When a delayed jobs report finally dropped Tuesday, economists and investors got their first real look under the hood of the U.S. labor market, and the engine is stalled. Payroll growth was modest in November at 64,000, while October showed a net decline of roughly 105,000 jobs, and the unemployment rate rose to a four-year high of 4.6%. Payroll growth hasn’t collapsed, but it hasn’t meaningfully advanced, either. The result is a labor market that’s drifting sideways, quietly losing momentum beneath the surface.

Economists say that kind of stall is more dangerous than it looks.

“There’s just no forward motion,” Moody’s Analytics chief economist Mark Zandi told Fortune. Job gains bounce slightly from month to month, but net hiring has gone essentially nowhere this year, he said, leaving the labor market “stuck in the mud.”

That stagnation explains why unemployment has continued to rise despite weak labor-force growth. Typically, joblessness climbs when layoffs surge or hiring freezes abruptly. This time, with neither happening, the economy has instead been failing a weaker benchmark, unable to create enough jobs just to absorb even modest population growth.

The dynamic mirrors a warning from analysts at Goldman Sachs, who suggested in October that the U.S. is settling into a phase of “jobless growth” where output rises despite flat hiring. Economists David Mericle and Pierfrancesco Mei wrote that productivity has essentially been doing the job of labor, echoing a prior analysis by Bank of America Research chief U.S. equity analyst Savita Subramanian. As employers increasingly turn to AI to reduce labor costs, this stalled period might turn into a  “a potentially long-lasting headwind to labor demand,” the Goldman economists wrote. 

The unemployment rate has risen by roughly six-tenths of a percentage point since the start of the year, a move that Zandi said carries weight, even if it unfolds gradually.

“You wouldn’t see unemployment rising if labor demand were okay,” Zandi said. “This tells us demand is weak too.”

At the same time, the economy is still growing. Output continues to expand, supported by what Fed Chair Jerome Powell has called “structural productivity gains” and heavy investment in artificial intelligence, which has allowed companies to produce more without adding much headcount. That dynamic has helped keep GDP positive, but it has also masked a labor market that is no longer providing the engine of growth it once did.

One of the clearest signs of that strain appeared beneath the headline payroll numbers. The number of people working part time for economic reasons jumped by nearly 1 million in November, rising to 5.5 million, as more workers reported having their hours cut or being unable to find full-time jobs. 

Businesses are “doing everything they can to avoid laying off workers,” Zandi said, noting that trimming hours and leaning more heavily on part-time or temporary labor is often the first step when demand begins to soften. He cautioned that the size of the increase was likely overstated by data noise related to the recent government shutdown, which disrupted survey collection. Even so, the direction of the decline is consistent with a broader cooling in labor demand.

Private-sector hiring, meanwhile, remains positive but weak. November’s gains offered little reassurance, and upcoming revisions could further soften the picture. Once those adjustments are made, Zandi expects overall job creation to look even closer to flat.

“It’s not hemorrhaging,” he said. “But it’s not creating jobs, either. It’s basically going sideways.”

That kind of stall can be as risky as an outright downturn. Rising unemployment tends to weigh on confidence, and over time that pressure can bleed into consumer spending. 

“The risks of the economy going into recession are uncomfortably high,” Zandi said. 

For now, the economy has avoided that outcome, in part because the AI boom has propped up investment and boosted household wealth through higher stock prices. Harvard economist Jason Furman even recently calculated that without investment into data centers, GDP growth would have been at a near standstill in the first half of 2025. Subtract that, and the economy would have to find another driver or, as Zandi suggests, run the risk of tipping into recession.

“We’re on the edge,” Zandi said. “We haven’t gone over yet. If that boost from AI wanes, then we’ve got a problem.”

“The modest job growth alongside robust GDP growth seen recently is likely to be normal to some degree in the years ahead,” Goldman’s economists also warned in October, speculating that many currently occupied jobs don’t actually need to be filled with human workers and the real toll won’t become apparent until companies get the cover provided by a recession to reduce force en masse. “History also suggests that the full consequences of AI for the labor market might not become apparent until a recession hits.”