These two charts on the AI spending boom have something for everyone

You’ve got to spend money to make money, or so I’ve been told.

Cliches become cliches because they simplify a truth we tend to see many examples of, like, for instance, megacap companies pouring hundreds of billions of dollars into their AI capabilities.

We are currently in the midst of a week where the market narrative could loosely be characterized as, “The AI capex boom is taking over the economy — and rightfully so!” based on economic data and high-profile earnings reports. Well, we have a pair of charts on the topic for your consideration.

First, let’s see how sales and capex for the hyperscalers (Alphabet, Amazon, Meta, Microsoft, and Oracle) have increased (in dollar terms) over the past four quarters:

Simply, you cannot look at this chart and reach an obvious “malinvestment” conclusion. The capital stock has a useful life beyond one year, and it’s been helping to juice annual sales to the tune of well over $150 billion.

On the other hand, you can observe some flattening out in terms of how much sales are increasing, to say nothing for the rate of growth. But of course, that’s for total revenues. Some of the most AI-sensitive parts of these companies’ businesses (like Microsoft’s Azure) are enjoying accelerating growth. In addition, rising capital intensity across an industry can be worrisome in and of itself. To quote the famous statistician and trader Nassim Taleb (again): “I’ve seen gluts not followed by shortages, but I’ve never seen a shortage not followed by a glut.”

Next, we can take a peek at how 12-month forward expectations for sales and revenues have evolved:

You can look at this chart and say, “80% of anticipated sales growth is expected to be plowed right into capex. When can I haz shareholder returns? Oh, and about that point on the aftermath of capex booms...”

Of course, there’s also no guarantee that expectations will be met, and companies may not be particularly nimble in dialing down investment in the face of slowing demand, particularly if the scope of the opportunity is as game-changing as management teams believe AI is and will be.

Or, optimists could observe that after clearing a few hyperscaler earnings, we’ve seen a nice inflection higher in 12-month forward sales expectations, indicating increased faith in these investments bearing fruit.

There was a point in time shortly following the market freak-out in April where analysts were on the verge of expecting that capex, in dollar terms, would rise by more than sales in the year ahead. That’s no longer close to the case.

Traders would be a lot more nervous about AI investments if they had to wait for them to pay off. While a full accounting of the ROI on all this spending will take time and be incredibly difficult to tease out, it certainly helps that there are early returns, and more expected to come in the near future.