AI Data Center Boom: Goldman Sachs Says Crowding Out Is Minimal 2026

The AI data center boom is pouring hundreds of billions of dollars into construction, and Goldman Sachs economists have now quantified how much that spending is crowding out other parts of the US economy. Their conclusion: the effect is real but small, and both the doomsayers and the cheerleaders are exaggerating.

According to an analysis by Jessica Rindels and David Mericle cited in Axios Macro, the massive capital expenditure on data centers does divert money, workers, and materials from other uses. But the scale of that diversion is far below what the loudest warnings suggest.

What Goldman Sachs Actually Found

Goldman identified two main channels through which AI investment displaces other economic activity. The first is physical: construction workers, concrete, steel, and land used for data centers are not available for other projects. The second is financial: hyperscalers have ramped up debt issuance tied to AI, which pushes up borrowing costs across the corporate sector.

The key number is that the financial crowding-out effect amounts to just 0.05% in higher financing costs. That is a rounding error for most businesses, not a suffocating burden.

• Physical crowding out is limited to specific local labor and materials markets, not the whole economy.
• Financial crowding out adds only about 0.05% to corporate borrowing costs.
• Claims that AI is driving most GDP growth are overstated.
• Claims that AI is strangling the rest of the economy are also overstated.
• The net effect is a modest reallocation, not a crisis.

Why the 0.05% Figure Matters

In a $27 trillion US economy, a 0.05% increase in financing costs is trivial at the aggregate level. For a small business taking out a loan, it might mean a few hundred dollars more per year on a $1 million credit line. That is not nothing, but it is far from the apocalyptic scenario some commentators have painted.

Goldman's analysis suggests that the AI investment wave is large enough to be visible in national accounts but not large enough to starve other sectors of capital. The bond market has absorbed the surge in AI-related debt without a broad surge in corporate yields.

The Two Competing Narratives, Fact-Checked

On one side, AI optimists argue that data center spending is a major engine of GDP growth. On the other, skeptics warn that this capex boom is cannibalizing investment in housing, manufacturing, and other productive assets. Goldman's numbers land in the middle.

The upshot is that the AI investment boom is a real phenomenon with real opportunity costs, but it is not the economy-swallowing monster that some fear. Nor is it the sole savior that others claim.

Who Is Affected by the AI Capex Boom

Construction firms in data center hubs are seeing labor shortages and higher wages, which is good for workers but costly for other builders. Utilities in those regions face rising electricity demand, which can push up local power prices. And companies that rely on debt markets are paying slightly more, though the effect is small.

For most consumers and businesses outside tech-heavy regions, the day-to-day impact is negligible. The AI boom is shaping the economy at the margin, not at the core. The background to this story is set out in Aragón Data Centers Boom: Where Does the Tax Money Go. A related thread runs through Goldman Sachs Executives Share $500M Bonus Pot in Record Payout. Earlier coverage in Spain Softens Data Center Rules to Attract AI Investment shows how the argument developed.

What to Watch Next

The picture could change if AI capex continues to accelerate. If hyperscalers keep issuing hundreds of billions in debt, the 0.05% figure could grow. If data center construction spreads to more regions, local labor and materials markets could tighten further. A closer look at the same subject appears in Waste Heat From Cars and Data Centers Could Power Cooling. Anyone following the wider picture can start with Data Centers Sell Waste Heat as Nearly Free Heating for Cities. The same theme surfaces in AI Investment Needs 9% of US GDP to Pay Off, Brookings Finds, which takes a different angle.

Goldman's analysis provides a useful baseline: for now, the crowding out is modest. But it is worth monitoring as the AI investment cycle matures. The debate over AI's economic impact is far from settled, but this report adds a crucial data point. Context on how this played out before sits in Australia Copyright Laws: AI Datacentre Investment Sparks Debate. A companion piece, Jeff Bezos Invests in AI Startup CuspAI for Materials Discovery, approaches the same subject from another angle.