Macro Moments Weekly: Moving the goalposts: AI, inflation, and the next rate cycle

Key takeaways:

1. AI-driven productivity gains are the best inflation hedge. Four consecutive quarters of US productivity growth suggest the current economic expansion can sustain itself without triggering broad inflation—as long as supply shocks remain contained to specific sectors.
2. Tech inflation is real but slow-burning. Soaring prices for DRAMs, NAND, and GPUs will eventually reach consumers, but because these are deferrable purchases, the inflationary impact will be spread over years rather than hitting all at once.
3. The bigger long-term risk is deflation, not inflation. Chinese export deflation continues unabated, oil prices are easing with the Hormuz truce, and demand-led inflation remains a distant concern—pointing to lower, not higher, interest rates by 2027.

Oh, to be a central banker, now the Iran war is over! The European Central Bank (ECB) pulled the trigger with a rate hike on 11 June, and while the US Federal Reserve (Fed) didn’t at its meeting on 17 June, the policy statement managed to exhibit every possible view, except that of the new Fed Chair Walsh, who was a bit vague, with no forward guidance. The Bank of England’s Monetary Policy Committee has sat on its hands, which is reasonable given the weakness in private sector wages and employment at present. The question that should be addressed is are we now looking at a US-led economic recovery? What impact does this recovery, driven by an industrial revolution, have on inflation?

The answer ought to be very little. The productivity gains from artificial intelligence (AI) are already evident, with four successive quarters of productivity gains in the US economy (per US labour productivity output per hour, non-farm business sector). The United States achieved greater productivity gains in the third quarter 2025 than the United Kingdom has since the start of 2022.¹ Rising productivity is the holy grail of inflation free growth, as long as there are no supply shocks.

Except that there are supply shocks. We can see the inflation in technology, and it is serious; prices for dynamic access memory (DRAMs) rose 80%-90% in the first quarter, and the Gartner group expects a rise of 125% this year.²

NAND prices are expected to go much further, with Gartner looking for 234% increase in 2026.³ Graphics processing units (GPUs) up a100% and cloud prices to follow. But whilst these will make life immediately more expensive for business especially those using more AI tokens, may force up the price of a new car or a robot vacuum, they don’t have the impact on the consumer that the pump price or the food shop does. 

So, the inflation in tech, which the consumer will have to pay for eventually, will not likely impact the wider economy just yet. But as you are looking at a new mobile phone, vacuum, or car, it’s going to cost you—eventually. The one advantage the consumer has is that they can, more often than not delay these purchases. So effectively the inflation will be spread across a number of years, whilst the productivity gains are not. They seem to be arriving fast and earlier, hence underpinning profits and in particular markets.

Which leaves us looking towards 2027. When does this growth produce broader demand led inflation? This is a concept we have not been familiar with much since 2009. There was a brief burst at the start of 2022, based on the return of demand coinciding with continued post covid disruption to supply lines. This was lost as it was mixed into the inflation cocktail created by the Russian invasion of Ukraine.

The key to prices since 2009 has been the deflation exerted by Chinese exports, and there seems to be nothing stopping that trend. Perhaps, behind the tariff wall, the US economy may feel this less than others. And as the world’s interest-rate benchmark, that is important, although it will feel the fall in oil prices that a truce in the Straits of Hormuz will bring.

It is hard to get bearish about long-term inflation outside of the United States at this point; in fact, the risk is to the downside. If that is the case, what is the possible direction of interest rates in 2027?

Parting shot

Social media is full of the fun that the Scottish football fans have had in Boston amid the World Cup—and that the bars have run dry. The Scots play their second fixture there as well. Four days later England is next up: Does this give a whole new meaning to the anthem: ‘The British are coming’? It is certainly no tea party!