Econbrowser author Menzie Chinn says an AI investment downturn may have a limited direct effect on U.S. GDP because imports offset much of its measured contribution, while falling stock wealth could slow consumer spending.
In a Nov. 6 post, Chinn compares computer and software investment with imports of computing equipment and semiconductors. The post’s figures credit Bureau of Economic Analysis (BEA) data and Chinn’s calculations; its household-wealth chart also draws on Federal Reserve Flow of Funds data.
His accounting analysis suggests the investment and import channels could nearly balance if their multiplier effects were similar. A multiplier describes how an initial spending change carries into broader economic activity. Adding available estimates for data-center construction through 2025 did not change the conclusion he draws from the comparison.
Chinn connects the AI capital-spending boom to expectations of future returns. He says those expectations could weaken if planned investment exceeds financing capacity or if views of the policy environment shift, leaving asset prices exposed to a reassessment.
He also considers a separate household-spending channel. In a scenario where stock-market values fail to rebound in the fourth quarter, lower household net worth would point to slower consumption growth if other economic conditions stayed the same. Chinn frames the result as weaker growth relative to what would otherwise have occurred. This channel depends on the stock-market path he assumes, apart from the investment-import comparison.
The post cites a 2026 estimate by Bigot and Espic that assigns roughly half of the 2% rise in U.S. consumption during 2025 to equity-market wealth. Chinn says the estimate uses a propensity to spend of 2.7 cents per additional dollar of equity wealth, or 0.027.
Consumption had increased 2.5% over the year ending in the third quarter, Chinn wrote. He said the wealth impact alone would still leave GDP growth above zero if other conditions held. He added that an abrupt market repricing related to AI could increase uncertainty and weigh on aggregate demand.
