Ask what’s driving the US economy forward in 2026, and the honest answer isn’t consumer spending, the traditional engine of American growth. It’s a much narrower story than that. According to Treasury Department data, AI-related investment accounted for roughly half of all US GDP growth in the first quarter of the year alone.
That’s a genuinely unusual place for the world’s largest economy to be. For most of the post-pandemic recovery, household spending did the heavy lifting. In 2026, leadership has shifted toward businesses pouring money into data centers, computing infrastructure, and software, a transition that’s reshaping what “economic growth” actually looks like under the hood, even when the topline number looks reasonably healthy.
The Headline Numbers, and Why They’re Deceiving
Most forecasts put US GDP growth for 2026 somewhere in the 1.8% to 2.3% range, modest by historical standards but still ahead of most other developed economies. The UK is tracking closer to 0.8%, the eurozone around 1.1%, Japan under 1%. On paper, the US looks like it’s holding up better than its peers.
Underneath that headline number, the composition tells a more uneven story. Business investment rose more than 10% in the first quarter, driven heavily by spending on new equipment and intellectual property, largely AI infrastructure. Meanwhile, consumer spending, historically the backbone of US growth, has been showing real strain. The personal savings rate fell to an unusually low 2.6% of after-tax income earlier this year, and inflation has continued to outpace wage growth for a lot of households, creating what economists are calling a real income squeeze: prices rising faster than paychecks, even as the stock market and asset prices for wealthier households climb.
That divide matters. A lot of the recent growth in consumer spending is coming disproportionately from higher-income households benefiting from rising asset values, while lower- and middle-income consumers are increasingly leaning on savings and credit to maintain their spending, a pattern that becomes harder to sustain the longer it continues.
Why AI Investment Specifically Is Doing So Much Work
The scale of AI-related spending right now is hard to overstate. It’s not just tech companies. Data center construction, energy infrastructure to power them, and the semiconductor and hardware supply chains feeding all of it have become a genuinely significant slice of the entire US economy’s growth story, not a side note within it.
Whether that investment is actually translating into productivity gains for the broader economy yet is a separate, more uncertain question. Most economists studying past technology cycles, the personal computer boom of the late 1980s and 90s is the most commonly cited comparison, point out that meaningful productivity gains from a new technology wave tend to show up years after the initial investment surge, not immediately alongside it. If that pattern holds, current AI spending may be laying the groundwork for productivity improvements that show up more clearly toward the end of the decade rather than right now.
That’s a reasonable long-term bet. It’s also a genuinely fragile short-term foundation. An economy where roughly half of growth traces back to one investment category is more exposed to a slowdown in that specific category than a more broadly distributed growth pattern would be. If AI-related capital spending cools, whether from a shift in corporate sentiment, financing costs, or simply investment cycles maturing, growth doesn’t have an obvious, equally large replacement source lined up behind it.
The Fed’s New Approach Is Adding Another Layer of Uncertainty
Kevin Warsh’s debut as Fed Chair this year has brought a noticeably different communication style than his predecessor, shorter statements, more policy discretion, and less forward guidance about where rates are headed next. For an economy this dependent on one investment category, a less predictable Fed adds a second layer of uncertainty on top of an already narrow growth base.
Inflation isn’t cooperating either. Core inflation is expected to land somewhere around 2.5% to 2.6% by the end of 2026, still above the Fed’s target, with tariffs and immigration-related labor constraints both contributing to price pressure from different directions. That combination, sticky inflation alongside investment-heavy but consumer-weak growth, is part of why some forecasts describe this as a “slow-growth, sticky-inflation” environment rather than a clean recovery story.
Why This Matters Beyond Economists’ Spreadsheets
For everyday households, this composition shift helps explain something that might otherwise feel confusing: headline GDP growth sounding fine while affordability remains one of the most consistently cited concerns in surveys heading toward the midterms. GDP growth and how an average household’s budget actually feels aren’t the same measurement, and in 2026 specifically, that gap is wider than usual because so much of the growth is concentrated in corporate AI infrastructure spending rather than the kind of broad-based wage and consumer spending growth that tends to be felt more directly.
It’s also worth watching simply because of what it signals about risk. An economy this reliant on one investment theme carrying growth is a economy where that theme’s health, AI capital spending specifically, deserves more attention as an economic indicator than it would in a more typical, broadly distributed expansion. Whether that concentration turns into the foundation for a longer boom or a narrower, more fragile growth story than the headline numbers suggest is likely to be one of the defining economic questions of the back half of this decade.
