By Teresa Rivas

The case for the artificial-intelligence trade is the technology is so revolutionary that eventually it will be well worth the trillions that have been invested to develop it.

But that argument means many investors don't bother to distinguish between AI-related companies with solid fundamentals and those on shakier ground.

In reality, the shares of most AI winners have traded on hopes for the far-flung future during this bull market, rather than the here, now, or near future, notes Trivariate Research President Adam Parker.

That's based on his work dating back to early 2023. As Parker details in a note Friday, his firm created three baskets of AI plays back then -- 10 semiconductor stocks, 19 software, and 20 "other" companies, which included a mix of names like Amazon.com, Dell Technologies, and Snowflake. Those baskets were then further divided into companies with high, low, and negative free cash flow.

After tracking these stocks, Parker found that the market doesn't care about free cash flow. The stocks in the "other" group have a median FCF yield -- free cash flow as a percentage of firm's market value -- of 4.3%, well above the rest of the market. Meanwhile, semiconductor companies now have significantly less FCF than those in other categories, yielding a paltry 1.7%. And yet semiconductor stocks have outperformed, as the clear winning group since the White House shook markets with its widespread tariffs announcement back in April.

In other words, AI believers aren't paying attention to free cash flow yield, a valuable metric to gauge a company's current performance.

"Semiconductors with low free cash flow have strongly outperformed those with high free cash flow, and Software with negative free cash flow (albeit a small sample size) have beaten those with positive free cash flow," Parker writes. "Among other AI companies, high- and low-free-cash-flow stocks have performed similarly."

Even near-term revenue growth forecasts haven't had much impact on stocks so far -- even though those can be predictive for a company's profit margins and expected growth. Parker looked at 12-month forward forecasted revenue growth versus FCF yield in each AI bucket, and, again, didn't see any aggregate relationship.

"We can only conclude that movements in AI stocks are driven by perceptions far out in the future, not by current capital spending and its impact on FCF, or the next year's revenue growth," he writes.

One might expect that this would be the case during the good times, and that investors would want the comfort of more well-capitalized companies with stronger near-term outlooks when things go sideways. But that hasn't proved to be the case, either: Shares of solid companies with strong revenue growth outlooks or higher FCF yields have struggled this year, in many cases just as much or more than those with weaker fundamentals.

Across the board, all stocks in Parker's analysis -- with the sole exception of MongoDB -- are down from their 2025 highs, ranging from 3.4% for Zoom Communications to nearly 60% for C3.ai.

It is true that the four companies with negative FCF yields -- Oracle, Alpha & Omega Semiconductor, SoundHound AI, and C3.ai -- are all off between 36% and 59% from their 2025 highs. Yet the three companies with double-digit FCF yields -- Concentrix, Sprinklr, and Five9 -- fared nearly as badly as a group.

In short, when it comes to the AI stocks that have powered this rally, the question isn't what have you done for me lately, but what will you do for me...eventually?

Write to Teresa Rivas at teresa.rivas@barrons.com

This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

December 06, 2025 02:00 ET (07:00 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.