How Tariffs, Oil, and AI Are Colliding Into the Most Complicated Inflation Picture in a Generation

Attempting to track American inflation at this time leads to a certain kind of fatigue. The story of pandemic supply shocks, stimulus money chasing too few goods, central banks raising rates, and prices eventually cooling used to be fairly readable. It hurts, but it makes sense. In 2026, things will be different. Prices are being pushed and pulled simultaneously by three distinct forces, each operating according to its own logic and schedule. These forces interact with one another in ways that make conventional economic forecasting seem almost archaic.

Since tariffs are where the immediate pressure is most noticeable, start there. Following a sharp increase to 6.5% in 2022, the highest level since 1981, US PCE inflation gradually decreased, reaching about 2.6% by 2024. The Federal Reserve was within striking distance of its 2% target after aggressively raising rates and keeping them there. The story of the “soft landing” was becoming more plausible. Subsequently, the disinflation stalled as the tariff architecture increased. Because US companies had accumulated pre-tariff inventory and could postpone the reckoning, they absorbed tariff costs into their margins rather than immediately passing them on to customers for the majority of 2025. There is currently a shortage of that inventory. According to Morningstar’s analysis, between 2025 and 2027, businesses will start passing on costs they have been covertly absorbing, resulting in cumulative price increases of about 4.5% for durable goods and 5.6% for non-durables. In other words, the worst of the inflation caused by tariffs is still to come.

Key Reference Data: Inflation Drivers 2025–2026

The uneven geography of the mechanism makes this more complex than a straightforward “tariffs cause higher prices” narrative. Tariffs act as a consumption tax inside the US, driving up the cost of imported goods, putting pressure on household spending, and now squeezing the corporate margins that have been shielding consumers from the full impact. However, something different is taking place outside of the US, especially in Europe and emerging markets. Products that are no longer able to be sold in the US at competitive prices are being redirected to other markets, which increases supply and lowers prices in those areas. In Stuttgart and Nairobi, the same policy that causes inflation in Chicago causes deflation. Any coordinated worldwide response is made more difficult by that truly unique dynamic.

The second force is oil, which came from geography and conflict rather than trade policy decisions. Energy price volatility has returned as a significant inflation driver at the wrong time due to the Iran War and its impact on shipping through the Strait of Hormuz. An extended period of high oil prices could “crimp” the AI boom, the WTO’s chief economist warned in March. This statement sounds almost ridiculously modern, but it is accurate. Consumers are not only affected by high energy costs at the pump. They affect supply chains, drive up manufacturing and transportation costs, drive up the cost of agricultural inputs like fertilizers, and drive up the running costs of the server farms and data centers that support the entire AI infrastructure buildout. One sign was the average fuel price in the US surpassing $4 per gallon for the first time in four years. Another, more significant one was Aramco’s warning of a possible “catastrophe” for the oil market should the Strait of Hormuz continue to be disrupted. The scenario of stagflation, which is characterized by high energy costs and slowing growth, is no longer theoretical. It is the main issue.

Then there’s AI, which is doing something unique in this picture: depending on your timeframe, it can act as both an inflationary and a disinflationary force simultaneously. The AI investment cycle is inflationary in the short run. The billions of dollars being spent on data centers, semiconductor fabrication capacity, server infrastructure, and power grid expansion are driving up the cost of the energy and components needed to build it all. Businesses are starting to pass those costs on, and the current spike in demand for AI-specific hardware partially reflects the chip supply bottlenecks that characterized 2021 and 2022. The Fed’s traditional inflation models are being complicated by AI investment, according to a University of Chicago economist, because it simultaneously increases investment spending (inflationary pressure) and promises productivity gains (disinflationary effect), and the two effects don’t happen at the same time.

In theory, the productivity payoff from AI would lower costs across industries, reduce labor inputs for many tasks, and provide the kind of efficiency gains that have historically helped keep prices in check. However, there is genuine uncertainty regarding the timeline. That long-term view is accurate. However, it won’t come this year. At the moment, the AI boom is primarily causing pressure on semiconductor supply, energy demand, and large capital flows that are warping investment patterns throughout the economy.

It’s difficult to ignore how different this environment feels from any of the inflationary periods that economists usually use as comparisons. Oil shocks and loose monetary policy contributed to the stagflation of the 1970s, but there was no corresponding increase in technological investment. The supply chain and demand were the main causes of the inflation in 2021–2022. All of those factors are present in the current situation at the same time, with a geopolitical conflict impacting the world’s most important shipping chokepoint and a trade war layered on top. Longer-term Treasury yields remain stubbornly high despite the Fed cutting rates by 1.75 percentage points since September 2024. In December, the 10-year was at 4.2%, indicating that bond markets are doubtful that inflation will decline as much as the central bank anticipates.

A background vulnerability that receives little attention is the personal savings rate, which stands at 4.8%, significantly lower than the pre-pandemic average of 7.3%. The spending pullback could significantly slow growth if asset prices decline and consumers attempt to replenish their savings buffers, but it would also theoretically help reduce demand-pull inflation. In fact, that is the best-case scenario. The pessimistic one is that consumers will continue to be hit by tariff-driven cost increases while oil prices remain high, energy demand remains high due to the AI buildout, and the Fed is caught between an economy that is clearly losing steam and inflation that is still above target.

Observing all of this at once gives the impression that the analytical frameworks designed to comprehend inflation in more typical times are being pushed beyond their practical bounds. By modifying the cost of borrowing, central banks were intended to control inflation. When excessive demand is the cause of the inflation, that works fairly well. When price increases stem from trade policy, geopolitical disruption, and an energy-intensive technology investment cycle—factors that higher interest rates cannot directly address and may even exacerbate by slowing the growth that could eventually cover all of this—it operates less smoothly.
There is no clear solution to this problem. That may be the most truthful thing to say about it.