On October 2, 2026, CompTIA reported that active tech job postings reached 625,633 in September, the highest total in more than three years and up 22.4% from a year earlier. Active postings requiring AI-related skills reached 349,821, up 32,327 from August. By that measure, employer demand for tech talent is at its strongest level in more than three years.
The same release told a different story about jobs actually filled. CompTIA estimates that tech occupation employment across all industries fell by roughly 6,000 in September, and that employment at tech sector companies fell by about 10,350. Demand was rising on paper, while tech employment fell.
The Bureau of Labor Statistics shows the same gap from another angle. In its Job Openings and Labor Turnover Survey (JOLTS) released September 29, 2026, the information industry posted 123,000 job openings in August, up from 78,000 in July. Hires in the same industry fell to 44,000, down from 62,000 in July and 69,000 a year earlier.
That is close to three openings for every hire in the information sector in August, against roughly one and a quarter openings per hire in July. For a VP of Engineering, the gap is not an abstraction. It is the distance between an approved requisition and a person shipping work, and every month it stays open is a month of roadmap you are not delivering.
A disclosure before we go further. Golabs recruits and staffs AI engineers and dedicated AI teams, so we benefit when companies look outside their own hiring pipeline. CompTIA sells certifications, Experis is a ManpowerGroup staffing brand, and Staffing Industry Analysts serves the staffing industry. We flag each source's interest where it applies.
Why Are Tech Job Postings Rising While Hiring Falls?
Because a job posting measures intent, not a hire. In September 2026, CompTIA counted 625,633 active tech postings, a three-year high, while tech occupation employment fell by about 6,000. In August, the information industry had 123,000 openings but only 44,000 hires. Employers want the skills, but in our experience, cautious budgets, tougher role profiles, and a slow-moving candidate pool keep roles open longer.
Two caveats keep this honest. First, the BLS information industry is not the same as tech occupations. It includes publishing, broadcasting, telecommunications, and data processing, and it excludes tech workers employed in banks, hospitals, or manufacturers. It is the closest federal industry proxy for tech hiring, not a precise one.
Second, the August JOLTS figures are preliminary and a single month. Information openings were higher a year earlier, at 141,000, and the industry series moves sharply from month to month. The next release, covering September, is scheduled for November 3, 2026. Treat August as a signal worth watching, not a trend.
With those caveats, sources point to the same direction. CompTIA's postings data, employment estimates, and JOLTS hires data all describe employers advertising more roles than they are filling. The more useful question for a hiring leader is why, because each cause has a different fix.
What Is Actually Slowing AI Hiring Down?
No single source measures the reasons directly, so this section combines what the data shows with what we see in our own recruiting work. We label which is which. Four causes come up again and again, and most stalled requisitions we review involve at least two of them at once.
Employers are more cautious about committing headcount. Experis's Q4 2026 tech talent outlook, released September 22, puts the US tech Net Employment Outlook at 37%, down 10 points from both the previous quarter and a year earlier. Most employers still plan to grow, with 53% expecting to increase staffing, 30% to hold, and 16% to reduce.
Kye Mitchell, President of Experis US, described it this way: "U.S. tech hiring is moderating, but this is a market getting more deliberate, not pulling back." In practice, a more deliberate market means more approval steps, more interview rounds, and longer debates over each candidate. That last observation is ours, not Experis's.
The role profile has become harder to fill. Experis found that US tech employers value human skills most: Professionalism and Work Ethic (44%), Critical Thinking and Problem-Solving (39%), and Adaptability and Willingness to Learn (37%). AI Literacy is the most sought technical skill at 34%. In our experience, employers now want both sets in one person, which narrows the pool.
CompTIA's data points the same way. Of the 349,821 active postings requiring AI-related skills, only 13,636 were in its AI and machine learning job titles group. Most AI demand sits inside existing roles such as software engineering, data analysis, and DevOps, which now carry an extra requirement. We covered this shift in our 2027 headcount planning guide.
Experienced people are staying put. The JOLTS quits rate for the information industry was 1.1% in August, against 1.2% a year earlier, and the quits rate across all industries was 1.9%. When fewer people leave their jobs voluntarily, fewer experienced candidates are on the market, and, in our experience, the ones who are often field several offers.
Pay expectations have moved faster than salary bands. Senior engineers with AI experience command a premium, and requisitions approved on last year's bands often stall at the offer stage. That pattern comes from our own searches, not from the sources above. Our guide to what you should pay AI engineers in 2027 covers current ranges.
Are Companies Filling the Gap With Contractors?
Signs point to this, though no source we found attributes the shift specifically to AI. Staffing Industry Analysts reported on September 3, 2026, that temporary staffing hours for IT occupations rose 1.4% year over year in the first quarter, 4.3% in the second and 5.3% in July and August, based on its SIA and Bullhorn Staffing Indicator.
SIA forecasts IT staffing revenue to grow 2% in 2026 to $38.1 billion, and a net 35% of IT staffing firms in its US Staffing Industry Pulse Survey reported rising bill rates over the previous three months. The article lists AI transformation projects among the areas driving demand, alongside data, cloud, cybersecurity, application development and integration.
The broader contingent market is large and dominated by project work. In an August 28, 2026 analysis, SIA valued the Americas contingent workforce at $4.6 trillion in 2025, with $4 trillion in the US, and statement-of-work contracts made up 73.3% of revenue. Adam Pode, SIA's senior research director, said the contingent workforce is "becoming less of a tactical response to peaks in demand."
The pattern fits a market where employers are unsure how many permanent seats AI work will need. Contract and statement-of-work capacity lets them deliver the project now and decide on permanent headcount later. That reading is ours, and it is also our business, which is exactly why we disclosed our interest at the top of this post.
How Do You Turn Open AI Roles Into Hires?
Start by treating time to fill as a delivery risk, not an HR metric. Every role on your plan should have a target start date tied to a roadmap milestone. These six practices are the ones we see shorten the gap most often, whether the seat ends up permanent, contract, or a mix of both.
Write the role around the work, not the title. Describe the problems the person will solve in the first six months, the systems they will touch, and the AI tools they will use. A clear scope shortens interviews and attracts candidates who can picture the job. Vague postings asking for every AI skill at once tend to stall, in our experience.
Shorten and structure the interview loop. Agree on the scorecard before the first interview, cap the number of rounds, and give every interviewer a defined area to assess. Experis's finding that employers prize human skills most is a reason to test them deliberately, not leave them to impressions. Our guide to vetting remote engineers covers verification.
Fix the salary band before you post. Benchmark the role against current data before it opens, not after the first offer is declined. If the budget cannot reach the market rate for the seniority you need, change the scope or the seniority of the role rather than leaving it open for months at a number no candidate will accept.
Set a deadline for borrowing capacity. Decide in advance how long a critical role can stay open before you bring in contract or partner capacity, for example, after 60 days without a finalist. That threshold is a planning choice, not an industry benchmark. Our AI staffing models guide compares the options.
Look inside before you look outside. In our experience, people who already know your systems, data, and customers can often move into AI-heavy roles faster than an external hire can ramp up, if you fund the training. Our 2027 AI skills plan guide covers when to build skills internally, buy them, or borrow them.
Use contract-to-hire for seats you cannot size yet. When you are unsure whether AI work needs a permanent seat, a contract engagement with a conversion option lets you learn before you commit. Dedicated AI teams and AI talent solutions are how we provide that capacity.
A caution about our own corner of the market: we found no independent 2026 data on Latin American AI engineering supply, pay, or time to hire that matches the US sources above. Any nearshore provider quoting precise savings or fill times, us included, is quoting its own book of business. Ask for those numbers in writing and check them against your own results.
Common Questions About Tech Job Postings and Hiring
Are tech job postings really at a three-year high?
Yes. CompTIA reported 625,633 active tech job postings in September 2026, the highest total in more than three years and up 22.4% year over year. Active postings requiring AI-related skills reached 349,821. Postings measure employer intent, though, and CompTIA estimated that tech occupation employment fell by roughly 6,000 the same month.
Why are companies posting jobs but not hiring?
Several factors overlap. Experis found the US tech hiring outlook fell 10 points for Q4 2026 as employers grew more deliberate. Roles now combine AI and human skills, which narrows the candidate pool, and low quit rates mean fewer experienced people are changing jobs. In our experience, outdated salary bands also stall many offers.
What does federal data show about tech hiring?
The closest federal measure is the BLS JOLTS survey. In August 2026, the information industry made 44,000 hires, down from 62,000 in July and 69,000 a year earlier, while openings rose to 123,000. The information industry is a proxy, not a match for tech occupations, and monthly figures are preliminary.
Should companies use contractors for AI roles?
For roles that are urgent or hard to size, contract capacity can be a sensible bridge. Staffing Industry Analysts reported IT temporary staffing hours up 5.3% year over year in July and August 2026. Set a clear time-to-fill threshold, and consider contract-to-hire so you can convert strong people once the need is clear.
If you have AI roles that have been open too long and want a second opinion on the scope, the loop, or the right mix of hiring and contract capacity, talk to the Golabs team. We will start with the work, and we will tell you when hiring is the better answer.