Capital investment is meh

Unlock the Editor’s Digest for free

Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.

Good morning. In case you have not been following the news, there is an election today. And it appears that investors have finally accepted that the two candidates are neck and neck. The dollar and the 10-year Treasury fell yesterday as investors started to unwind their “Trump trades”. Let’s hold hands and jump into the abyss of uncertainty together. Email us: robert.armstrong@ft.com and aiden.reiter@ft.com.

Capital investment

There is a broad Wall Street consensus, shared by many academic economists, that investment is going to rise in the medium to long term, globally and in the US. The world needs a new, green power grid; the AI revolution requires massive infrastructure; governments are enthusiastic about industrial policy. This means higher investment, higher rates and higher growth.

At the same time, there is a view that corporations in the US underinvest, spending their cash on profit-flattering share repurchases or acquisitions designed to reduce competition, rather than on productivity-increasing investment. 

The two views are not directly opposed. Government and private investment must be distinguished. Within private investment the picture varies by sector. No one accuses the Magnificent Seven tech stocks of underinvesting, for example — on the contrary (see below). But the situation bears some looking at. Is corporate America investing more?

Here is capital investment, by category, as a portion of GDP over the past 50 years or so. Investment in buildings has been stable relative to the size of the economy for 30 years. Spending on equipment has been falling by fits and starts. Spending on software and intangibles has steadily grown. Put the three together and capital expenditures are stable at about 12-14 per cent of GDP over the long term.  

Capex is not a huge part of the economy. It would have to grow very fast indeed to noticeably change the rate of GDP growth. The issue is whether investment is growing fast enough to support increasing productivity. And you will notice to the right of the chart that in the past few years, investment is not expanding faster than the economy — not even software investment.

Recently, there has been only one growth story in investment, and it has been equipment. Here is year-over-year growth in the three investment categories. See how the light blue line has perked up this year:

It’s a pretty good guess that the rise in equipment investment comes down to hardware for AI data centres. Some evidence of that comes from the capex at S&P 500 companies. S&P capex surged from 2021-23 but is flat over the past 12 months, except for the Magnificent Seven, which are charging ahead with AI investment:

The broad picture is pretty clear. Corporate investment is not growing as a portion of GDP. The only type of investment that has been accelerating recently has been equipment, and that looks to be an AI effect. The same pattern is evident at large companies: investment is stagnant this year, except for Big Tech’s AI budgets.

The Mag 7: high expectations, meet high investment

Last week, Meta beat Wall Street’s earnings expectations by a comfortable margin. But its stock fell and has been trading poorly since. The culprit was AI-related capital expenditures.

Capital expenditures in the quarter were more than $9.2 billion, and according to CEO Mark Zuckerberg, the company “continue[s] to expect significant capital expenditure growth in 2025”. Zuckerberg acknowledged that news of higher infrastructure investment “is maybe not what investors want to hear in the near term”. But executives were vague about how and when all the capital expenditure would transform into revenue.

Meta is feeling a squeeze that other members of the Magnificent Seven may experience soon. The stocks are expensive, and investors expect astonishing levels of revenue and earnings growth every quarter. At the same time, the investment requirements for staying in the AI game are tremendous, and the spending is not yet generating much revenue and profit. Unless AI products start bringing in money soon, this dilemma is only going to become more acute.

Microsoft and Apple also beat earnings expectations last week, and also saw their stocks slide. Microsoft CEO Satya Nadella said the company expected capital expenditures, already double what they were last year, to increase going forward. He attributed the growth in Azure revenue to AI capabilities. But investors seemed unsure. According to Gregg Moskowitz at Mizuho, “most of the Azure F1Q upside was driven by greater than expected in-period revenue recognition”, and Azure forward guidance was below expectations. Apple, whose AI capex is small by comparison, but rising, gave cautious forward guidance, as the use case for its Apple Intelligence AI system remains unclear. 

Microsoft, Amazon and Google all sell computing capacity to other companies who want to run AI applications. Nvidia sells AI chips to everyone involved. But “selling shovels in the gold rush” pushes off the revenue problem rather than solving it. Someone has to pay to use AI if the AI economy is going to keep churning, and AI stocks are going to keep rising. With the market riding high, those who are confident that AI will be a money spinner need to worry about when the spinning will start.

(Reiter)

One good read

The life of Q.

FT Unhedged podcast

Can’t get enough of Unhedged? Listen to our new podcast, for a 15-minute dive into the latest markets news and financial headlines, twice a week. Catch up on past editions of the newsletter here.