US consumer price data for July showed an overall slowdown in inflation, but detailed breakdowns reveal a notable trend: rising AI-related investment is pushing up prices for technology goods. Headline CPI rose 0.1 percent month-on-month and annual inflation eased from 3.5 percent to 3.4 percent; core inflation (excluding food and energy) was 0.2 percent month-on-month and 2.5 percent year-on-year. Despite these favourable headline figures, prices in information technology categories have jumped.
Key figures
- Information technology commodities rose 1.4 percent in July month-on-month. Bank of America (BofA) estimates this contributed roughly 12 basis points to core goods inflation.
- Prices for computers, peripherals and smart home devices increased 3.5 percent month-on-month in July.
- Smartphone prices rose 1.1 percent month-on-month.
- Computer software and accessories were 21.2 percent higher year-on-year.
These developments stand out because many electronic products have tended to decline in price over recent decades.
What is driving the price increases?
The primary drivers are growing business demand for AI and large-scale data-centre construction. AI infrastructure requires not only accelerators (for example, Nvidia GPUs) but also large volumes of DRAM, NAND flash memory, processors, networking equipment and storage. Chipmakers are redirecting capacity toward higher-margin server markets, leaving fewer components for PCs and smartphones.
TrendForce projects traditional DRAM contract prices could rise another 13–18 percent quarter-on-quarter in Q3, while NAND flash prices could increase 10–15 percent. The research firm says rising component costs will force notebook and smartphone makers to raise consumer prices.
The macro scale of AI investment
AI-related capital spending has become material at the macro level. Goldman Sachs estimates that the largest hyperscalers’ capex in 2026 could approach $750 billion. Morgan Stanley’s CEO stated in July that global data-centre investment expected for 2026 has risen from a prior estimate of $575 billion to about $850 billion. Such demand shocks affect not only semiconductors but also power grids, construction and financing markets.
Demand-side effects and statistical caveats
BofA notes the inflationary channel is not only via production costs. Rising tech equity prices have created wealth gains for some US households, which can support consumption and raise demand. However, IT goods still have a small weight in the CPI basket: in July the "information technology commodities" category accounted for only 0.736 percent of the CPI, and computer software and accessories just 0.031 percent. Thus, even double-digit price increases in these categories have a limited direct impact on the headline CPI.
The Federal Reserve has also highlighted measurement issues. The appearance of AI features, continuous product updates and subscription business models complicate the proper handling of quality changes in inflation statistics. The Fed’s analysis suggests that with certain methodological adjustments more than half of the reported software price rise could disappear, implying some of the observed increases reflect better or different products rather than pure price inflation.
What policymakers say
Kevin Warsh, the Fed nominee, acknowledged in a July congressional hearing that the AI investment boom could push measured prices higher over the next 12 months. Fed Governor Lisa Cook also cited AI investments among upside inflation risks. At the same time, both—and other officials—stress that AI could be disinflationary over the longer term if it generates substantial productivity gains.
Short-run versus long-run dynamics
Analysts describe a two-phase economic effect of AI. The first phase requires building infrastructure, producing chips, expanding power capacity, hiring engineers and mobilising large amounts of capital. If these investments grow faster than supply can adapt, they will raise certain prices. The second phase would see productivity gains from AI reduce unit costs. Current data suggest the US economy is still largely in the first, investment-heavy phase, increasing near-term inflationary risk.
If higher tech prices spill over through additional channels—data-centre energy demand, construction capacity constraints, higher wages, and greater financing needs—AI could become a measurable macro-level inflation factor. If supply adapts quickly and productivity gains materialise, the current price pressures may be temporary.
Conclusion
July’s figures indicate that the AI revolution is making parts of the economy more expensive in the short term rather than cheaper. The technology’s long-term disinflationary promise remains significant, but the cost of building the necessary infrastructure must be paid first—and a growing share of that cost is already being felt by consumers.
Tags: inflation, Federal Reserve, artificial intelligence, technology prices, DRAM, NAND, data centres, Bank of America, Goldman Sachs, Morgan Stanley, Kevin Warsh, Lisa Cook
