AI Boom Fuels Inflation Through Infrastructure Costs Before Delivering Savings

While artificial intelligence promises future cost reductions and productivity gains, its initial rollout is driving significant inflation due to massive infrastructure investments. Tech giants like Amazon and Microsoft are projected to spend around $410 billion on capital expenditures by 2026, with the total industry investment approaching $700-800 billion annually. This spending fuels demand for physical resources such as land, construction, millions of computer chips, transformers, power lines, cooling systems, generators, and water, creating bottlenecks in supply chains.

The surge in demand for electricity is a major consequence, with U.S. power consumption expected to rise by approximately 2% annually in 2026 and 2027, largely due to server farms. These data centers could account for half of the increased demand by 2030. Expanding the power grid and manufacturing essential components like transformers are lagging behind this rapid growth, leading to increased costs for electricity and other resources like copper, which is crucial for cables, transformers, and cooling systems.

In Israel, the demand for server farm connections has reached approximately 27,000 megawatts, far exceeding the country's average consumption of around 9,000 megawatts. This situation raises concerns about who will bear the cost of infrastructure upgrades, with some U.S. states already requiring data centers to make minimum consumption and payment commitments.

The article explains that the AI revolution is currently in a "demand shock" phase, characterized by increased economic activity and inflation, rather than the "supply shock" phase where productivity gains would lead to cost reductions. The International Monetary Fund estimates that AI could add 0.1 to 0.8 percentage points to global potential growth in the long term, but the immediate impact is inflationary.

Eventually, increased investment in hardware and more efficient AI models are expected to lower computation costs. Furthermore, as AI tools become commonplace, businesses will likely pass on savings to consumers, and the labor market may see a moderation in wage increases as certain tasks become automated. However, these productivity benefits are anticipated to materialize years after the initial inflationary pressures from infrastructure build-out.