Artificial intelligence has moved from a technology story to a macroeconomic one. The latest evidence points in two directions at once: AI-related investment is helping support growth and exports, while the infrastructure required to run it is adding to energy demand, emissions and inflationary pressure.

That tension is becoming harder for governments and central banks to ignore. The Financial Times reports that the AI boom is helping sustain growth above long-run averages and is contributing to stronger European export performance, even as the bloc’s economies diverge. Separately, S&P Global says global inflation remained elevated in October, with fuel, energy and supply constraints feeding price increases across goods and services. The Financial Times and S&P Global

Why investors are betting on AI

The immediate economic argument is straightforward. Building data centers, purchasing advanced chips and expanding electricity networks require enormous capital spending. That spending creates demand for construction, engineering, software, semiconductors and power generation. It also promises productivity gains if companies can use AI to produce more with fewer hours of labor or lower operating costs.

Financial markets are increasingly treating those gains as a possible answer to weak underlying growth. The Financial Times has reported that policymakers and economists see the technology as capable of supporting higher U.S. growth without necessarily reigniting inflation, provided productivity improves quickly enough. The European Commission’s improved growth outlook, as reported by the paper, has also been linked partly to AI-driven exports. The Financial Times

That optimism has limits. Productivity gains are not automatic, and their distribution is uncertain. Large technology companies can finance data centers and absorb the cost of experimentation; smaller firms may face higher prices for computing and electricity. Workers may benefit from new tools, but some occupations could see reduced demand or weaker bargaining power. The economic upside therefore depends not only on technical progress, but on competition, training and the distribution of returns.

The hidden cost is power

AI systems require large data centers that consume electricity continuously. A Harvard study reported by the Associated Press found that the 400 largest U.S. data centers already produced more heat-trapping gas in one year than Denmark, before the full impact of the AI boom was reflected. The comparison captures the scale of the sector’s climate footprint, although the precise emissions burden varies according to the electricity sources used by each facility. Associated Press reporting

This creates a policy conflict. Governments want domestic computing capacity because it is increasingly associated with national security, industrial competitiveness and technological sovereignty. Yet rapid construction can strain local grids, increase demand for gas-fired generation and delay decarbonization if renewable capacity and transmission lines do not keep pace.

Utilities and technology companies argue that new data centers can help finance grid upgrades and stimulate clean-energy investment. Critics respond that the costs may be shifted to households through higher rates, while the benefits accrue mainly to shareholders and highly skilled workers. Both claims can be true: new demand can support infrastructure investment, but poorly designed pricing can socialize the expense.

Why central banks are watching

The inflation question is especially important. S&P Global says price growth remained above central-bank targets in all major developed economies in October and notes that several central banks have adopted more hawkish positions. The Bank of England raised rates in November for the first time in more than three years, while the Bank of Canada signaled a tougher stance. S&P Global

AI can eventually reduce inflation if productivity lowers costs. In the short term, however, the investment boom may increase demand for scarce chips, electricity, land, construction workers and specialized engineers. Higher energy prices can spread through the wider economy, especially when fuel costs are already passing into services and manufactured goods.

This helps explain why policymakers face a difficult sequencing problem. Tight monetary policy may restrain excessive demand but also make long-term infrastructure more expensive. Subsidies can accelerate clean power and domestic technology production, but they may worsen fiscal pressure or encourage overinvestment. Restrictions could reduce environmental damage, yet excessive regulation could push computing projects to countries with weaker standards.

What comes next

The next phase will be defined less by headline promises than by bottlenecks. Governments will need clearer rules on who pays for grid connections, how emissions are measured and whether data centers receive preferential access to scarce electricity. Investors will demand evidence that AI revenues can justify the cost of chips, facilities and power contracts.

The most credible policy path combines competition rules, transparent energy pricing, accelerated transmission construction and firm emissions standards. It also requires acknowledging uncertainty: the productivity gains may be substantial, but they may arrive unevenly and later than markets expect.

AI is therefore neither a simple growth miracle nor merely an environmental liability. It is a test of whether governments can convert technological investment into broad productivity gains without allowing infrastructure costs, climate damage and inflation to become the bill paid by everyone else.

Sources