United States Economic Forecast: Q3 2026
AI investments keep US economic growth in gear, for now
The AI buildout helps the United States to maintain relatively strong growth despite near-term inflation and demographic headwinds
Article
•
6-min read
•
29 September 2026
•
Deloitte Global Economics Research Center
Since our last edition, we have continued to see a rapidly changing global economic and policy environment and recognize that conditions remain highly fluid. So, the three scenarios we present in this forecast are not meant to be precise estimates of where the US economy will end up. Instead, they are built on explicit assumptions to help guide thinking on the future.
Our baseline forecast reflects our best assessment of the path economic variables will take. Our downside and upside scenarios reflect plausible alternatives for the US economy should our assumptions prove to be either too optimistic or too pessimistic, respectively.1
Scenarios
Baseline
Higher energy costs, rising interest rates, weaker population growth, and geopolitical tensions have thus far been insufficient to significantly derail growth in real gross domestic product, which was up 2.1% from a year earlier in the second quarter. Domestic demand in the private sector, that is, real final sales to private domestic purchasers—was even stronger at 2.7%.
The strength of the economy is primarily due to business investments related to artificial intelligence and the positive wealth effects from strong equity price gains. As a result, we have raised our outlook for business investment, consumer spending, and therefore economic growth.
While we anticipate stronger growth compared to our June forecast, we still expect real GDP growth to moderate by a bit over the next two years. Although strong, the rate of growth in business investment is expected to begin moderating in 2027. Consumer spending is likely to slow as well. We expect elevated inflation and rising interest rates to weigh on both investment and consumer spending. Meanwhile, consumer spending growth has outpaced post-tax income growth for more than two years, suggesting consumers are likely to pull back on spending soon.
The public sector is also expected to provide a modest headwind to growth in the near term. Fiscal policy had been a tailwind at the start of this year, with tax cuts supporting income and spending growth. However, it is expected to turn more contractionary as certain tax credits are phased out.2 International trade will also do little to spur growth as relatively modest global growth and trade tensions weigh on exports. We expect real GDP growth to slow modestly from 2.1% in 2026 to 1.9% in 2028.
Beyond the next two years, real GDP growth is expected to accelerate, supported in part by anticipated improvements in productivity associated with AI investments. At the same time, we expect inflation to move lower, bringing interest rates down as well, which will further fuel the economy. We expect real GDP growth to accelerate to 2.3% by 2031—up slightly from the 2.2% growth forecasted last year.
Downside: Open-weight and foreign models outcompete
Our downside scenario assumes that the return on AI-related investments in the United States is far lower than what investors currently expect. Consumers and businesses primarily opt for lower-cost, open-weight models for their needs. In addition, we assume foreign models gain market share relative to US models in much of the rest of the world, limiting external demand for US models.
In this scenario, we assume an equity-market correction of 35%, bringing the cyclically adjusted price-to-earnings ratio down to its average between 2003 and 2005. Beginning in the third quarter of 2027, business investment declines about 12% over the course of two years, pushing investment levels down to where they were roughly three years prior to the contraction. Default rates rise on the rapidly growing debt that was issued to finance the AI buildout. This initially contributes to a broad tightening of financial conditions as loans are written down and lenders become more risk-averse.
Just as positive wealth effects supported consumer spending in early 2026,3 the negative wealth effects created by the drop in equity prices lead to a pullback in consumer spending. The unemployment rate rises to 7.2% amid the slowdown in economic activity, further restraining consumer spending. Real GDP grows by just 0.2% in 2027 before contracting by 1.2% in 2028.
As the recession begins to take hold, we assume the US Federal Reserve embarks on a series of interest rate cuts. This helps stabilize the economy and equity markets. However, the subsequent recovery is assumed to be much slower than after previous downturns. This is partially due to relatively small fiscal stimulus amid concerns over rapidly growing government debt and reinvigorated inflation resulting from large stimulus programs.
Unlike previous innovation cycles, the investments made in the United States do not retain their usefulness in the future. US businesses and consumers opt for cheaper, investment-light models, leaving frontier model developers with a much smaller customer base. Furthermore, foreign firms are expected to capture a larger market share outside of the United States. This reduces the economic value of some of the capital invested in more expensive frontier models and limits business investment from rebounding more quickly once the economy stabilizes. Real GDP growth during the recovery from 2029 through 2031 averages just 1.5% per year.
Upside: Inflation retrenches while AI investment runs ahead
We assume that the average US tariff rate falls to about 5% by early next year, as additional exemptions are granted and courts limit the scope of certain tariffs. Geopolitical tensions also moderate, reducing pressure on commodity prices. Oil prices are expected to stay below the baseline, with Brent crude averaging US$86 per barrel in 2026 and US$67 per barrel in 2027. Stronger net migration results in the adult population being roughly 868,000 higher than in the baseline by 2031. Business investment, supported largely by AI-related activity, remains stronger than in the baseline throughout the forecast period, while AI-related productivity gains are assumed to begin providing a larger boost to growth starting in 2028.
Consumer spending remains relatively strong, supported by faster population growth and continued gains in equity markets. Stronger investment and economic activity are assumed to generate sufficient labor demand to absorb the additional workers entering the labor force through higher migration. As the labor market tightens and productivity growth strengthens, faster real wage growth is expected to provide an additional boost to household spending. The unemployment rate remains below the baseline throughout the forecast period and eventually stabilizes at around 4.1%.
Although stronger in aggregate, business investment is likely to remain uneven across sectors. Elevated long-term interest rates continue to restrain investment in industries with less exposure to AI. However, lower tariffs are expected to reduce the cost of imported inputs and capital goods, particularly for capital-intensive businesses, helping to offset some of the drag from elevated financing costs.
Lower tariffs and oil prices are expected to help contain inflation despite stronger economic activity. We expect core inflation to remain marginally below the baseline through the third quarter of 2027. Thereafter, stronger demand from business investment and consumer spending is expected to place modest upward pressure on core inflation through 2029, although stronger productivity growth limits the extent of those pressures.
Lower near-term inflation reduces the likelihood that the Fed will need to raise interest rates again this year. At the same time, strong demand and a relatively tight labor market keep rates unchanged until the start of 2028.
Sectors
Labor market
Labor market data is sending mixed signals. Since January 2026, the establishment survey by the US Bureau of Labor Statistics shows nonfarm payroll employment grew by 643,000. The household survey’s employment measure, meanwhile, dropped by 1.2 million.
Given the establishment survey has a larger sample size and is more consistent with some private sector data, we view its employment figures are the better signal. This pace of employment growth is still relatively modest, with monthly payroll gains averaging just 80,375 over the last eight months, bringing total employment just 0.4% above its level a year earlier.
This modest pace of employment growth is still likely stronger than the growth of the US working-age population, which the Census Bureau projects to be unchanged in 2026 and 2027. Such a weak demographic outlook is one reason why we anticipate employment growth to moderate, while remaining positive through the end of 2027. The working-age population, along with employment, is expected to modestly accelerate from 2028 through 2031. We anticipate employment growth in 2031 will be 0.1 percentage point higher than our June forecast.
A slowdown in employment growth along with more modest declines in the workforce participation rate will allow the unemployment rate to drift upward from an expected 4.2% average in 2026 to 4.3% in 2028.
From there, the unemployment rate is expected to fall back toward 4.1% by 2031. The slight uptick in the unemployment rate will allow wage growth to moderate slightly before accelerating in 2029. We assume that AI-related advances contribute to an acceleration in economywide productivity growth by this point, allowing for stronger wage growth. It is important to note that we have, so far, seen little evidence of this occurring and acknowledge that both the timing and magnitude of these gains from AI remain highly uncertain.
Consumer spending
Consumer spending continues to show upward movement. Positive wealth effects and tax cuts have likely bolstered consumer spending this year. In nominal terms, consumer spending grew 5.9% from a year earlier in July. At the same time, disposable income was up a more modest 4.5%. Aggregate wage growth was up just 3.5% over the same period. This dynamic of consumer spending rising more quickly than income has persisted since June 2024, although the gap has widened this year. This, in turn, has pushed saving rates to very low levels.
The US personal saving rate has been at 3% or lower since April 2026. Prior to this, a rate of 3% or lower occurred in only 41 of the 807 months for which we have available data. Almost all of those months of low saving rates occurred between 2005 and 2008, when the housing bubble was generating positive wealth effects that ultimately turned negative when the bubble burst.
That is not to say that the United States is necessarily experiencing an asset bubble today, but such low saving rates are unusual. We expect consumer spending to become better aligned with income growth in the quarters ahead, allowing the saving rate to move higher. The convergence between spending and income will likely need to come from the spending side. After all, hourly wage growth is decelerating, and demographics make it difficult for employment to accelerate.
We now expect real consumer spending to grow by 2% this year before decelerating to 1.4% next year. However, we also expect wage and employment growth to pick up thereafter, allowing consumer spending to accelerate as well. By 2031, real consumer spending is expected to grow by 2.1%.
Prices
Similar to the labor market data, inflation data has also been sending mixed messages. While the core consumer price index (CPI), which excludes food and energy prices, was up 2.4% in August, the core personal consumption expenditures (PCE) price index increased 3.4% in the same period.
This is notable for two reasons: One is that the CPI is historically the measure with higher inflation readings, not the PCE price index. In addition, the two indices are moving in opposite directions. Core PCE prices are accelerating while the core CPI is decelerating. Only some of the gap between these two measures is expected to diminish after a methodological change to the PCE price index is published on September 30.4
On a sequential basis, we expect core inflation to moderate until 2029, when it is projected to return to the US Fed’s 2% target.5 The persistence of above-trend inflation over this period will likely come from two sources: The first is the ongoing AI buildout, which has caused the price of computer software and accessories to rise 21.2% from a year earlier in July. This is a stark change for a price index that was typically in the negative prior to the pandemic.
Some of the strength in these tech-related price indices is likely due to methodological issues, but some upward inflationary pressure is still expected to remain even after correcting for them.
The second driver of inflation will likely come from housing. After decelerating through late 2025, the cost of housing as measured in the consumer price index has stabilized. Private sector measures show that rental costs are beginning to accelerate again, which will likely add to inflationary pressure, albeit with a lag.
Tariffs appear to have been one factor contributing to inflationary pressure. However, one study shows that most of the tariffs that were implemented last year have already been passed on to consumers.6 Although tariff policy continues to evolve, we expect that the average tariff rate will remain relatively stable and that additional price pressures from tariffs will be minimal.
We assume crude oil prices will move downward over the next two years, allowing headline inflation to come down as well. However, ongoing conflict in the Middle East makes this assumption highly uncertain. Another risk to the forecast comes from elevated inflation expectations,7 which could cause actual inflation to become more persistent than we anticipate in the baseline.
Business investment
One of the most substantive changes made to the forecast has been the increase in the pace of business investment. We had previously anticipated that real business investment would grow by 6% this year and 4.9% next year. We now expect it to grow by 6.6% this year and another 5.1% next year. The stronger forecast partly reflects upward revisions to historical data and a stronger-than-expected second quarter.
Another factor behind the revised outlook is recent announcements from hyperscalers around their capital expenditure intentions. Although those intentions indicate growth of about 40% next year to more than US$1.1 trillion, this is still a slower pace than the roughly 75% growth expected this year.
These sizeable investment plans appear to be contributing meaningfully to overall business investment trends in the United States. In the second quarter, real business investment in information-processing equipment was up 22.2% from a year earlier. Software investment was up 10.8% over the same period.
While we expect strong growth in AI-related investments, other investment types are expected to remain relatively weak. Real investment in structures was 4.5% lower than a year earlier in the second quarter, while investment in transportation equipment was down 6.1%. Higher borrowing costs and rapidly rising input costs may continue to weigh on the economic case for stronger capital expenditures outside of technology-related sectors. Elevated uncertainty will also likely hold down investment growth. We expect investment in both structures and transportation equipment to contract modestly into next year.
Monetary policy and interest rates
After delivering three quarter-point interest rate cuts in the second half of last year, the Federal Reserve reversed course and raised interest rates at its September meeting. We now expect the Fed to hike once more before the end of this year. As consumer spending growth and inflation moderate in 2027, we expect the Fed will begin to cut rates again at the end of that year.
We have significantly raised our expectations for the 10-year Treasury yield. Longer-term bond yields have moved notably higher. Only part of this has been due to expectations of another rate hike from the Fed. Other factors, including higher expected debt issuance related to the AI buildout and relatively loose fiscal policy stances in several large economies, may also be raising the supply of debt and lifting yields in the process. We now expect the 10-year Treasury yield to remain above 4.5% until the end of 2027. As the Fed restarts its rate cuts, we expect the yield to converge toward 4.2%.
Higher long-term yields in the US Treasury market have led to higher mortgage rates. We now expect 30-year fixed-rate mortgages to remain above 6.5% through the end of 2027. Furthermore, we no longer expect the mortgage rate to fall below 6% before the end of 2031. As a result, residential investment is expected to remain relatively weak, contracting by 3.4% this year and growing by less than 1% in 2027. Housing construction will likely be further constrained by rapidly rising prices of construction materials and construction wages.
By
Michael Wolf
Deloitte United States
Rohini Sanyal
Deloitte India
Editorial (including production and copyediting): Arpan Saha, Preetha Devan, and Anu Augustine
Design: Harry Wedel
Cover image by: Rahul Bodiga
Knowledge services: Rohan Singh
Visit the Deloitte Global Economics Research Center
Access more insights for the consumer spending, housing, business investment, globalization & international trade, fiscal & monetary policy, sustainability, equity, & climate, labor markets and prices & inflation sectors.