by Daniel Brouse
December 13, 2025
Introduction
The U.S. economy in 2025 is shaped by three powerful and interacting forces: tariff implementation, restrictive immigration policies, and the rapid adoption of artificial intelligence (AI). Each has distinct economic effects, but taken together they create a complex macroeconomic picture with both upside and risk.
1. Tariffs: A Drag on Growth
Tariffs function effectively as taxes on imported goods and inputs. Higher tariffs increase costs for U.S. producers and consumers, reduce trade volumes, and create uncertainty in supply chains. Empirical research shows that higher tariff rates tend to be negatively correlated with real GDP growth over time, with lagged effects that can take months or years to fully materialize.
Recent macroeconomic forecasts attribute part of the slowdown in U.S. GDP growth to tariff effects. The Organisation for Economic Co-operation and Development (OECD) notes that rising tariff rates and lower net immigration have weakened growth momentum, even as other sectors contribute positively.
Tariffs also distort reported GDP through accounting effects. When imports fall sharply—as has happened recently—net exports (exports minus imports) mechanically add to GDP growth even when underlying domestic demand remains weak.
Taken together, tariffs are acting as a drag on growth, raising costs and undermining business and consumer demand, even as headline numbers may temporarily mask underlying weakness.
2. Immigration Policy: Labor Market Constraints
Restrictive immigration policies reduce labor supply directly and indirectly. While precise impacts are difficult to quantify in real time, recent scholarly and policy analysis suggests immigration contributes significantly to labor force growth, especially in sectors such as technology, construction, and services.
For example, forecasts indicate that both tariff hikes and lower net immigration contributed to a slower labor market, reducing payroll growth and increasing unemployment rates.
Immigrants accounted for an estimated 80% of U.S. GDP growth from 2020–2025, underscoring their outsized contribution to labor supply and overall economic expansion. A sharp reduction or elimination of immigration therefore constrains the labor force, lowers potential output, and slows GDP growth. While official estimates vary, the underlying mechanism is well established in macroeconomic models and supported by recent forecasts: reduced immigration leads directly to weaker economic growth.
Equally important is the spending behavior of immigrant households. A large share of earned income is spent quickly on goods and services, providing an immediate boost to aggregate demand. In recent months, multiple consumer-facing industries have reported noticeable declines in sales following reductions in immigrant participation in the economy, including beer brands, food producers, and businesses tied to cultural and community celebrations.
3. AI Adoption and Capital Spending: A Temporary Boon?
AI-related investment—particularly in data centers, hardware, and information technology—has become a dominant driver of U.S. GDP growth. Multiple analyses find that AI infrastructure spending alone explained an exceptionally high share of reported GDP growth in 2025. Some estimates show AI-related data center investment accounted for as much as 92% of total GDP growth in the first half of 2025, despite representing a much smaller share of overall GDP.
Capital spending in AI is substantial: hyperscale firms like Microsoft, Google, Meta, Amazon, and Nvidia are expanding data center capacity and hardware investment by the hundreds of billions of dollars annually. These investments act like short-term stimulus, lifting real GDP growth above what it would have been otherwise.
However, long-term returns on all that AI investment remain uncertain from the standpoint of actual productivity realized in final goods and services. Goldman Sachs analysts note that significant portions of AI investment aren’t fully reflected in official GDP statistics because of how intermediate inputs are treated, making accurate measurement difficult.
Moreover, traditional economic models suggest that capital spending without commensurate returns can lead to misallocation, just as excessive leverage did before past financial crashes. If AI capital expenditures do not generate sustainable earnings and broader productivity gains, the long-term macro impact may be weaker than headline growth suggests.
Perhaps most concerning is the emergence of circular financing among hyperscalers. Companies such as NVIDIA are effectively extending credit—directly or indirectly—to other large technology firms so they can purchase NVIDIA’s chips and related infrastructure. With no proven, scalable business model yet established for monetizing many AI applications, this financing structure carries significant risk and could result in losses measured in the hundreds of billions, if not trillions, of dollars. Reflecting these concerns, industry analysts now estimate that at least one-third of all announced data center projects are unlikely to ever be completed.
4. What It All Means
Taken together, these three forces create a fragile economic structure:
- Tariffs slow trade, raise prices, and dampen business investment.
- Immigration restrictions weaken labor supply and potential output.
- AI spending is boosting GDP figures in the short term, but its sustainability and true economic impact remain heavily debated. The industry’s highly leveraged structure raises the risk of a bubble bursting, potentially triggering sharp declines in stock prices. Recent weeks of heightened market volatility suggest that investors are beginning to recognize and price in these risks.
The Organization for Economic Co-operation and Development now projects modest GDP growth in 2025 and 2026, in part because of strong AI-related investment, but also warns that tariff and immigration effects will continue to weigh on growth.
Without AI investment, the economy would likely be much weaker and possibly in contraction, as private domestic demand outside technology sectors has struggled to expand.
5. Risks on the Horizon
There are several risks:
- Investment sustainability: If hyperscale investments fail to deliver strong returns, both corporate profits and stock valuations could suffer.
- Leverage and risk: Much of the market’s valuation of AI-linked firms depends on optimistic long-term returns on borrowed capital.
- Labor displacement: AI-driven automation could reduce employment opportunities without corresponding growth in new job categories.
- Collapse of circular financing: If one hyperscaler defaults on obligations to another, the interconnected debt structure could trigger a cascading, domino-style failure across the sector.
These risks could amplify the negative effects of tariff-induced cost pressures and labor shortages, creating headwinds for long-term economic resilience.
Conclusion
In 2025, the U.S. economy appears to be a highly concentrated growth story: dependent on AI-related capital spending to offset structural headwinds caused by tariffs and restrictive immigration policies. While headline GDP growth persists, underlying fundamentals point to slowing labor markets, rising costs, and a high degree of reliance on a narrow set of technology investments. The sustainability of this model remains uncertain and will depend critically on whether AI investments translate into broad-based productivity gains and meaningful returns in the years ahead.