America's Economy Is Accelerating With the Brakes Still On

Analysis

The American economy accelerated in the third quarter despite costly energy, elevated financing rates and war-related oil disruptions. The Energy Information Administration documented oil prices rising alongside Middle East fighting. The larger question is what happens if those pressures ease just as investment in artificial intelligence begins producing lasting gains across the economy.

The Bureau of Economic Analysis reported 3.4% annualized real GDP growth in the third quarter of 2026, up from 2.2% in the second. An especially revealing measure, real final sales to private domestic purchasers—what households and private businesses buy, excluding inventories and government—grew at a 4.9% annualized rate.

Manufacturing indicators are also strengthening. The Census Bureau measured $677.2 billion in September factory orders, 1.8% above August. Orders are measured in current dollars, so part of that increase may reflect prices rather than more goods. The Institute for Supply Management reported its manufacturing index at 55.5 in October, the tenth consecutive month of expansion, with production at 59.0. Readings above 50 generally signal expansion; a production reading of 59.0 indicates that more surveyed manufacturers reported rising output than falling output.

The first act: Building the new economy

AI is often described as software. Its expansion is increasingly a physical construction project.

A new AI data center needs semiconductors, transformers, power generation, transmission equipment, cooling systems, networking hardware and buildings. Producing those inputs creates demand for industries far beyond the technology companies paying the bills. The Congressional Budget Office identifies generative AI as one driver of U.S. business investment in 2026. AI infrastructure is therefore one plausible source of demand for machinery and electrical equipment, though the aggregate factory-order figures do not isolate AI-related purchases.

That investment can lift GDP immediately, before a single customer realizes the full productivity benefit of AI. It is construction and capital formation in the present, with the possibility of a much larger return in the future.

Cost pressures remain important. The ISM survey's prices index was 75.2 in October, down from 77.9 in September but still signaling widespread increases in factory input prices. The index measures how widely prices are rising, rather than how much they have risen. Interest expenses raise the hurdle for new projects, while disruptions to oil markets affect transport and household budgets.

The second act: What if the investment starts paying for itself?

Consider an engineer who can design more equipment in a workday, a factory that reduces downtime with predictive maintenance, or a logistics company that moves the same freight with less wasted fuel and labor. In each case, the economy can produce more without needing a proportionate increase in workers or resources.

That is productivity growth: more real output for each hour worked. It is the mechanism that could raise America's economic speed limit.

The Bureau of Labor Statistics reported that nonfarm business labor productivity was 2.2% higher than a year earlier in the second quarter. Growth within that quarter was slower, at 1.4% annualized. Productivity is improving, but it has not reached the pace needed to support 4–5% GDP growth year after year with a slowly growing workforce.

The arithmetic clarifies the ambition. BLS projections put long-run labor-force growth near 0.3% annually. At that pace, sustained 4% real GDP growth would require approximately 3.7% yearly output-per-worker growth; sustained 5% growth would require roughly 4.7%. These are illustrative scenarios, not forecasts.

History supplies a useful comparison. BLS estimates the U.S. economy averaged 3.4% real GDP growth from 1995 to 2005, a decade that benefited from an earlier information-technology revolution. The next AI-driven productivity era could prove even more powerful—but only if the benefits spread beyond the companies building and selling the infrastructure.

The economic stakes

A 2% economy grows about 22% over ten years. A 4% economy grows 48%, while a 5% economy grows 63%. The difference is not a temporary sugar rush; it is a much larger capacity to pay wages, improve living standards, support public services and finance investment.

The Congressional Budget Office projects average real growth of 1.8% from 2027 through 2036; the BLS long-range projection is 2.1% for 2025–35. Sustained 4% growth would therefore require a genuine productivity break from either baseline, not simply lower oil prices or a few interest-rate cuts.

Lower energy costs, calmer trade routes and more affordable financing could help unleash projects that are already economically close to viable. The larger opportunity requires the combination: fewer external constraints and persistent, measurable gains in productivity.

What would change our view

Watch for productivity spreading into healthcare, manufacturing, logistics and ordinary business services; stronger real wages; and higher output without a renewed inflation spiral. If AI spending keeps growing but unit costs fail to fall, electricity shortages intensify or household purchasing power continues to erode, the investment boom may be much less transformative than its construction activity suggests.

America is accelerating with the brakes still on. The next question is whether AI will build a bigger engine.

Primary sources

BEA, third-quarter 2026 GDP advance estimate , October 29, 2026.

Census Bureau, manufactured-goods new orders , September 2026 data released November 3.

ISM, October 2026 manufacturing report , November 2026.

BLS, Q2 2026 productivity and costs , September 3, 2026.

EIA, third-quarter 2026 crude oil prices and Middle East conflict , October 5, 2026.

BLS, 2025–35 labor-force and productivity projections , October 2026.

CBO, Budget and Economic Outlook 2026–2036 , February 2026.