The AI buildout is adding resilience to the economy at a time when consumption is softening and rates remain elevated, and shows some independence to variables like interest rates, labor markets and even trade shocks.
Throughout economic history, different variables have served as reliable bellwethers of growth. Housing played this role for decades due to its cyclical nature and interest rate sensitivity. With globalization, manufacturing PMIs and trade flows became key. In recent decades, consumers have taken center stage, with retail sales and employment data in focus.
Today, evidence suggests a new bellwether may be emerging: artificial intelligence. In the first half of 2025, AI-related capital expenditures contributed 1.1% to GDP growth, outpacing the U.S. consumer as an engine of expansion.
Investors have become accustomed to AI as a dominant market theme, but it is also becoming an economic one?
Where AI is showing up in the data
The clearest signal is in business investment. In Q2, tech-related categories contributed 4.3 percentage points to overall investment growth, offsetting declines elsewhere. Hardware has led, with investment in computers and related equipment up 41% on the year, reflecting a surge of orders for servers and GPU systems. Data center construction hit a record $40 billion annual rate in June, up 30% from last year—a bright spot in an otherwise challenged construction environment.1
This investment surge has been driven by the hyperscalers (Meta, Alphabet, Microsoft, Amazon and Oracle), projected to allocate $342 billion to capex in 20251, a 62% increase from last year’s 67%. Private companies like OpenAI and Anthropic are also making similar investments to support further frontier model development.
From a GDP perspective, the impact is still modest, but this should evolve. Official data primarily reflect the first phase of AI investment, emphasizing chips, servers and networking equipment. This next phase is targeting supporting infrastructure such as power plants and grid upgrades, which can take years to plan, permit and build. Early signs of this phase are emerging, but the full impact is likely ahead.
The investment boom is real, but the growth impact will likely be less dramatic
Not every AI dollar will translate directly to U.S. GDP. Much investment goes toward imported technology goods, which subtracts from GDP, and efforts to reshore manufacturing capacity will involve a long transition process. Data centers also employ few workers once built2, especially compared to a factory or office campus, limiting their multiplier effect through wage-driven consumption. Grid capacity and permitting bottlenecks may also limit expansion despite willing and available capital.
Another risk is that while hyperscaler capex has been supported by strong operating cash flows, historical precedent suggests tech investment cycles can be volatile. If projected demand falls short of forecasts, spending could turn on a dime.
Innovation as a cushion
The AI buildout is adding resilience to the economy at a time when consumption is softening and rates remain elevated, and shows some independence to variables like interest rates, labor markets and even trade shocks. But while this wave lends support to parts of the economy benefitting from AI investment, it could also add vulnerability if underlying fundamentals remain weak. This makes diversification all the more crucial, so that portfolios are positioned for both the promise of innovation and the persistence of macro risk.