The Federal Reserve has an inflation problem again, but it is not quite the same one it faced in 2022. US consumer prices rose 0.6% in September, pushing annual inflation to 3.7%, while core inflation increased 0.3% over the month. Gasoline prices alone jumped 9.1%, accounting for more than half of the monthly increase. The producer side looks even more uncomfortable. PPI rose 1.3% in September and 6.3% year over year, with energy prices up 10.8% and transportation and warehousing costs rising sharply. The obvious conclusion would be that the Fed needs to raise rates again. The reality is more complicated. Kevin Warsh declined to give markets any clear signal about the 27–28 October meeting during his appearance in Bangkok, beyond repeating the Fed’s commitment to price stability. Investors interpreted the absence of hawkish guidance as meaningful, cutting the implied probability of an October increase to below 15%. That does not mean the inflation problem has disappeared. It means the Fed is becoming more interested in understanding what kind of inflation it is fighting.
The latest acceleration began with energy. The war in the Middle East pushed oil, petrol and diesel prices sharply higher. Those increases are now spreading through transport, logistics and eventually the prices of goods delivered to American consumers. Producer prices show the mechanism clearly. Energy becomes more expensive, transport costs increase, manufacturers and wholesalers face higher input costs, and eventually some of those costs are passed on. The concern is therefore no longer simply that petrol became expensive. The question is whether expensive petrol makes everything expensive. Early signs suggest some transmission is occurring. Core CPI has surprised on the upside for two consecutive months, while prices for several goods and services have strengthened. September also produced the largest monthly increase in overall consumer prices since April. But the evidence is not yet sufficient to conclude that America has entered a new broad-based inflation cycle.
Warsh and several colleagues appear willing to distinguish between a temporary price-level shock and persistent underlying inflation. Monetary policy works against excessive demand; it is far less effective against a geopolitical disruption to oil supply. The Fed can make mortgages more expensive. It cannot reopen Hormuz. Raising rates aggressively against an energy shock therefore carries an obvious risk: inflation may eventually fall as oil supply normalises, while the central bank has unnecessarily weakened the rest of the economy. Yet doing nothing carries another risk. If businesses assume energy costs will remain high, they raise prices. If employees respond to falling purchasing power by demanding larger wage increases, temporary inflation can become persistent inflation. At that point, the distinction between an external energy shock and domestic inflation disappears. This is the second-round effect the Fed needs to prevent.
For the moment, the US economy is making that judgement particularly difficult. Demand remains resilient. Retail sales continue to grow, manufacturing indicators have strengthened, and major US banks report record net interest income, while credit deterioration remains remarkably limited. Higher rates are hurting certain sectors, particularly housing and highly leveraged borrowers, but they have not produced a general collapse in demand. That matters because the Fed is dealing with an economy that can absorb much higher borrowing costs than many expected. Artificial intelligence adds another complication. Warsh describes AI as potentially the most consequential economic innovation of his lifetime and ultimately sees it as a positive supply shock: higher productivity should allow the economy to produce more without generating equivalent inflation. Before AI increases supply, it is increasing demand. Data centres require enormous amounts of construction, electricity, computing equipment, semiconductors and capital. That investment boom is happening while the economy is already near full employment and energy markets are constrained. AI may eventually be deflationary. For the moment, parts of it are inflationary.
The Fed therefore faces three inflation forces at once. Energy is creating an external supply shock. Tariffs continue to affect imported goods prices. AI investment is creating an internal demand shock that may eventually become a productivity shock. Using one interest rate to respond to all three is necessarily imprecise. This explains why October increasingly looks like a pause rather than a change of direction. The September rate increase demonstrated that the Fed was willing to respond to renewed inflation pressure. Waiting in October would let policymakers assess whether the energy shock is spreading into underlying prices without prematurely committing to another sequence of increases. A pause should therefore not be confused with the end of tightening. September CPI showed annual inflation at 3.7%, above the Fed’s target, while producer inflation suggests that further cost pressure remains in the pipeline. Another increase later this year remains entirely plausible if those pressures persist.
But the more interesting question is what happens afterwards. If energy prices stabilise, core inflation moderates, and the bond market keeps long-term financing conditions tight, the Fed may find that one more increase is enough. New York Fed President John Williams has already indicated he can envisage another hike this year without assuming another tightening cycle in 2027. If energy inflation instead migrates into wages, services, and inflation expectations, while AI investment continues to stimulate demand, the market may once again underestimate how high rates need to remain. That is why Warsh’s silence matters. The Fed no longer needs to convince markets that it is prepared to fight inflation. September demonstrated that. It now needs to avoid fighting the wrong inflation by tightening too much. America’s inflation problem has returned, but its causes are fragmented between geopolitics, trade policy and an extraordinary investment boom. The next Fed decision will therefore be less about whether inflation is too high, and more about something much harder to determine: how much of that inflation interest rates can actually cure.