It’s an AI market now

Good morning. On Tuesday I asked readers if they could name examples of non-finance companies that had diversified into finance successfully. One reader bravely defended the record of GE Capital, despite its bad end. Others stuck up for the automakers’ big captive finance operations. Several noted that Starbucks basically takes billions in deposits in the form of payments for gift cards. Others looked abroad, to the big payment businesses built by Alibaba and Safaricom. Most interesting: the Manhattan Company was founded in 1799 by Aaron Burr, ostensibly as a water utility with a sideline in banking. By 1808 it was out of the water business, in 1955 it merged with Chase National, and is now part of a little bank called JPMorgan Chase. Email me: robert.armstrong@ft.com

How the market changed

It worries people that so much of the US stock market’s gains come from one company. Two years ago Nvidia was worth $400bn. Now it is worth $3.3tn and is the most valuable company in the world. If something should happen to the AI chipmaker, what will become of the stock market generally? The worry is justified, but the story is complex. 

I date the start of the current rally in the S&P 500 to late October of last year, when a three-month, 10 per cent drawdown came to an end. Since then the market has risen 33 per cent, increasing in value by about $12tn. Those gains can be neatly broken down into four buckets. Nvidia is alone in one bucket. The next contains the fabulous five tech companies (Microsoft, Alphabet, Amazon, Apple and Meta — sorry Tesla, you’re out). Then there is a group of 10 semiconductor industry companies that have gone bananas in the reflected glory of Nvidia and on hopes for a general digital investment boom. The final bucket contains the remaining 484 companies in the index. This chart shows the dollar contribution of each bucket to that $12tn gain in the S&P over the period:

One semiconductor company contributing almost 20 per cent of an index’s gains over an extended period is a bit unsettling. That another 10 per cent of the gains come from companies in the same industry makes it worse. Broadcom, Qualcomm, Micron, Applied Materials, Advanced Micro Devices, Lam Research, KLA, Texas Instruments, Analog Devices and NXP are up an average of 84 per cent since October (whether they all actually stand to benefit from the AI boom is not clear, at least not to me). Another 25 per cent of the gains come from the fab five, which all, to a greater or lesser extent, have risen on the promise of the same technology that has boosted Nvidia and the semiconductor companies. Some 56 per cent of the market’s gains are linked to AI, then. What happens if the market, collectively, decides that there is less profit in the AI business than is currently priced in? 

The reassuring thing is that, taking all of these companies out, the rest of the market has done very well. The other 484 companies in the index have risen in value by 20 per cent. They are having a big rally, too, just not as huge as when the AI mob is included. 

Here are the 10 largest contributors to the non-AI rally, along with their share price changes:

That is a nice mix of companies: entertainment, staples, banking, tech, home improvement, payments. Large, high-quality companies of many sorts have done really well since October. In that sense, this has been a broad rally.

It is not so simple, however, because the past eight months contain two distinct periods.

The broad rally continued only until late March. The thinning since then has been significant, as this chart of the S&P 500 market weight and equal weight indices shows. The equal weight index is down over the past 12 weeks: 

Looking back at the company buckets, the $2tn in gains since March are distributed very differently:

All of the gains in the market — and more — are from companies touched by AI. The S&P 500 as a whole is up more than 4 per cent. The non-AI 484 is down 2 per cent. 

Here are the 10 companies that have been the biggest drag on the S&P in the more recent period, along with their share price returns: 

It is interesting that Intel, a chip company the market has decided will not be helped by AI, is the worst offender. More to the point, Intel and the other three worst performers — Disney, Accenture and Salesforce — were all crushed after giving disappointing earnings targets when they reported first-quarter earnings. In this market, companies without an AI halo can’t afford to miss expectations. 

And if the idea gets around that AI is not going to be a great business for everyone involved, this rally will end badly.

One good read

On central bank’s losses.

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.