investment.

Coming full circle, all that money spent (i.e., by the Mag7) better not be for nothing. Or even for less than what analysts are planning for in terms of future revenue.

Kostin was direct. “US equity market exceptionalism is not predestined,” he said, adding that “faith in US exceptionalism has been shaken following the recent announcement that China AI program DeepSeek replicated the performance of existing Made-in-America AI models for just a fraction of the cost.”