Is an “AI Bubble” Masking Deeper Economic Problems?

The race for artificial general intelligence has become an all-gas, no-brakes affair. As companies across the globe race to develop computerized superintelligence, capital investment has flowed to the sector at a pace unprecedented in the modern era, propping up an otherwise weak economy in what some economists are calling an AI bubble.

In the most recent quarter, wage growth for the lowest quartile was cut in half from its most recent peak in late 2022, while wages for the top quartile grew by more than 30% above that of the bottom. That suggests that the bulk of newly created wealth continues to accrue at the very top of the economy. Not good. In a consumption-based economy in which 70% of the GDP comes from consumer spending, you always want that growth targeted at the bottom in order to feed consumers into the middle class.

The number of Americans jobless for 27 weeks or more has climbed above 1.8 million, the highest level since 2017 (excluding COVID). The median duration of unemployment has increased to 10.2 weeks, signaling a cooling U.S. labor market. While overall unemployment remains relatively low, this persistence of long-term joblessness suggests deeper structural economic issues, including reduced business investment outside of AI and caution around new hiring, especially in industries where tariffs have created additional costs or uncertainty about future costs.

AI capital expenditure accounted for 1.2% of infrastructure capital expenditures as a percentage of GDP in the first quarter. That's a higher percentage than peak telecom spending as the country built out its broadband network. In fact, you would have to go all the way back to the late 19th century, when around 6% of capital expenditure was targeted at the buildout of our rail systems, to find an example of one infrastructure investment sucking up so much capital. At the rate AI’s share is growing, it is possible that it could even surpass that historic statistic.

The worst part about the AI bubble is that, as artificial intelligence improves, it will become capable of replacing an increasing number of jobs currently performed by humans. However, a close second is the way it holds back finite investment capital from other sectors in the economy, making them less able to expand in a way that will create new jobs or even protect existing ones.

Railroads and broadband were long-term infrastructure investments that had a profound and dynamic impact on the economy in terms of job creation. AI data centers are limited-use, asset-intensive facilities that leverage declining-cost technology curves (which exhibit a consistent reduction in per-unit production costs as cumulative production increases) and require frequent hardware upgrades to maintain margins. Massive amounts of space and resources (including chips, energy, and water) are devoted to training future models as companies and countries vie to establish dominance in the field, with the competition intensified by concerns about how the winners may be able to dominate the losers not only economically but also militarily.

So, even before AI can take over most jobs and render a previously unthinkable percentage of our population unemployed (something we are not even having serious conversations about how to address), it is already driving job losses by sucking up so much investment.

The recently released second-quarter U.S. GDP report had the economy growing at three percent, which was better than the 2.3 percent that had been expected. However, with IT spending representing a significant portion of that growth (Google, Amazon, Meta, and Microsoft alone accounted for about half of such investments) and most of it directed toward AI-related endeavors, it is essential to ask what that “better than expected” GDP report means to the average American. 

To summarize my concerns with this data, it was bad enough when we seemed woefully unprepared to deal with the inevitability of AI replacing a staggering number of human workers. However, we are now seeing that faith in the promise of AI is diverting enormous sums of potential investment from elsewhere in the economy, where it is desperately needed. That could leave us with an even more hollowed-out economy for workers displaced by AI to attempt to reenter.

Given that we are already battling persistent long-term unemployment, a rapid spike would send resounding shockwaves throughout the economy. It would seem prudent to be openly planning for the management of such realities before they come to fruition. Again, we are not even having conversations about something that will almost certainly cause the most significant shock to labor markets in human history.

The companies developing AI and their investors only seem to see dollar signs. But even a mountain of money may be of little comfort in the midst of the sort of societal collapse such a financial apocalypse would be likely to cause.

Dennis "Mitch" Maley is an editor and columnist for The Bradenton Times and the host of our weekly podcast. With over two decades of experience as a journalist, he has covered Manatee County government since 2010. He is a graduate of Shippensburg University and later served as a Captain in the U.S. Army. Click here for his bio. Mitch is also the author of three novels and a short story collection available here. He can be reached at editor@thebradentontimes.com.