AI Will Change Everything. AI Stocks Won’t.

‍AI will be a transformative, revolutionary technology. Nonetheless, Americans are pessimistic about it, and they’re not wrong. There’s every chance AI triggers human harm or does serious economic damage somewhere—whether through labor displacement, misinformation at scale, security failures, or just the gradual human consequences of a world where bureaucracy, surveillance, and automation get dramatically cheaper. To borrow a line from the Economist, catastrophe is not if, it’s when and let’s just hope we get Three Mile Island, not Chernobyl.

Since ChatGPT was released in November 2022, 75% of the entire stock market gains (a ~21% total return) have been solely from AI stocks. It’s very likely that most of your portfolio returns since then are AI stock-driven. (Note: AI stocks refer to companies that are investing in AI, building AI, and transforming their businesses around AI. It’s not only the companies that train LLMs like OpenAI and Anthropic.)

Despite the scale of change AI promises to deliver, the history of technological revolutions (canals, railroads, automobiles, and the internet) suggests AI stocks will succeed economically but underperform long-term investor expectations. This is counterintuitive. Why wouldn’t it be smart to invest in companies at the forefront of a technological revolution? 

Because inevitably, a stock market asset bubble forms. This is happening now at digital speed because of financial democratization, a meme-driven investor market, and rapidly shared information across huge groups. Many people with the same “genius” idea pour money into hot AI stocks, raising their price far above what is justified by fundamental measures like profitability.   ‍

It works in the beginning. The stock keeps going up, and things look good because there is always a greater fool willing to pay a higher price – but eventually even the fools come to their senses and stocks start to fall. Two other downsides are that economic competition erodes profits faster than humans predict and behaviorally, consumer investors systematically overpay for technology stocks due to excitement. Paying too much for stock up front reduces long term returns.

This all leads to painful stock market washouts, lower than expected future returns, and frustrated investors. At Lifetime Financial, our portfolios modestly and systematically deemphasize pricey technology stocks, and we already see significant evidence of a bubble in AI stocks. 

How do we avoid the bubble? Again, the answer is counterintuitive. We can’t avoid it. The best strategy is to ride the bubble out like a roller coaster on which we are stuck.

Why can’t we just sell now, and get back later when it’s safe?  Because we really don’t know the final shape of the bubble: how tall the climb will be, how long the dip will last, and how quickly the market eventually recovers. We might only be a quarter of the way to the top of this bubble, and those future returns are needed to offset the upcoming losses of the steep drop to follow. In other words, avoiding a bubble requires being right twice. The first time you need to sell high at the market top, and then a second time to get back in at the bottom. These are extraordinarily difficult predictions to make and identify in real time. For long-term investing, riding out a bear market is statistically the strongest strategy, and that’s how we do it at Lifetime.

Still, given the massive impact on the economy, AI stocks must be a good investment at some point, right?  Again, the (counterintuitive) answer is no. Only a few of the early leaders in a technological revolution become the dominant profitable players later. This is immortalized in the 4% Rule by financial expert Hendrik Bessembinder (Arizona State University), who analyzed the entire database of U.S. common stocks from 1926 to 2020 and found that just (4.2%) of companies accounted for all the net wealth creation in the U.S. stock market over that 94-year period. Only a few companies end up dominating long-term stock market returns, and it’s impossible to pick the winners ahead of time. Most companies will fail or be absorbed by a competitor. The only effective strategy is to own them all via an index fund – also a core Lifetime Financial investment strategy.

If it weren’t for AI excitement, it would look stormy ahead for the market. On the plus side, corporate earnings and employment remain strong, both of which are critical core economic indicators. However, we face heavy inflationary stimuli (the Iran war, tariffs and immigration policy), which prevent the Federal Reserve from lowering rates, and in turn, keep housing costs high. Additionally, there are uncertainties regarding the private credit selloff, a swelling federal budget deficit, the dollar is softening, and decades of supply chain integration are being chaotically unwound. 

Humans tend to overestimate what technology will change in five years and underestimate how things will change in 10 or 15 years. The modern version of this theory is the Gartner Hype Cycle, but the phenomenon was coined in the 1970s by Roy Amara as Amara’s Law.  Our over/underestimation stems from a) linear thinking when technology curves are exponential, b) the marketing and media hype cycle unreasonably raising our expectations, and c) society and our infrastructure take longer to change than we think. 

In the long run, we are all going to have to stay open to change and be kind to those who are forced to harshly adapt to the new economy. The next 15 years will bring just as many or more challenging events as the last 15. Buckle up.