I’m undertaking a 1000-day reinvention project, blogging here daily to track my progress. In Monday Money, I write about money management.
The stock market is back to its seemingly inevitable march upwards, on the back of the AI trade, but meanwhile the companies that should stand to benefit from AI — companies in sectors outside of tech — aren’t seeing improvements to profit margins.
When is the AI productivity boom going to arrive?
The payoff could be delayed
In The Buyers of AI Are Still Waiting for the Payoff, Apollo Chief Economist Torsten Slok writes that only AI sellers are benefiting from AI so far:
Health care margins have halved since 2015; consumer staples are stuck near 6% and consumer discretionary near 8%; energy and materials have given back most of their 2022-23 gains; real estate is going sideways; and the only genuine improvements look like an ordinary cyclical recovery rather than a technology-driven step change.
The bottom line is that the AI capex boom is so far only showing up in the sellers’ margins, not the buyers’.
This is important because the longer it takes the S&P 493 to generate ROI, the bigger the downside risks to an economy and a market this concentrated in the AI trade.
Slok writes that consensus estimates call for surging hyperscaler free cash flow by 2028, but this payoff might take longer or never happen. One factor is the pressure Chinese AI models are putting on token prices of U.S. AI models. Because of soaring AI costs, U.S. companies have been shopping around for alternatives that are cheaper than OpenAI or Anthropic offerings, and are increasingly turning to Chinese models like DeepSeek and Z.ai.
What could this mean for the U.S. stock market? Right now, the Mag 7 makes up 32 to 34% of the S&P 500 index, with Nvidia leading the group at 7 to 7.5% of the total. If you add in the broader semiconductor industry, that gets you to about 44% of the total. That means choosing what seems like it should be a broad-market fund is actually almost half AI-dependent tech.
Slok suggests what might happen if the AI payoff comes slower than expected. Cash flows and earnings could disappoint as committed capex and depreciation for previous expenditures rise while no free cash flow boost arrives. We could see a Mag 7 sell-off taking the market down. Balance sheets will deteriorate and credit risk will rise, as hyperscalers depend more and more on debt funding for their capex.
What to do as an individual investor
What’s an investor to do? Many might think, “I can get out before a crash” but, in fact, most people tend to sell after equities have fallen (because otherwise how do you know that the top is in?)
Instead, now is the time to start tilting towards investments that aren’t so heavily dependent upon a quick payoff for AI capex.
How to do that? You have to be cautious, because some things you might think will get you out of the AI trade will keep you in it — like a market cap weighted emerging market fund such as $EEM, which is 15% Taiwan Semiconductor, and another almost 13% Samsung and SK Hynix (two AI hardware darlings).
One way to lean towards less pricey areas of the market is by choosing either value-tilted ETFs or by moving towards geographic regions with lower CAPE ratios.
Here’s an updated global market valuation table from Meb Faber’s Idea Farm.
Note the countries at the top by CAPE (the Shiller P/E which divides a stock index’s price by the average of its inflation-adjusted earnings over the past 10 years): Taiwan at 51.6, Korea at 43.9, and the U.S. at 40.3.
In the dotcom run-up, the S&P 500 index reached a record 44.2 in December 1999. The actual top of the market didn’t arrive until March of 2000.
Of course there’s no reason the stock market can’t get even more richly valued this time around, with more savvy investors, algorithmic trading, and megacap companies driving it.
Still, who wants to chance another almost 50% drawdown (what was seen after the dotcom bubble burst) or worse, almost 57% (during the Great Financial Crisis)?
Some people will tell you “the Fed will never let that happen again.” But the Fed isn’t entirely in control of what happens.
So it makes sense, right now, I think to lean towards sectors of the market that aren’t dependent on the AI trade. Instead of $SPY (market cap weighted S&P 500), $RSP (equal weighted S&P 500). Instead of $EEM (market cap weighted emerging market), $AVES (a value-based emerging market ETF). Instead of $QQQ, how about $FNDX?
I’m leaning towards less AI and more value right now. I might miss out on some of the run up, but I’ll have prepositioned for a bubble burst.