ICIMS Insights September Workforce Report: U.S. and EMEA hiring slow as AI skills race heats up

September 18, 2026

 

Trent Cotton

6 min read

Hiring is cooling on both sides of the Atlantic but not at the same speed.  

• In the U.S., the gap between job openings and hires keeps widening even as growth slows.  

• In EMEA, hires just posted their sharpest single-month drop of the year.  

Underneath both, the AI skills race is intensifying, and candidates are building those skills faster than employers are training for them. Get the full breakdown below and download the complete ICIMS Insights September Workforce Report at www.icims.com/insights. 

Two speeds, one strain 

Demand was accelerating on both sides of the Atlantic last month, and the labor market answered with strain instead of speed. That demand has now turned, but the strain has not gone anywhere. In the U.S., job openings are no longer surging ahead of hiring capacity. In EMEA, the whole system just cooled at once.  

This month, ICIMS also turns its spotlight on the AI skills race and what the market’s response reveals about what skills are in most demand.  

The U.S. gap is not closing. It is just moving slower 

Job openings in the U.S. finished August 13% above the August 2025 baseline. Month-over-month growth fell from 7% in June to 4% in July to just 1% in August. Three consecutive months of slowdown could be normal pacing after a hiring surge. It could also be an early signal that demand is turning. 

Hires did not keep pace. They closed the month up only 2% against the August 2025 baseline and posted their first back-to-back monthly decline of the year, down 1% in August after a 3% drop in July. Openings are now growing roughly 11 points faster than hires on a year-over-year basis, a wider spread than July’s. 

Applications tell a more hopeful story, up for a third straight month after May’s steep decline. But a recovering top of funnel does not fix a stalled bottom of funnel. The constraint has shifted from finding candidates to converting them. 

EMEA cooled faster, and the calendar only explains part of it 

Job openings in EMEA finished August 21% above the August 2025 baseline, but growth has now been negative for two straight months. Hires tell the sharper story. After remaining above baseline from March through July, hires fell 20 points in a single month, the largest one-month move either region has seen all year. 

August is peak vacation season across much of Europe and hiring slows region-wide, which could contribute to the decline in the three metrics. However, it does not fully explain a drop this size. 

Applications reinforce the concern. They fell for the third straight month in EMEA, down 7% in August alone, moving in the opposite direction from the U.S., where applications have been climbing. Employers are pulling back and candidates are disengaging at the same time, and neither side is offsetting the other. 

The AI skills race has a talent pool problem, not a demand problem 

Skills-based hiring is having a moment, especially as organizations rethink how they evaluate candidates for an AI-upgraded workforce. ​This report draws on two data partners to understand what that shift looks like: 

• Lightcast’s job postings and skills data shows where and how fast AI requirements are spreading across occupations and industries.​ 

• ICIMS commissioned a survey in August 2026 of 1,000 U.S. adults to understand how they are responding to the AI skills race. We found candidates are building AI skills largely on their own.​ 

AI-related postings are still a small share of overall hiring: 4% in the U.S., 2.7% in the UK, and 1.2% in France. But six occupations, including generative AI engineer, machine learning engineer and data scientist, rank in the top 10 for AI skill demand in every region studied. Every employer hiring for these roles is drawing from the same narrow pool. 

Adoption is uneven by industry, too. Finance leads three of the four regions, which tracks with where the work sits: risk decisions and fraud detection are exactly the kind of pattern-driven work AI tools handle well today. Healthcare is the outlier. It has the largest share of newly prominent AI skills of any sector measured and the lowest AI hiring rate of the three. Requirements are arriving faster than hiring capability can absorb them. 

Candidates are not waiting for employers to catch up. According to the ICIMS survey, self-teaching for AI climbed from 22% to 30% in a year, while reported employer-provided training for AI barely moved. But confidence is outrunning depth. Sixty-seven percent of job seekers say they feel ready to meet an AI requirement, yet only 25% call themselves genuinely skilled. Most of what candidates have is familiarity with a tool like ChatGPT, not the specialized capability roles actually need. A generalist confidence score cannot tell you whether someone can do the specific work in front of them. 

That gap is the real story this month. The market is not short on people learning AI skills. It is short on employers who can tell the difference between exposure and expertise. 

Get the full report 

Download the ICIMS Insights September Workforce Report at www.icims.com/insights for the full U.S. and EMEA data, the complete AI skills race analysis and what it means for your hiring strategy heading into Fall.