AI infrastructure investment raises inflationary pressures Analysts: The Federal Reserve faces new challenges

With the reopening of the Hormuz Strait, oil prices are falling sharply. However, at a time when energy prices are dropping, the Federal Reserve may face a new inflation challenge: artificial intelligence (AI). Tech giants are competing to build AI infrastructure, purchasing and manufacturing millions of specialized AI computing chips, and constructing data centers equipped with liquid cooling systems. TD Cowen predicts that major hyperscale cloud service providers' capital expenditures will reach $745 billion this year, and will continue to invest over $1 trillion in 2027 and 2028. According to its calculations, the spending of these giants will rise to about 3% of GDP next year, significantly higher than the level of less than 0.5% in 2020. The impact is beginning to show in inflation data and in the rising demand for construction workers. Demand for components used to manufacture chips and build infrastructure is extremely strong, affecting prices and spreading to other industries. Memory and storage chips, in high demand in the AI field, are also used in consumer electronics products such as video games, cars, and smartphones. Apple announced this week that it will raise prices for iPads and MacBooks due to the rising costs of memory and storage. Its impact may continue to widen. Oscar Munoz, chief economist at TD Securities, said, "The initial stage of AI infrastructure development will be inflationary because demand will create pressure in a situation where the economic supply is fixed; but in the coming years, as productivity improves and the economic pie grows, it may ultimately become a force inhibiting inflation." Munoz said this means that the neutral rate is now higher, and the Federal Reserve will have to determine whether to raise rates as a result. The neutral rate is the benchmark interest rate level that neither stimulates nor suppresses economic growth. He said, "This presents a challenge for the Federal Reserve. The massive capital demand for AI, even if only sustained during the construction phase, is pushing up short-term neutral rates, regardless of whether it ultimately proves to have anti-inflation effects in the medium to long term." When will AI boost productivity? These developments contradict statements made by Federal Reserve Chairman Kevin Walsh and others earlier. They have said that AI will boost productivity, lower inflation, and give the Federal Reserve room to cut interest rates. Walsh did not confirm whether he still holds this view during his first press conference on June 17, merely stating that the Federal Reserve still has work to do on price stability, which is the current focus. But Greg Daco, chief economist at EY, believes there is a misunderstanding about technological revolutions, that technological revolutions will boost productivity and alleviate price pressures. He said this result will indeed happen, but before that, prices will rise. Daco pointed out that the first phase of any technological revolution will bring a rise in inflation because this phase is capital intensive. The initial wave of investment puts pressure on resources, with a large demand chasing limited supply. "As long as we're still in the first stage, the investment stage, we will see these pressures. In the next one to two years, these inflation pressures will gradually pass on to consumers, and then gradually fade as investment growth peaks," Daco said. Only then will we enter the second stage, where stronger productivity starts to appear. Daco pointed out that once the infrastructure and talent are in place, it will drive productivity growth. He said, "It takes time for companies to adopt this technology, integrate it into existing processes, create new processes, and have the right talent. It's only when all these elements are in place that productivity will show. It's a process that takes several years." Munoz also pointed out that Wash's view and the current rise in prices can coexist. He said that the final second stage may see a decrease in inflation, as the application of AI throughout the economy may increase output and production capacity. "As economic productivity improves, the cost savings can be passed on to consumers in the form of lower prices. Therefore, both narratives may be true, just at different times," Munoz said. This article is reprinted from "Financial Union". GMTEight Editor: Chen Siyu.