US AI adoption hits 20.6% as job market impact stays narrow

Artificial-intelligence adoption among US companies reached 20.6 percent in June, up 1.1 percentage points from May, yet the technology's impact on the labor market remains confined to specific job categories, Goldman Sachs analysts said.

"AI's labor market impact remains visible but narrow," Sarah Dong, a macroeconomic research analyst at Goldman Sachs, and Joseph Briggs, an economist at the bank, wrote in a note Tuesday. Employment drags have emerged in marketing, graphic design, customer service and some tech occupations where AI use cases are established, they said, but those constraints have been offset by job growth in construction as tech companies race to build data centers.

The Census Bureau's Business Trends and Outlook Survey shows AI penetration rising from 20.6 percent in June to an expected 24 percent by year-end. Information, professional services and education companies lead adoption, with some finance firms reporting usage near 80 percent and publishing companies above 50 percent. Among businesses with more than 150 employees, the adoption rate stands at 41 percent.

The data challenges the narrative that generative AI is already driving broad-based white-collar layoffs. Challenger, Gray & Christmas reported Wednesday that US employers announced 45,849 job cuts in June, down more than 50 percent from 97,006 in May. While technology remains the sector with the most concentrated cuts, the overall pace of layoffs has cooled. Dong and Briggs found no statistically significant correlation between AI adoption and unemployment figures, even as tech's share of total employment has declined compared with the pre-2022 hiring boom.

Productivity gains emerge, but remain uneven

In the few areas where generative AI has been deployed, productivity improvements are showing up in the data. Academics have documented a 23 percent gain, while client anecdotes cited by Goldman Sachs suggest a rise of about 34 percent. The divergence reflects the early stage of measurement — companies are still determining how to quantify efficiency gains from tools deployed across marketing, customer service and software development workflows.

A separate study from Ramp, conducted with labor market analytics firm Revelio Labs, analyzed AI spending and employment records for 21,559 US companies between 2021 and early 2026. The researchers found that firms with the highest AI spending intensity increased overall employment by roughly 10 percent after adopting the technology, with entry-level hiring rising about 12 percent. Low-intensity adopters saw no statistically significant change. The gains emerged gradually over six to 12 months, suggesting companies require time to integrate AI into workflows before realizing productivity benefits.

The Ramp researchers cautioned that AI adopters were already larger, faster-growing and more likely to be venture-backed before deploying the technology, making simple comparisons with non-adopters misleading. The study defined adoption as three consecutive months of at least $100 in AI vendor spending, with intensity measured by spend per employee during the first three months after deployment.

What this means for investors

For investors tracking the AI trade, the data points in two directions. The 20.6 percent adoption floor confirms that enterprise AI integration is accelerating, which supports demand for infrastructure providers such as Nvidia Corp. and Microsoft Corp., as well as AI platform companies. At the same time, the narrow labor market impact suggests the disruptive phase of AI — the period when job displacement fears could trigger regulatory pushback — remains years away, reducing near-term policy risk for the sector.

The productivity gains, if they scale beyond the early adopter cohort, could eventually pressure margins at companies that fail to integrate AI tools. But for now, the data from both Goldman Sachs and Ramp suggests that AI investment is complementing workforce growth rather than replacing workers — a dynamic that, if sustained, would support continued capital spending on AI infrastructure without the political backlash that mass displacement would invite.

This article is for informational purposes only and does not constitute investment advice.