President Donald Trump played cheerleader for the US economy in a prime-time address last night. But in the end, facts on the ground will matter more than words on a teleprompter.
And what might those facts be? Here’s one take on the coming year: In a newly released 2026 outlook, “Sturdy Growth, Stagnant Jobs, Stable Prices,” the bank Goldman Sachs forecasts US GDP growth of 2.6 percent, well above the 2.0 percent consensus and 2025’s projected 2.1 percent, alongside inflation easing back toward the Federal Reserve’s two percent target. (Goldman argues that once temporary tariff effects wash out, underlying inflation is already close to target.)
What’s more, a) the tariff drag should fade, b) consumers are set to receive roughly $100 billion in tax refunds from the One Big Beautiful Bill Act, and c) easier financial conditions—driven by Fed rate cuts, deregulation, and spillovers from the AI investment boom—are providing an additional growth impulse. So some tailwinds.
Most interesting, perhaps, is what isn’t doing the economic lifting. Don’t tell Time magazine, but artificial intelligence so far barely registers in measured GDP. Goldman estimates that AI’s direct contribution to the level of output is at just 0.2 percent. Even treating semiconductor purchases as investments rather than intermediate inputs lifts that to only 0.3 to 0.4 percent. Also, because most AI hardware is imported, data center construction can actually subtract from headline growth in the national accounts.
That doesn’t mean AI is an economic nothingburger. Hardly. But its topline impact is showing up first in the labor market, and not even a whole lot there. A survey of the bank’s investment-banking clients points to expected headcount reductions of four percent over the next year and 11 percent over three years. So far, the pain has been narrowly concentrated in tech, where employment has slipped below its pre-ChatGPT trend. But broader cost-cutting could follow, the bank warns.
That said, Goldman expects the jobless rate to stabilize around 4.5 percent in 2026, but offers this caveat: “In fact, we could easily imagine further unemployment rate increases in the near term if either productivity-enabling AI applications arrive more quickly than expected or company management teams increase their focus on lowering labor costs in 2026.” But for now, no signs of an impending robopocalypse.
Thankfully, Goldman also argues that AI will eventually deliver meaningful productivity gains, justifying continued capital spending by business. (Don’t forget: Productivity growth is the long-run driver of higher income and living standards.) Yet the stock market may already have priced in a large share of those future benefits, especially for firms building models and infrastructure.
So the risk: If monetization disappoints, takes longer than hoped, or if AI ultimately delivers more value to the companies that use the technology rather than those that build and sell it, today’s lofty valuations leave little margin for error, the bank concludes.
There’s your comprehensive bubble risk, I guess. AI may yet transform the economy—but markets appear to be pricing in the ending of that techno-optimist story before the middle.
The post Why AI Is Juicing Stocks, Not GDP (Yet) appeared first on American Enterprise Institute – AEI.
