Is the AI Boom Depressing the Rest of the Economy?

Over the past four years, Dallas has transformed into the center of the U.S. data-center economy thanks to its affordable land, friendly regulations, ample workforce, efficient transportation, and robust power grid. Presently, approximately 125 data centers are operational in the metro area, with an additional 19 under construction and 53 planned. This building boom has raised wages for blue-collar workers and maintained the local unemployment rate near its historical minimum of 4.1%.

However, Dallas and its suburbs are also experiencing a construction downturn. Despite high rents and a growing population, new housing developments are scarce: the number of new homes began to decline by 12% year-over-year, and the number of planned multifamily units has plummeted by 43% over the last three years. Moreover, office construction is also faltering. This prompts the question: is the AI boom offsetting these other downturns or causing them?

Investors are currently more enthusiastic about AI than they have been about any technology since the railroad. If they weren't funneling a large portion of their funds into AI firms, the nation's employment levels and economic output might be lower. Conversely, if they weren't directing their investments to AI firms, they might be spreading them across other sectors.

AI is certainly contributing to GDP growth. In 2020, American businesses invested $3.6 trillion, adjusted for inflation, in durable assets like software, warehouses, and trucks. In 2021, they invested $4.4 trillion. Major tech companies, including Oracle, Microsoft, Amazon, Meta, and Alphabet, are allocating $800 billion toward capital expenditures this year alone – a sum that has increased eightfold over the past decade. This tech-driven stimulus dwarfs the one provided by the U.S. government during the Great Recession.

This surge in private-sector investment has impacted national accounts. In the second quarter of this year, household consumption increased at a 3.8% annualized rate, while private investment across all sectors, except for real estate, grew at a 12.7% pace. Equipment spending rose by 21%, computer purchases added 0.7 percentage points to GDP, and "information processing equipment" boosted GDP by a significant 0.9 percentage points.

The early AI boom was primarily financed by wealthy venture capitalists and profitable tech firms. However, they now require more computing power to refine their models, leading them to build more data centers and chip factories, which necessitate substantial additional liquidity.

The growing demand for loans has driven up borrowing costs, causing crowding out of non-tech enterprises. Data centers are being constructed at a rate of 543, with an additional 2,513 in the pipeline. Spending on the construction of data centers has increased by 65% over the past year. At the same time, spending on housing, retail outlets, education facilities, offices, hotels, motels, and manufacturing plants has decreased by 10%, 13%, 14%, 15%, and 24%, respectively.

Tech enterprises are not just making it more expensive for non-tech firms to finance construction projects – they are also bidding up the cost of labor, land, and materials. They are pushing landlords and institutional investors towards Treasuries or stocks instead of real estate ventures. Families are paying more for home loans, and small businesses and individuals looking to launch a new venture or enroll in a training program are facing increased costs as well. In a roundabout way, AI-driven firms might even contribute to higher credit-card delinquencies.

However, the impact of AI on the economy may not be as significant outside of the construction and real-estate sectors, which are already under pressure due to interest rate hikes. Nonetheless, AI is undoubtedly reshaping the American economy, raising both questions and concerns about its long-term effects.