Artificial intelligence has become a powerful engine of U.S. economic growth this year, even as its impact on employment remains modest for now, according to a new report from the Bank of America Institute, the bank’s think tank.
AI fuels economic expansion
AI-related investments, particularly in software, computing and data infrastructure, have been a major driver of U.S. GDP growth in 2025. According to the October 22 report, “AI-related capital expenditures, particularly in software and computing, have significantly boosted GDP growth in 2025, contributing up to 1.3 percentage points in Q2.”
BofA Global Research estimated that new AI investment alone could have accounted for up to 1.2 percentage points of growth in the first quarter and 1.3 percentage points in the second, helping the economy rebound after a soft start to the year.
“After a slight decline in Q1, GDP growth rebounded in Q2 with 3.8% for an annualized rate of 1.6% in 1H,” the report said. “One major reason for the resiliency in growth to date is the investment being poured into technology- and AI-related categories.”
Small businesses join the AI wave
Adoption of AI and digital services is no longer confined to large corporations. Drawing on internal transaction data, the Institute found that small business payments to technology services rose 6.9% year over year in September, led by strong gains in manufacturing and construction sectors.
“Small firms have continued to prioritize tech spending amongst the overall uncertain economic backdrop,” the report said. This suggests that the diffusion of AI-driven productivity tools is broadening across industries.
Data centers drive growth, power demand
Much of the current investment surge is concentrated in software, computing and peripheral equipment, but data centers stand out as one of the few bright spots in the otherwise sluggish structures category.
“Data centers are one of the lone bright spots of large and expensive structures investment,” the Institute wrote, even though their share of total investment remains “tiny in magnitude.”
The boom, however, has come with side effects.
“Consumers have seen energy prices outpace overall prices this year,” the report said, citing energy services inflation running well ahead of the headline CPI, particularly in regions with high data-center demand.
The authors added that rising utility investment could turn into another tailwind for capital spending next year, especially if favorable tax incentives under the “One Big Beautiful Bill Act” spur broader infrastructure expansion.
AI reshaping productivity more than jobs
Despite fears of mass displacement, Bank of America Institute data show little evidence that AI has caused widespread job losses.
“When comparing AI usage levels to year-to-date employment changes in major industries, there is a slightly negative correlation between higher AI usage and employment growth,” the report said, adding that “the relationship is insignificant.”
In fact, white-collar sectors including finance, professional services and information are experiencing a positive link between AI usage and employment growth.
“These findings suggest that AI is playing out mostly as a productivity story for white-collar workers,” according to the Institute.
Future labor outlook: winners and losers
While the long-term impact of AI on jobs remains uncertain, the Bureau of Labor Statistics projects a 17.9% increase in employment for software developers and a 17.1% gain for personal financial advisors between 2023 and 2033. At the same time, roles such as credit analysts, claims adjusters, and paralegals face greater automation risk.
“AI is expected to primarily affect occupations whose core tasks can be most easily replicated by generative AI in its current form,” according to the report.
The Bank of America Institute said that while AI investment has “room to run,” the sustainability of growth depends on whether the investment wave expands beyond technology and data infrastructure. If capital spending broadens to utilities and other sectors, AI could help offset some of the labor supply constraints facing the U.S. economy.
But if that doesn’t materialize, “consumers will continue to bear part of the cost of powering data centers,” the report said.