AI Investment Drives About One-Third of U.S. GDP Growth

ING Economic and Financial Analysis estimates technology investment linked to the AI boom accounted for 36% of U.S. year-over-year GDP growth in the second quarter of 2026 after adjusting for net technology imports, providing a measure of the investment cycle’s contribution to economic output.

Key Takeaways

• ING estimates technology investment contributed 36% of U.S. year-over-year GDP growth in Q2 2026 after adjusting for net technology imports.
• Real spending on computing, peripheral equipment and software increased 62% from Q1 2022.
• Technology-related imports rose to about $60 billion per month, reducing the domestic contribution of technology investment.
• ING estimates technology investment accounted for about one-third of the current U.S. GDP growth rate.
• ING cautions that its estimate may include technology spending that is not directly tied to AI.

ING Estimates Technology Investment at 36% of Q2 GDP Growth

AI investment and related technology spending accounted for an estimated 36% of U.S. year-over-year GDP growth in the second quarter of 2026 after ING Economic and Financial Analysis adjusted its calculation for net imports of computers, peripheral equipment and semiconductors.

The estimate uses a narrower measure of technology investment that includes computing and peripheral equipment and software. ING calculated the contribution after accounting for technology products purchased from abroad, which subtract from domestic GDP.

The 36% figure is lower than broader estimates that include total information-processing investment, data-center construction and software. Under that broader measure, technology investment represented 50.2% of year-over-year GDP growth in the second quarter, according to ING’s calculations.

A narrower calculation that excludes the import adjustment produced a 44% contribution to second-quarter GDP growth. ING then subtracted net imports and adjusted the import values for price changes, producing the 36% estimate.

ING said the adjusted calculation represents its fairest measure of the contribution of technology investment to U.S. economic growth. The analysis also estimates that technology investment accounted for about one-third of the current year-over-year U.S. GDP growth rate.

The calculation does not establish that all of the investment was directly caused by artificial intelligence. ING said it cannot fully separate AI-related spending from other technology investment because available economic data do not provide enough detail for every category.

For context on the broader U.S. growth picture, recent U.S. productivity growth analysis has also examined the relationship between capital use, productivity and economic output.

Real Spending on Computing and Software Rises 62%

Real spending on computing, peripheral equipment and software increased 62% cumulatively from the first quarter of 2022, according to ING’s analysis.

Data-center construction and electric-power generation were among the areas where ING identified construction-spending growth.

The analysis also found that other forms of nonresidential business investment weakened during the same period. Nonresidential business investment declined on a year-over-year basis for six consecutive quarters between the fourth quarter of 2024 and the first quarter of 2026.

ING’s calculation therefore attributes a substantial portion of recent U.S. economic growth to technology-related capital spending rather than treating all business investment as contributing equally.

Data-center construction is part of the investment cycle, but ING assigns a relatively smaller direct GDP contribution to the physical structures because the major cost of a data center is the computer equipment installed inside it.

Computing and peripheral equipment also include products that are not specifically used for artificial intelligence. Office computers and other conventional equipment are included in the available economic data, making it impossible to isolate every dollar of AI-related spending.

Software spending presents a similar measurement issue. ING includes software because of the increase in spending associated with the technology investment cycle but acknowledges that not all software purchases are AI-related.

The analysis therefore treats the increase in technology spending as an indicator of the AI investment cycle while recognizing the limits of the available categories.

Technology Imports Reduce the Domestic GDP Contribution

Technology imports significantly affect the calculation because imported products do not add to U.S. domestic production in the same way as goods produced domestically.

ING said imports of technology-related products increased to about $60 billion per month, roughly triple their value two years earlier. U.S. exports of those products also increased, but at a slower pace, rising from about $9 billion per month to $17 billion.

As a result, the U.S. trade deficit in computers, peripherals and semiconductors increased from about $20 billion per month to more than $40 billion.

ING adjusted the import figures for changes in prices before calculating their effect on real economic growth. The analysis used a producer-price measure covering electronic components and accessories to account for changes in semiconductor and technology-product prices.

The adjustment matters because a large increase in spending on imported technology does not translate directly into an equivalent increase in U.S. domestic output.

Technology-related services also form part of the analysis. ING used the computer-services category, which includes foreign AI subscriptions and token purchases as well as other digital services.

Computer-services exports were running at about $22 billion on an annualized basis in the second quarter, compared with about $19 billion of imports. Export volumes increased 15% year over year, while import volumes increased 7.5%.

ING acknowledged that attributing all of the observed growth in computer services to AI would be too broad, but said the category represents a relatively small portion of overall economic activity.

AI Investment Shows Limited Direct Consumer Spending Impact

Direct consumer spending on AI services remains a small part of overall household spending, according to the ING analysis.

Data cited by ING from Bank of America and PNC Bank indicated that about 2% to 3% of U.S. households were spending money on AI tools, with typical monthly spending of $20 to $30.

That level of direct spending remains much smaller than household spending on established services such as internet, television and mobile-phone subscriptions.

ING also examined an indirect consumer effect through household financial wealth. The analysis linked the rise in technology-related financial assets to increased household wealth and estimated that this wealth effect has had a larger current impact on consumer spending than direct purchases of AI subscriptions.

Household holdings of financial assets increased from $109 trillion to $142 trillion over the three-and-a-half-year period following the November 2022 release of ChatGPT, according to figures cited in the analysis.

The distribution of those assets is uneven. Federal Reserve data cited by ING show that the highest-income 20% of U.S. households hold 72% of the nation’s wealth, while the bottom 60% hold 15%.

For additional context on household spending differences, recent K-shaped economy analysis examined differences in spending patterns and wealth accumulation across income groups.

ING estimated that the increase in household equity wealth could have generated roughly $500 billion in cumulative additional consumer spending since the fourth quarter of 2022. The estimate was based on an assumed propensity to consume from wealth and was used to calculate a potential contribution to consumer spending and GDP.

The analysis estimated that the wealth effect could have contributed about 0.44 percentage point to GDP in the second quarter of 2026. Direct AI subscriptions were not the primary source of that estimated consumer contribution.

AI Investment and the Labor Market

The Federal Reserve’s Beige Book reported that most districts had not seen a significant overall staffing impact from AI, although some businesses said AI-related productivity improvements allowed them to delay or reduce hiring.

ING also cited LinkedIn data showing that entry-level hiring for graduates had fallen 17% since 2019. The analysis cited a July unemployment rate of 9.7% for recent graduates ages 20 to 24, compared with 2.7% for graduates overall.

Capital spending can increase measured economic output without producing an equivalent increase in hiring.

ING Estimates AI-Related Investment Accounts for One-Third of Growth

ING estimates that technology investment associated with the AI investment cycle accounts for about one-third of the current year-over-year U.S. GDP growth rate.

The estimate is based on the narrower technology-investment measure that includes computing, peripheral equipment and software and subtracts net imports of computers, peripherals and semiconductors.

ING’s four-quarter average for that adjusted measure was 37%, compared with the 36% contribution calculated for the second quarter of 2026.

The analysis also includes an important qualification: the available economic data cannot fully distinguish AI investment from spending on older or conventional technology.

Computers used in offices, non-AI software and other technology products remain part of the measured investment categories. ING therefore said its calculations may overstate the portion of technology spending directly attributable to artificial intelligence.

The analysis nevertheless estimates that the increase in technology capital spending since the release of ChatGPT is large enough that only a marginal downward revision would be needed to its broader estimate of AI-related activity.

Frequently Asked Questions

How much did AI investment contribute to U.S. GDP growth in Q2 2026?

ING estimates that its narrower measure of technology investment contributed 36% of U.S. year-over-year GDP growth in the second quarter of 2026 after adjusting for net technology imports.

What technology investments did ING include in its calculation?

ING’s narrower calculation includes computing and peripheral equipment and software. The analysis also considers data-center construction but assigns a smaller direct contribution to the physical structures.

How did technology imports affect AI investment’s GDP contribution?

Technology imports reduce the domestic GDP contribution because imported goods are not produced in the United States. ING adjusted for net imports of computers, peripherals and semiconductors before arriving at its 36% estimate.

How much has U.S. spending on computing and software increased?

Real spending on computing, peripheral equipment and software increased 62% cumulatively from the first quarter of 2022, according to ING’s analysis.

Does direct consumer spending on AI significantly affect GDP growth?

ING estimates that direct household spending on AI subscriptions remains limited, with about 2% to 3% of households spending on the tools. Its analysis assigns a larger current consumer effect to wealth gains associated with technology-related financial assets.