Fly or Fall: Big Economic News Coming

“Greed … is good. Greed is right. Greed works. Greed clarifies and cuts through and captures the essence of the evolutionary spirit. Greed in all its forms, greed for life, for money, for love, knowledge, has marked the upward surge of mankind”

- Gordon Gekko

Wednesday is going to be a big, almost fateful day for the U.S. economy. The Bureau of Economic Analysis releases its second, more comprehensive estimate of GDP for Q2. That estimate will include corporate profits, which—given the current nervousness surrounding Big AI—will be a big piece of news in itself.

On top of that, Nvidia releases its Q2 report, which some analysts predict will cause the NVDA stock to fall by up to 20 percent.

The fate of the stock itself is one story; the macroeconomic repercussions are another. And yet, the two are intertwined: capital formation in the AI industry, where Nvidia is a major player, is growing so fast that next year it could account for 20 percent of all capital formation in the U.S. economy. This means that Big AI is gobbling up investment financing that could go to other industrial sectors, thus weakening the U.S. economy generally at a point where it needs the opposite.

It also means that the fate of Nvidia and other Big AI corporations is increasingly tied to the fate of the entire U.S. economy—and vice versa.

Media is of course abuzz over this; looking at their anticipatory conversation, overall expectations lean pessimistic. The aforementioned predictions of a possible 20-percent drop on the stock market is the wrapping paper; the package of expectational substance is built around a growing realization that Big AI’s revenue is not catching up with its rocketing costs for production, capital formation, and R&D.

Just look at the expectations of how Nvidia’s earnings need to evolve over the next couple of years in order to keep financial analysts happy. Their earnings per share, EPS, were at $2.09 for Q2 this year; compare that to $15-20 EPS for 2027-2028:

"Because many of the debates around AI infrastructure spend/return on investment and credit risk are out of NVDA's hands, we think the numbers are more important than the narrative and coming out of this call, we expect investors to gain greater confidence in a path to $15+ EPS in 2027E and $20 in 2028E — numbers that should keep the stock grinding higher," UBS analyst Tim Arcuri said in a note.

Let’s be clear about what analysts mean by ‘expectations’. When they say they expect a stock to deliver a certain EPS, or for that matter the stock to reach so-and-so values, they are not sharing objective estimates based on tedious, dispassionate examination of tons of statistics. They are simply stating what the company in question must achieve in order for the analyst and his employer to still recommend that their customers keep buying the stock—or even keep the stock at all.

Although the analyst cited above does not say so openly, his statement is nevertheless a demand that Nvidia deliver a seven-to-tenfold increase in its EPS over the next two years. This demand should be understood against the backdrop of Nvidia’s earnings per share over the past four quarters, which have grown quarter-to-quarter at an average of 33.6 percent.

That is an aggressive growth rate. If they could keep it up, they would handily meet those EPS expectations. The question, of course, is if they can do it. There are three reasons to believe they won’t.

Why is this important? Because if there is a prevailing expectation among analysts and portfolio managers that the center stock of Big AI must up its EPS at least seven times in 12 months, and they fall short of doing so with even a moderate margin, there will be a big sell-off of their stock.

If that sell-off happens, we can expect a major setback on the stock market as a whole. We can also expect shockwaves through the financial industry and repercussions for the dollar. At that point the bubble-bursting process sets a fiscal crisis in motion. In a portfolio-saving effort, investors will sell off U.S. sovereign debt, in which they are already losing faith.

How probable is this scenario? I don’t use numbers when I discuss probability, but when I say something is probable, it means it is more likely to happen than not to happen. Here are three reasons why the first link in this chain of events from Big AI revenue shortfalls to a fiscal crisis, is probable.

First, we have the growing resistance to data centers, which is going to raise the marginal cost of adding another center. Sooner or later, that marginal cost will be prohibitive. Big AI depends heavily on expanding data-center capacity to grow its data-service production—and its revenue.

Secondly, there is growing competition from other chip makers—Broadcom is one of them—who will inevitably nibble away at Nvidia’s dominant market position. Competition forces a reorientation of Nvidia' corporate leadership, from sales expansion to lean-cost production reforms. These two are not mutually exclusive, but the latter puts heavier demands on the former: once corporate leadership starts slimming its organization, people within it start protecting their turfs in order not to lose their jobs. When that happens, they lose—at least to some degree—focus on what makes the company grow and thrive.

I don’t want to exaggerate this point, but it needs to be on this list, especially if Nvidia’s Q2 report is more than a little disappointing.

Third, the chips that are at the AI hardware epicenter are becoming prohibitively expensive. The industry has already resorted to lease programs for their buyers, allowing them to ‘rent’ computer chips for a limited period of time. In effect, this is a subscription program: you pay monthly to have a chip maker supply you with the latest-tech chip and then replace it in, say, two years at no other cost than the regular monthly fee.

To us Joe Sixpacks out there, leasing a computer chip seems insanely complicated, like leasing the engine in your car. But it is a financing product for buyers of computer chips that has emerged in a free-market environment, and there was clearly a need for it. The problem is that not even with this solution in place, chip prices are still outpacing chip buyers’ financing ability.

Normally, at this point in a market the producers and sellers would scale back to avoid driving prices lower. Big AI with Nvidia at the helm is going to try something else, a financing scheme for their buyers that the Wall Street Journal refers to as and “exotic money pipeline”:

Nvidia CEO Jensen Huang is running into a problem: Many of his customers can’t afford to buy his company’s coveted AI-powering chips. That explains why Huang teamed up with an array of Wall Street firms on a $500 billion plan that will theoretically standardize chip financing, creating asset-backed pools of capital for AI companies—while leaving Nvidia partly on the hook if things go wrong.

A pool backed why what assets? The computer chips that Nvidia produces.

Yes, the chips themselves are going to be an asset class of their own. The idea that Nvidia is apparently spearheading is to sell bonds on the corporate bond market with Nvidia’s (to use them as the example) own product as collateral for the bond.

Those chips are already part of the collateral for Nvidia stock and for Nvidia corporate bonds. Making them collateral for a third financial product would be like Ford issuing bonds based on its F-150—a comparison that puts the finger on the absurdity with this idea. Or, in the words of the Wall Street Journal:

The AI boom, and Nvidia’s financial momentum, can’t keep up at this blistering pace unless those kinds of companies get the hardware they need.

Nvidia would use the money raised through this third-rail borrowing scheme to give loans to its own customers. Nvidia and other AI conglomerates have a history of engaging in this type of circular financing:

One company injects capital into another, who then turns around and funnels that capital into the original investor’s chips, cloud contracts or data center buildouts. Known as circular financing, these “round-trip” transactions are shaping the entire AI industry — and, by extension, the global economy.

The danger? The Built-In article puts it bluntly:

At a system-wide level, any mismatch between expectations and reality can trigger broader instability, with one failure cascading through interconnected contracts, valuations and credit markets — the kind of chain reaction that fuels concerns about a trillion-dollar AI bubble.

Interestingly, CNBC warned about exactly this already back in October. Now that Nvidia is leading the proposal of a new AI-industry circular-financing model, where they would borrow money for the third time on the same asset, it can only be interpreted as a sign that their revenue cannot catch up with their costs. When that realization sets in, all the speculative hot air that has gone into the Big AI stocks will burst—and creditors will come knocking on the door to get their money back.

Long story short, it looks increasingly as if Nvidia will be unable to reach the EPS goal that analysts demand. Although no analyst has said openly (as far as I have been able to tell) that Nvidia must reach $15 EPS next year or he will put a sell recommendation on it, I read the entirety of the news buzz around the company as precisely that. If that does indeed happen; if Big AI as a whole loses just enough steam to deflate a speculative bubble; the repercussions for the U.S. economy will be significant.

How significant? Let us get back to that when we have the revised Q2 GDP plus corporate-profit figures from the Bureau of Economic Analysis on Wednesday.