How much of the market is AI?

Unlock the Editor’s Digest for free

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

Good morning. The rest of this week will be busy. Lots of consumer stocks report: Home Depot today, then Target, TJX, Walmart and BJ’s later in the week. Nvidia, the biggest company there is, reports earnings on Wednesday. And long-awaited delayed government data starts rolling in on Thursday, starting with September’s payrolls report. Buckle up, friends. And email us: unhedged@ft.com

Market breadth

Everyone is worried about stock market concentration — the outsize contribution of just a few stocks to profit growth and share performance. Concentration in general is not the worry just now; instead, it is concentration in stocks associated with artificial intelligence that has everyone jumping at shadows. At a high level, the worries are absolutely justified. Consider the stonking outperformance of the S&P 500 weighted by capitalisation against its equal-weight sibling over the past three years. The big stocks are doing a lot of the work:

This bears an unpleasant resemblance to what happened to the S&P 500 between 1998 and late 2000, before the gap closed to catastrophic effect:

Yes, we know, tech stocks had lower revenues and correspondingly higher valuations 24 years ago. But that only provides so much comfort in a world where AI infrastructure spending drives a big chunk of corporate revenue. Part of the worry about an AI bubble is that it might be a revenue bubble, not just a stock price bubble. 

Other ways of looking at market breadth are a bit less alarming. One standard measure is the number of S&P 500 stocks trading above their 200-day moving average. More than half of the index is currently trading above its 200-DMA, below the average of the past decade, but not near the lows: 

Out of the nearly 280 stocks trading above their 200-day moving average, about 240 names are not in the tech sector.

Another way of assessing market breadth is looking at how many companies in the S&P 500 are outperforming the index. This metric paints a more concerning picture, Kevin Gordon of Schwab pointed out to Unhedged. Year to date, only about a third of S&P 500 constituents are beating the index.

According to Gordon, that

. . . in and of itself it looks worrisome, and you could point to that and say it’s a negative development for the market. But at the same time, in this megacap concentrated world, it’s become more of the norm for fewer companies to outperform the index. Take 2023 as an example; we were seeing really low percentages of companies outperforming the index, and a lot of bears came out at the time and said it was a warning sign for the bull market to end. But, of course, that didn’t happen.

To focus more on the concentration in AI stocks, we tried — unsystematically and undoubtedly imperfectly — to pick out a list of stocks in the S&P 500 that are closely associated (mostly for better, but in a few cases for worse) with the AI boom. We came up with 17 names (data from S&P Capital IQ):

As of Friday’s close, the S&P 500 — not including dividends — has added about $7.5tn in value, or about a 14 per cent return. The names in the above list contributed about $4.9tn, or about two-thirds of that. That’s a lot! But looking in percentage terms, the other 483 stocks returned about 7 per cent this year. That’s in line with the long-term average return on the index. Without the big AI stocks, the market is hardly a wasteland.

Of course, the influence of the AI boom has extended, perceptibly but ambiguously, to other parts of the index. Many utilities and industrial stocks, for example, have received a boost from prospects for significant data centre construction and high energy demand. Other semiconductor and tech stocks have benefited obliquely, too. Are, say, Constellation Energy, Corning and Caterpillar AI stocks now? How about IBM, Uber or Cisco? We took 13 of these ambiguous cases out, too (all of them have had stonking years) and the remainder of the market is still up 6 per cent.

To sum up: yes, the market is very concentrated and that worries us. Without the AI narrative, returns this year would be much lower. But the non-AI bits of the market are OK, overall, which is, in a small way, reassuring.

Corrections

Two things went wrong in yesterday’s letter. One: we said that Nvidia’s valuation implied 8 per cent gross margins forever. A zero got lost; we meant 80. Two: we did some bad maths. We said that holding stocks through the 1970s would have left investors with a loss of nearly half their money in real terms. It would have been a 20 per cent real loss, as several numerate readers pointed out. These errors pain us and we apologise.

One good read

Unhedged’s career plan, if the whole market commentary thing doesn’t work out.

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.