US job openings fell to their lowest level in over five years in December 2025, according to the Bureau of Labor Statistics’ Job Openings and Labor Turnover Survey. The number of job openings decreased by 386,000 to 6.542 million, marking the lowest point since September 2020. Additionally, data for November was revised downwards, showing 6.928 million openings instead of the previously reported 7.146 million. Despite a rise in unemployment claims, which increased by 22,000 to a seasonally adjusted 231,000 for the week ending January 31, the overall labor market remains stable. Chief economist Carl Weinberg noted there are no signs of widespread layoffs typically associated with a weakening labor market, indicating a “low hire, low fire” environment.
Why do we care?
The job openings data isn’t about recession. It’s about your end customers freezing hiring while maintaining headcount.
Here’s the specific harm I’m watching for. MSPs who built their growth models on seat expansion—new employees mean new devices, new licenses, new onboarding projects—are forecasting into a headwind they haven’t recognized. Your customers aren’t posting job openings. That means your “new seat” pipeline is fiction.
But there’s a second-order effect that’s more dangerous. The “low hire, low fire” environment creates productivity pressure. Companies are asking existing staff to do more.
That should create demand for automation and AI-assisted workflows—exactly what partners are chasing with those ad hoc AI practices as we just covered.
Except here’s the trap: your customers want productivity tools but they’re also in capital preservation mode. They’ll buy efficiency gains framed as operational expense. They won’t approve capital projects requiring budget committee sign-off.
So you have partners with immature AI practices trying to sell into customers who want AI productivity gains but won’t fund capital projects. The mismatch is structural.
The MSPs who navigate this will do two things. First, reforecast without headcount growth assumptions—your seat-based pipeline is unreliable. Second, reframe everything as operational expense with measurable productivity ROI.