Inflation, AI, and the new demand cycle

AI is no longer just a productivity story. It is becoming a demand story – with implications for inflation. In this insight, we explore what the data shows about this shift.

• Inflation remains above target, with both headline and core measures rising
• AI investment is emerging as a new demand driver, particularly in semiconductors, energy and infrastructure
• Capex at scale is supporting pricing pressures in the near term
• Long-term impacts remain uncertain, with potential productivity gains offsetting inflation over time

Recent US inflation data indicate that price pressures remain above the Federal Reserve's long-term target. The Consumer Price Index (CPI) increased 3.8% year-on-year in April, compared with 3.3% in March, while core inflation, which excludes food and energy, rose to 2.8% from 2.6%.

Energy prices contributed significantly to the monthly increase, while shelter costs continued to exert upward pressure on overall inflation.

Other indicators also point to ongoing pricing pressures across parts of the economy. Producer prices increased at an annual rate of 6%, their highest level since late 2022, while import and export prices also moved higher. Market participants have cited a range of factors, including developments in energy markets and supply chain concerns, as contributors to these price pressures.

Financial markets responded by reassessing inflation expectations and the potential path of monetary policy. Treasury yields moved higher, while long-dated Treasury Inflation-Protected Securities (TIPS) yields reached their highest levels in more than a year. Attention has also focused on developments at the Federal Reserve following the appointment of Kevin Warsh as Chair.

Warsh has previously expressed support for maintaining price stability through monetary policy when inflation risks are elevated. More recently, he has highlighted the potential benefits of technological innovation, productivity improvements and regulatory efficiency, particularly those associated with AI. While continuing to emphasise the importance of an independent and data-driven approach to monetary policy, he has suggested that supply-side improvements could help moderate inflation over time. He has also noted that temporary inflationary pressures arising from energy prices, tariffs or geopolitical developments may not always require a strong policy response.

This discussion comes at a time when AI is becoming an increasingly significant area of investment within the US economy. The rapid expansion of AI infrastructure is reshaping capital allocation across multiple sectors and influencing economic activity in a variety of ways.

Demand for semiconductors, memory products and other electronic components has increased as technology companies expand computing capacity. Prices for computer software and accessories, which had generally trended lower over many years, have risen by nearly 14% over the past 12 months. Wholesale prices for electronic components increased by 28% over the same period.

The impact is also visible in international trade data. Because a substantial proportion of AI-related hardware is manufactured overseas, US imports of technology products have increased significantly. Imports of computers more than doubled during the first quarter of 2026 compared with the same period a year earlier. Semiconductor imports rose by 40%, while imports of computer accessories increased by 37%. These developments indicate that AI-related investment is becoming a more significant source of demand within the economy, although its broader impact on inflation remains subject to ongoing assessment.

Federal Reserve officials have acknowledged that strong AI-related investment could contribute to higher input costs across a range of industries. One area receiving particular attention is electricity demand. 

After more than a decade of relatively modest growth, US electricity production increased by 2.5% in 2024 and 2.4% in 2025, with further growth recorded in early 2026. Data centres supporting AI model training and deployment are widely viewed as a contributing factor. 

Consumer electricity prices rose 4.6% year-on-year in March, potentially reflecting a combination of factors, including increased demand.

Investment spending by major technology companies remains elevated. In 2025, leading AI-focused firms, including Google, OpenAI, Anthropic, Meta, Amazon and Oracle, collectively committed an estimated $300 billion to capital expenditure across semiconductor supply chains, energy infrastructure and specialised labour. That momentum has continued into 2026, with the five largest US hyperscalers – Microsoft, Alphabet, Meta, Oracle and Amazon – projecting combined capital expenditure of approximately $720 billion this year alone.

Much of this spending is being supported by growing revenues from AI-related products and services. Microsoft recently reported that its annualised AI revenue run rate exceeded $37 billion, representing growth of more than 120% year-on-year. Alphabet has highlighted increased adoption of AI-enabled services, while Amazon Web Services recorded its fastest expansion in more than three years, supported in part by growing AI workloads and strategic partnerships.

Corporate earnings have remained relatively resilient despite higher interest rates and tighter financial conditions. Nvidia continues to report strong customer demand and healthy balance sheet performance, while Anthropic has reported significant revenue growth and is expected to achieve its first quarter of operating profitability. These developments have coincided with stronger performance among several large technology companies, even as bond markets continue to reflect concerns about inflation and elevated yields.

Questions remain regarding whether current investment levels will generate sufficient long-term returns. Some analysts estimate that hundreds of billions of dollars in recurring annual revenue may ultimately be required to justify the scale of capital expenditure currently underway across the AI ecosystem. 

Rising hardware costs also present challenges, with companies reporting increased spending on chips, memory products and specialised components amid sustained demand.

Over the longer term, many economists continue to view AI as a potentially disinflationary force. By automating administrative processes, improving operational efficiency and optimising supply chains, AI could support productivity growth across the economy. Historically, higher productivity has enabled stronger economic growth and wage gains without necessarily generating sustained inflationary pressures.

The timing and scale of these benefits, however, remain uncertain. While the effects of infrastructure investment are already visible in some sectors, productivity gains associated with AI may take longer to emerge. As a result, the relationship between AI investment and inflation is unlikely to be straightforward. Large-scale investment in data centres, semiconductors and energy infrastructure may continue to support demand and influence pricing dynamics in the near term, even if broader productivity benefits materialise over a longer horizon.