AI investment boom faces dot-com comparison as Mag-7 index retreats
An Econbrowser analysis says a 12.2% fall in Bloomberg’s Mag-7 index raises the risk that AI-linked capital spending could slow after a lag.
By Ingrid Halvorsen · Staff Writer
· 3 min read
Bloomberg’s Magnificent Seven index is 12.2% below its May 29 peak, according to an Econbrowser analysis citing Bloomberg data, putting renewed attention on whether equity weakness could later feed into AI-related investment. The comparison drawn by the analysis is the dot-com cycle: after technology shares fell from their early-2000 highs, nonresidential fixed investment reached a peak later that year.
The Econbrowser analysis said the U.S. stock market topped out at the end of the first quarter of 2000, while nonresidential fixed investment peaked in the fourth quarter of that year. It used that historical lag to consider whether a similar pattern could follow a potential high in AI-linked equity prices.
The current data do not yet show a broad investment rollover. Figures cited from the Bureau of Economic Analysis, the Atlanta Fed’s GDPNow estimate dated July 2, 2026, and the author’s calculations show equipment and software investment continuing to rise in real, seasonally adjusted annual-rate terms. The analysis described the trajectory in equipment and software spending as still strongly upward.
Why the equity-investment link matters
In equity-led booms, stock prices can influence capital spending by changing firms’ financing conditions and management incentives. Higher market valuations can make it easier for companies to issue shares, borrow on favorable terms or justify large capital budgets. A sustained decline can have the opposite effect, particularly when investment depends on external financing or high expected returns.
The Econbrowser analysis said the analogy with 2000 has limits. It noted that many of the Magnificent Seven companies have enough retained earnings to fund investment internally, reducing their immediate dependence on equity issuance. It also said large companies are only now turning to debt markets, another difference from a funding cycle driven primarily by stock-market valuations.
That distinction matters for AI infrastructure. Spending on data centers, computing equipment and software can continue for a time even after share prices weaken if companies fund projects from operating cash flow or existing balance-sheet capacity. The analysis did not identify a current spending peak, and it noted that the Bloomberg index could recover if policy rates fall or policy uncertainty declines.
Forecasters still see growth
Using a hypothetical three-quarter lag between an equity-market peak and an investment peak, the Econbrowser analysis compared the present cycle with the 2000 episode. The comparison normalized investment around a possible 2027 first-quarter peak and the actual 2000 fourth-quarter peak, using BEA data, the Atlanta Fed’s GDPNow estimate and professional forecaster projections.
The analysis said professional forecasters assume investment will keep growing. It also said that if the recent high in AI-related equity prices proves to be a peak, that would point to a possible downturn in investment ahead.
The conclusion remains conditional. The cited market decline is confirmed by Bloomberg’s index level relative to May 29, while the investment downturn scenario depends on whether the equity move represents a durable peak and whether the historical lag from the dot-com period is a useful guide for the AI spending cycle.
This story draws on original reporting from Econbrowser.
