Policy

AI Investment Adds 1% to U.S. GDP Growth, CRS Warns of Bubble Risk

Congressional researchers say the infrastructure boom is fueling expansion but straining electricity grids, raising import bills, and potentially outpacing economic returns.

Omega Editorial · November 16, 2026 · 3 min read

Key takeaways

  • AI investment added approximately 1 percentage point to U.S. GDP growth in Q2 2025, according to Congressional Research Service estimates.
  • Imported AI equipment and supply constraints in semiconductors and electricity are creating inflationary pressure and limiting net domestic economic gains.
  • A Bank for International Settlements paper suggests investment may already exceed economically efficient levels by 50%, raising bubble concerns.
  • House-passed legislation requiring data centers to pay for their electricity infrastructure costs has stalled in the Senate.
  • CRS warns that if AI spending outpaces returns, resulting defaults and bankruptcies could ripple through the U.S. and global economies.

AI spending drives measurable economic expansion

Artificial intelligence investment contributed approximately 1 percentage point to U.S. economic growth in the second quarter of 2025 and 0.7 percentage points in the first quarter of 2026, according to a Congressional Research Service analysis. The figures mark a notable acceleration from prior years, though the agency cautions that alternative measurement approaches could yield different results.

The expansion reflects a surge in data center construction, semiconductor manufacturing facilities, and related infrastructure. Yet the growth comes with complications: much of the equipment fueling the boom is imported, and because imports subtract from gross domestic product calculations, the net contribution to domestic output is smaller than raw investment figures suggest. Federal Reserve researchers found global trade in AI-related products jumped 65% in the first half of 2025 compared to the previous year.

Why it matters

The CRS report arrives as policymakers debate whether the AI buildout represents sustainable economic transformation or speculative overinvestment. If spending outpaces the technology's ability to generate returns, the resulting financial losses could trigger defaults, bankruptcies, and broader economic contraction. The infrastructure demands are already forcing difficult questions about who pays for expanded electricity capacity and whether competition is driving capital allocation beyond economically efficient levels.

Inflation and interest rate pressures emerge

The infrastructure push is creating supply constraints in semiconductors and electricity, adding inflationary pressure. Goldman Sachs estimates these bottlenecks could increase inflation by as much as 0.5 percentage points at their peak, though the Congressional Research Service notes the effect is expected to be temporary.

Much of the expansion is debt-financed, potentially pushing interest rates higher. The CRS report indicates this borrowing contributes to a widening U.S. current account deficit, though Federal Reserve Bank of New York researchers attribute most of the deficit growth to federal budget shortfalls rather than business investment patterns.

Bubble concerns and economic efficiency

A Bank for International Settlements working paper estimates AI investment could reach 1.5 times the economically efficient level, suggesting competitive dynamics may be driving spending beyond what underlying demand justifies. Goldman Sachs projects AI investment will hit 2.8% of U.S. GDP by 2028.

The Congressional Research Service warns that if the boom develops into a financial bubble, the eventual correction could impose substantial losses on investors and lenders while reducing investment across the broader economy.

Congressional response stalls in Senate

The House passed legislation in September 2026 addressing one consequence of the expansion: the burden new data centers place on electricity ratepayers. H.R. 9340, the Ratepayer Protection Act, would require state utility regulators to ensure large electricity customers bear the additional costs their demand creates. The bill passed 417-3 but failed a procedural vote in the Senate, falling short of the 60 votes needed to advance.

The Congressional Research Service concludes that the long-term economic effects will depend on whether companies can generate sufficient returns to justify current spending levels without triggering financial losses or a broader slowdown.

These findings were first reported by Legis1, citing the Congressional Research Service report.

  • #ai investment
  • #economic growth
  • #congressional research service
  • #infrastructure spending
  • #financial bubble risk
  • #data centers

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Policy