Artificial intelligence is widely expected to make businesses more productive, reduce operating costs and eventually make many goods and services cheaper. Yet the technology could create a less obvious economic problem during the transition: it may raise inflation before its productivity benefits become large enough to reduce prices. The warning from Swiss National Bank governing board member Petra Tschudin reflects a growing concern among central banks that the enormous investment required to build the artificial intelligence economy could generate demand and supply pressures at the same time.

Tschudin said artificial intelligence could create upward pressure on prices in the short and medium term because investment is being redirected toward chips, computing infrastructure and other technologies needed for artificial intelligence. Those shifts can create shortages in other parts of the economy and push up input costs. She also argued that productivity gains could eventually reduce prices, but that a one-time improvement in productivity would not automatically create sustained deflation.

The concern is becoming more relevant as artificial intelligence investment reaches unprecedented levels. Companies are spending heavily on data centres, advanced processors, electricity infrastructure and networks, while governments and businesses are competing for many of the same physical resources. The Bank for International Settlements has warned that the artificial intelligence boom is simultaneously stimulating demand and expanding potential supply, making its ultimate effect on inflation unusually difficult for central banks to assess.

This creates an important distinction between the long-term promise of artificial intelligence and its short-term economic consequences. A technology can eventually make production more efficient while initially making the economy more expensive to build. The inflation question therefore depends heavily on timing: how quickly investment demand rises, how quickly infrastructure and energy supply respond, and how rapidly productivity improvements spread through the wider economy.

AI Is Creating a Powerful Investment Shock

The first inflationary channel comes from the extraordinary amount of capital being directed toward artificial intelligence. Technology companies are building data centres at a scale that requires enormous quantities of advanced processors, servers, electrical equipment, construction materials and power. That spending creates additional demand for workers, industrial inputs and financing even before artificial intelligence produces significant productivity gains elsewhere in the economy.

The scale of the investment is already large enough to influence national economic activity. The five largest American companies investing in artificial intelligence data centres are projected to approach $1 trillion in combined capital spending by 2027, compared with about $200 billion in 2024. Such spending supports construction, manufacturing, energy production and technology supply chains, creating an economic stimulus that can increase demand for goods and services.

The inflationary effect becomes stronger when supply cannot expand as quickly as demand. Semiconductor manufacturing is an obvious example because advanced processors are essential to artificial intelligence systems and cannot be produced rapidly simply because prices rise. The same problem can occur with electricity, data-centre construction, specialised equipment and some raw materials.

Central banks are therefore paying attention to the possibility that artificial intelligence investment could crowd out other economic activity. The Bank of England has noted that artificial intelligence infrastructure requires substantial investment and that shortages involving electricity connections and labour could constrain the expansion of data-centre capacity. The technology may create new productive capacity in the future, but the construction of that capacity can generate immediate pressure on scarce resources.

This is particularly important because inflation does not require the entire economy to experience shortages simultaneously. If artificial intelligence investment creates substantial demand in a few critical industries, higher prices for those inputs can spread through supply chains. Producers facing more expensive components, electricity or equipment may eventually pass some of those costs to customers.

Chips, Energy and Infrastructure Could Spread Price Pressure

Semiconductors are among the clearest examples of how artificial intelligence can create inflationary pressure. Advanced computing systems require large quantities of high-performance chips, while demand for memory and other electronic components has also increased as data-centre construction accelerates. When artificial intelligence companies compete with traditional electronics manufacturers for limited capacity, prices can rise across industries rather than only within artificial intelligence.

The pressure is already visible in some technology-related prices. Recent economic analysis has identified rising prices for electronic components and computer-related equipment as areas where artificial intelligence investment is leaving a measurable mark. Printed circuit boards, for example, have experienced significant price increases as demand from artificial intelligence infrastructure combines with higher input costs.

Energy presents an even broader challenge. Data centres require large and continuous quantities of electricity, and the expansion of artificial intelligence computing is increasing demand for power in several major technology markets. If electricity generation and transmission capacity do not expand quickly enough, higher demand can put upward pressure on energy prices. European Central Bank Executive Board member Philip Lane has specifically identified rising energy demand during the artificial intelligence expansion as a potential source of inflationary pressure until energy supply catches up.

The effect can extend beyond the technology sector. Higher electricity demand can increase operating costs for manufacturers, transport companies and other energy-intensive businesses. If construction of data centres competes with other projects for steel, electrical equipment or skilled labour, those costs can also spread beyond the technology industry.

This is why the inflationary consequences of artificial intelligence cannot be measured simply by examining the prices of artificial intelligence products. The more important question is how much the technology is changing demand for the resources needed to build and operate it. The wider those supply constraints become, the greater the possibility that artificial intelligence investment will influence overall price levels.

Productivity Gains May Arrive Too Slowly

The strongest argument against the inflation concern is that artificial intelligence could dramatically increase productivity. If businesses can produce more with fewer resources, their costs should eventually decline. Increased productivity can also expand the economy’s productive capacity, allowing supply to grow without corresponding increases in prices. This is the long-term mechanism through which artificial intelligence could become disinflationary.

The difficulty is that productivity gains do not necessarily arrive at the same time as investment spending. Businesses may spend years installing computing systems, redesigning workflows, training employees and integrating artificial intelligence before measurable productivity improvements become widespread. During that period, investment demand can be immediate while the supply benefits remain gradual.

Research from the Bank for International Settlements illustrates this tension. Its analysis finds that artificial intelligence can raise output, consumption and investment, but the effect on inflation depends on how quickly productivity improvements emerge compared with the demand created by investment and higher incomes. The institution therefore does not treat artificial intelligence as automatically inflationary or deflationary.

There is another reason the productivity argument may not immediately translate into lower consumer prices. Companies do not have to pass every efficiency gain directly to customers. Some productivity gains can instead appear as higher profits, increased wages, greater investment or larger market share. Competitive pressure may eventually force businesses to share some of the benefits through lower prices, but that process is not guaranteed to happen uniformly across industries.

Tschudin’s point about recurring price declines is therefore important. A productivity improvement can reduce the price of producing a particular good without causing the overall economy to enter sustained deflation. For artificial intelligence to create a persistent downward force on general inflation, productivity improvements would have to spread broadly and repeatedly across the economy rather than simply produce occasional reductions in individual prices.

Central Banks Face a More Complicated Inflation Problem

The unusual combination of demand and supply effects is creating a difficult problem for monetary policymakers. Normally, stronger investment and economic demand can signal inflationary pressure, while stronger productivity can increase supply and reduce it. Artificial intelligence can produce both simultaneously, making it harder to determine whether rising economic activity is likely to produce lasting inflation or temporary pressure before a later productivity expansion.

The Bank for International Settlements has warned that this two-sided effect could complicate central banks’ assessment of underlying economic conditions. Artificial intelligence investment is already boosting trade, financial markets and economic activity, while the eventual productivity payoff remains uncertain and uneven between countries and industries.

That uncertainty matters for interest rates. If central banks interpret artificial intelligence investment as purely inflationary, they could keep monetary policy tighter for longer than necessary if productivity gains eventually increase supply substantially. If they assume that artificial intelligence will quickly reduce inflation and ease policy too early, they could underestimate the inflation created by investment, energy demand and supply shortages.

The challenge is particularly difficult because artificial intelligence investment is concentrated in specific countries and industries. The United States and China are receiving large amounts of investment, while other economies may experience the effects indirectly through higher global demand for chips, energy and industrial inputs. The inflationary impact can therefore differ significantly between economies even when they are exposed to the same technology.

For central banks, artificial intelligence is consequently becoming more than a productivity story. It is a new source of uncertainty about the relationship between investment, supply, demand and prices. The technology could ultimately increase productive capacity enough to reduce inflationary pressure, but the transition itself may be expensive.

The key economic issue is therefore not whether artificial intelligence will eventually make some activities cheaper. That outcome is plausible and already visible in productivity improvements in some applications. The harder question is whether the supply of chips, electricity, infrastructure, labour and other inputs can expand quickly enough to prevent the investment boom from generating broader price pressure before those productivity gains arrive.

If infrastructure expands rapidly and productivity spreads across the wider economy, artificial intelligence could become a powerful source of disinflation over time. If demand for computing and energy consistently outruns supply, however, the technology could remain an inflationary force for longer than policymakers currently expect. That uncertainty explains why central banks are increasingly treating artificial intelligence not simply as a technological revolution, but as a potentially important new factor in the future path of inflation and interest rates.

(Adapted from Forexfactory.com)

Categories: Economy & Finance, Regulations & Legal, Strategy, Uncategorized