Everyone loves the AI productivity narrative, but Goldman Sachs is asking the tough question: when does this become a labor market problem?
According to a note from strategist Lee Coppersmith, the transition from a productivity story to a labor story is now underway. While the headline jobs numbers have held up so far, the underlying trend is clear. Companies are accelerating adoption and explicitly framing AI as a mechanism to slow hiring or cut costs entirely.
Goldman sees this becoming visible in the data next year. Their base case calls for job losses in AI-exposed industries to run at a clip of roughly 20,000 per month in 2026 though it’s “potentially” offset by AI-related job creation elsewhere.
“So far, the impact on headline employment has been modest, but the direction is clear. Adoption accelerated last year, particularly in roles most amenable to automation, and many companies are explicitly framing AI as a tool to reduce labor costs or slow hiring,” he writes.
For markets, this is the danger zone. Coppersmith argues that the uncertainty surrounding this shift is one of the few macro risks that isn’t priced in. If the labor market starts to crack because of automation rather than the business cycle, it adds a layer of complexity to the macro outlook that traders aren’t currently hedging for.
I tend to think there are some opportunities in high-cost, low margin industries. Think old economy stuff and places with high human costs, like banking and insurance. One company already seems to be aggressively lowering headcount: Amazon. So far that hasn’t helped the shares but at some points, doing the same thing with less workers add to the bottom line.
Couple that with a Fed that’s hell-bent on cutting rates if unemploment rises and there’s a compelling case to eventually buy a dip.
