Global Outlook 2026 - US: The AI Haves and Have Nots

The economy is running at two speeds. AI and related sectors are thriving, whereas most other areas are stagnant or contracting. This creates a challenging environment for policy-making and economic forecasting. The Trump administration is fully committed to AI, implementing an Action Plan that reduces regulatory oversight and encourages investment across the supply chain. Meanwhile, the Federal Reserve has resumed easing its policy rate to support a weakening labour market. However, it remains unclear how this shift in monetary policy will affect a bifurcated economy, where supply-side factors are significant. The outlook is further muddied by uncertainty over the timing and magnitude of trade and immigration policy impacts; both are expected to increase prices and suppress growth, but have yet to fully be reflected in headline data. We expect 2026 to be characterised by a further widening of this ‘K-shaped’ economy, amplified by trade and immigration policy. Downside risk to growth is limited by fiscal and monetary easing. This keeps the economy growing at a solid pace of slightly above 2% but comes at the cost of higher inflation and larger fiscal deficits. Policy is unlikely to narrow the divide.

The two faces of the US

Investment has surged on the prospect of AI market dominance. Our piece on AI describes how its introduction and development have already influenced the US economy and speculates on future impact. AI has already supported headline economic growth, both directly through investment and indirectly through consumption, via wealth effects from soaring equity prices. This wealth effect is evident when comparing the University of Michigan’s sentiment index for non-stock owners and the top 20% of equity owners. While both have deteriorated since last year, the sentiment gap between the groups is widening again. Sentiment among stock owners correlates strongly with returns on the ‘Magnificent Seven’, and these gains have supported consumption. The top 10% of income households now account for 49.2% of all consumption, compared to 46% in 2022 and a low of 41% after the financial crisis. If this share fell back to 46%, with the remaining 90% of consumption fixed, total consumption would drop by 6%.

Meanwhile, sentiment among non-stock owners has fallen to its lowest level in at least 25 years. Without support from equity gains, these households rely on real wage growth for consumption growth. According to Atlanta Fed data, wage growth for the lowest quartile, previously strong due to labour market tightness since 2021, has now dipped below that of the top quartile for the first time since 2014. Overall, wage growth has returned to pre-Covid levels in nominal terms, but is substantially lower in real terms. While overall unemployment has not risen much, unemployment among some groups which serve as an early indicator has increased in recent months. Lower income households are also expected to face a significant hit to disposable income due to rising healthcare costs under the OBBBA.

The real economy, and investment in particular, mirrors this divide. Investment in data centres has surged by over 50% per year, while other forms of fixed and residential investment are contracting. Estimates of capital expenditure by the four largest hyperscalers (builders of infrastructure) suggest their influence on headline growth figures will continue in the near term, but gradually fade. It is not clear that the expected easing by the Fed will be enough to revive the rest of the economy, especially household consumption. Fiscal policy may provide a moderate boost to investment through bonus depreciation extensions, but if anything, redistributional effects only makes things harder for most households.

Implications for 2026 and beyond

First and foremost, we expect to see the impact of tariffs and immigration policy in headline data in 2026. Goods prices have risen, but full pass-through of tariffs has yet to occur. As import prices have not increased and consumer price rises have been limited, it is mostly US firms paying the tariffs. Businesses need time to react to such shocks. Initially, uncertainty over the permanence of tariffs led firms to build inventories, limiting immediate price rises. Now that tariffs seem permanent, companies face decisions about sourcing and production, which takes time, but costs will eventually be passed on to consumers.

The full impact of immigration policy has also yet to be seen. Labour supply growth has stalled, explaining much of the current labour market dynamics. However, the effects of the ICE crackdown on illegal labour are not yet evident. Homeland Security reports over two million people have left, but economic data shows little evidence of this. Production has not declined, nor has there been a substantial increase in native or legal employment in sectors that rely heavily on illegal labour. There is no clear evidence for replacement by capital inputs or AI. It is likely that deportation figures are overstated and most illegal workers remain. Similar to the businesses above, these families also need time to react, and more substantial departures are likely in 2026. This will lead to tightness in relevant sectors, with wage pressures balancing between the weaker negotiating position of remaining illegal employees and the stronger position of legal employees. We expect this to result in some upward price pressure, particularly in food and other sectors.

Both trade and immigration policy will therefore raise prices and suppress growth in the coming year. Still, the economy should hold up due to substantial private sector investment, predominantly in AI, and accommodative fiscal and monetary policy. The Fed’s rate cuts this year and into 2026 will help somewhat, though the impact will be limited, with markets already pricing in likely cuts, broadly higher risk premia in long-term rates, and the lags of monetary policy. It is also unclear how much demand stimulus can support the labour market, given that much of the weakness is supply-driven. On the fiscal side, we expect unintended but substantial stimulus relative to 2025. The administration’s cut to the IRS budget has a large multiplier for tax revenues. We also expect tariff revenues to decline as firms adjust. The total reduction in tax and tariff revenues could exceed half a percentage point of GDP. Both fiscal and monetary easing will benefit the broader economy, though the advantages are likely to be concentrated among the ‘AI Haves’. Lower tariffs will reduce inflationary pressure, but the IRS budget cut will likely favour higher incomes. Given the relative insensitivity of AI investments to policy rates, monetary easing may provide more stimulus to the rest of the economy. Still, lower rates will amplify the wealth effects described above and increase inflationary pressures in an economy where much inflation is already demand-driven.

This suggests the US will continue to grow at a solid pace, but at the cost of higher deficits and inflation. We have upgraded our growth forecasts and increased the expected peak and persistence of inflation. Despite higher inflation, we expect the Fed to continue easing into 2026. This is consistent with historical Fed behaviour in similar situations, where sustaining ‘inclusive full employment’ is seen as the best solution to uneven growth, although inflation was typically at or below target in those easing periods. The dovish FOMC rotation next year may use the windfall from lower oil prices on headline inflation to argue that inflation remains contained and that employment is the bigger worry. Continued fiscal and monetary stimulus will, therefore, keep inflation running above target for the remainder of our projection horizon. If policy easing ignites labour demand without supply, risks of second-round inflation effects beyond our base case increase. If labour demand does not pick up, the weaker part of the economy risks entering stall speed and contracting. The economic resilience of the AI Haves is fragile due to its sheer concentration. Both segments of the economy tread a narrow path, and political pressure on the Fed makes it substantially harder to navigate.