Earnings season is upon us, and Wall Street expects companies in the S&P 500 Index that aren’t tied to artificial intelligence to rake in profits. That would be good news if it pointed to a broadening out of earnings growth, but instead the gauge is in a Pyrrhic boom created by President Donald Trump’s war with Iran. It’s a bump that we’ll pay for later.

Energy industry net income is expected to soar about 135% in the third quarter from a year earlier, while the aerospace and defense industry may see profits jump by about 50%. Add to that the projected 39% jump in the materials sector — including war-impacted petrochemicals, fertilizers and metals — and you’ve explained away most of the non-AI strength.

The war may continue to be a source of some idiosyncratic profits for as long as it continues, but at what cost? The “everything else” category of adjusted S&P 500 earnings, those not linked to the war or AI, is expected to have grown just about 4.5% last quarter, barely outstripping the 3.7% pace of inflation.

War earnings are inherently parasitic. The economic tradeoff is higher fuel prices and borrowing costs, which hamper consumer sentiment and crowd out other types of consumption. The resultant voter frustration may cost Republicans control of the House and Senate in next month’s midterm elections.

We got a taste of this in the short-lived earnings recession of 2022 and early 2023, before AI became the dominant market narrative, displacing Russia’s invasion of Ukraine and the Federal Reserve’s interest-rate hikes. AI is still a major source of market ballast, but the consumer economy has grown shakier under the weight of repeated economic surprises, including Trump’s 2025 tariffs. Inflation has outstripped growth in hourly earnings for the past six months, leaving lower-income consumers to rack up credit card debt or cut into savings to finance their spending.

History shows that the deleterious effects of energy shocks are usually somewhat slow to manifest themselves in aggregate earnings and output. Consumers and businesses can typically brush off higher fuel prices for a while, but they run out of cushion when conflicts drag on for a long period. A Federal Reserve Bank of Dallas paper found that industrial production typically declines with a delay of about 10 months, with household appliances, furniture and automobiles among the sectors hit the hardest. Of these, only purchases of motor vehicles turn noticeably sluggish in the early days, but a lack of demand tends to hit the entire non-petroleum economy eventually.

The retreat may have come early in some sectors this time around, only about a half year after Trump joined Israeli airstrikes on Iran. The S&P 500 consumer discretionary sector is expected to see just 1.8% growth this quarter, and consumer staples probably grew about 1.3%. Real estate was essentially flat, and health care earnings remain in contraction.

Going forward, rate-sensitive sectors including housing and autos will probably get worse if inflation doesn’t abate soon. The National Association of Home Builders/Wells Fargo gauge of housing market conditions fell to the lowest since June 2012 in a report published last week. If these aren’t early signs of a war-related drag, then they show that demand was weakening under the surface even before Iran.

Won’t AI help the economy power through all of this? Analysts expect net income for the S&P 500 information technology sector to grow 67% this quarter, bringing index-wide earnings growth to 26%! Not only does the AI boom reach deep into other sectors through data centers — including utilities and industrials — but it has contributed to an increase of about $52 trillion in household net worth since the end of 2022, when ChatGPT was released. That, in turn, can continue to cushion consumption, if it holds.

No one can reliably time a massive boom like AI, but billionaire Ken Griffin last week joined Bridgewater founder Ray Dalio in suggesting that we were in a bubble. Fixed-income investors are suddenly demanding much higher yields on debt from riskier AI-related issuers including CoreWeave Inc. Whatever you think of the doomer thesis, there’s no question that narrow breadth raises the odds of turbulence. After all, so much of the economy and market now rest on the ability of a few hyperscalers to prove AI is profitable and worth further investments in more chips and infrastructure.

That’s why the recent jump in energy and materials earnings is a false victory. The price is increased pressure on households that were already running out of runway and a consumer earnings base that will probably soften further from here, leaving a razor-thin cushion to buoy the market if Griffin and Dalio are right.