Major technological innovations possess the capacity to reshape economies by generating opportunities, employment, and even entirely new industries, while significantly enhancing productivity and growth. However, the expectation of longer-term gains frequently comes at the cost of short- and medium-term discomfort. In the context of artificial intelligence, this has encompassed job losses, a deceleration in wage growth, an expansion of wealth inequality, and notably in recent months, an uptick in inflation. Recent data indicates that the substantial interest and investment in AI adoption, estimated at approximately $750 billion for this year, have contributed to an increase in various prices, thereby elevating overall inflation. “The higher inflation means that households must spend just over $375 more to purchase the same goods and services as they did this time last year due to AI’s inflationary impact,” Mark Zandi wrote. The positive aspect is that AI currently accounts for only a minor fraction (approximately 0.2 percentage points) of total inflation, with its effects confined to a select few categories. However, the less favourable news is that AI is contributing to an increase in inflation, thereby exacerbating existing issues related to affordability. Furthermore, these price pressures are anticipated to persist for the foreseeable future, and there is a significant possibility that they may expand. That potential dynamic has Federal Reserve officials, including the central bank’s new chairman, vigilant.
Here is an examination of the instances where AI has already manifested in inflation and potential future occurrences. AI data centers exhibit significant energy consumption, particularly in terms of electricity and water. The swift growth of these large facilities poses a risk of overburdening power grids and potentially escalating prices further. That is primarily due to demand exceeding supply. Data center facilities can be constructed or expanded at a rate that is two to three times faster than the development of new electricity generation systems required to support them, as highlighted by PJM Interconnection, the largest grid operator in the United States. When one considers the necessity of retiring coal plants alongside the rising demands for electrification, the impact of extreme weather events, and the challenges posed by ageing infrastructure, it becomes evident that these factors collectively exacerbate the disparity between supply and demand. “Data centers are demanding huge amounts of power, and that’s tending to crowd out the electricity available to distribute to residents; it’s also led to wholesale electricity prices being bid up, because data centers are willing to pay the price that providers ask them for, and that ends up also raising the prices for residential electricity costs,” Pooja Sriram told. The US Consumer Price Index data indicates that residential electricity prices increased at approximately double the rate in 2025 compared to the historical average observed in previous years, she noted. During the initial five months of this year, data from the Bureau of Labour Statistics indicates that electricity prices have been rising at a rate surpassing that of 2025. “I think that is one of the clearest imprints of AI data center demand driving up residential electricity costs,” she said.
Electricity prices experienced an unanticipated decline of 1% in June; however, they remain above the general inflation rate, reflecting a 4% increase compared to the same period last year, according to the latest CPI data. The data centers’ appetites, however, aren’t fully sated with power alone. The substantial expansions have resulted in a significant increase in demand for memory chips, which has greatly benefited manufacturers, according to Sriram. “The issue is not just the demand; the issue is the supply side for those memory chips has been very constrained,” she said. The trillions of dollars pursuing AI-related components have led storage and memory suppliers to focus their wafer-manufacturing capabilities on high-bandwidth and high-speed (and highly profitable) memory products demanded by data centers, she stated. “What that does has basically diverted (the production of) the memory chips that you need for consumer products toward very specific high-performance memory chips that the data centers need,” she said. The pricing pressures of these and other components have been most evident at the producer level. The Producer Price Index chart for the semiconductor and other electronic component manufacturing industry exhibits a hockey stick shape. As of June, wholesale prices in that category increased by 26% compared to the previous year, marking a significant change from June 2025 when prices were declining at an annual rate of 0.8%, according to PPI data. “Why this matters for the end consumer is, at the end of the day, our laptops, computers, iPhones, iPads all have some sort of memory chip embedded in that hardware, and those chips have become quite expensive,” she said.
Late last month, Apple increased the prices for several of its most popular products by approximately 20%. In a statement, the company highlighted that AI data centers generated a “extraordinary surge” in demand for memory and storage. Sony increased the price of its PlayStation console earlier this year, while Microsoft raised the price of its Xbox consoles by approximately 25% last month for comparable reasons. “The entire consumer electronics industry is grappling with the ongoing components crisis, yet the repercussions are especially severe for consoles,” Microsoft stated. Computers and related hardware have historically represented a significantly deflationary segment within the Consumer Price Index. This trend is largely attributable to technological advancements, which enable consumers to obtain greater value for their expenditures. (For example, a $1,500 computer in 2025 is expected to exhibit greater power than a $1,500 model from the previous year, leading the BLS to categorise this as a price decline.) For the first half of 2026, however, data from the Bureau of Labour Statistics indicates that computers and related products have undergone price inflation. “We’re in the early innings of these consumer price pressures and the pass-through from higher producer prices, higher import prices and greater demand, especially for AI-led investment,” said Gregory Daco. Incorporating AI functionalities into business applications invariably influences the pricing structure of software.
For instance, Microsoft increased the prices of personal Office 365 subscriptions by 43% in February, with a 30% hike for family plans, following a decade of stable pricing. The new feature: Copilot, Microsoft’s new AI tool. Data centers are also influencing the availability of other essential construction inputs, including copper and electrical wiring, along with the labour force. “If you’re looking for data that was conclusive (about AI’s effect on the economy and inflation), albeit a bit more subtle, you would look and see whether wages in construction were going up more than wages in the rest of the economy,” said Thierry Wizman. “Because if in fact there is upward pressure, straining resources of the economy because of the AI data center buildout, you would see it in wages as well – specifically in the wages of labor that would be working these projects,” he added. So far, the available national-level wage data indicates a “robust divergence” between the construction sector and the aggregate, he noted. Regional data could prove even more telling, he noted, emphasising the significance of monitoring construction wages in regions with a high concentration of data centers. That will require additional patience, as the more localised data is delayed due to the necessities of collection and modelling. “We’re having a problem with housing in the country these days; people talk about it as being unaffordable,” Wizman said. “It could be the case that the fact wages in construction have been rising a lot is putting upward pressure on houses as well.”