Belated Comments On The AI Boom

Animal spirits have been soaring in the area of Artificial Intelligence (AI), and the surge in spending has shown up in the figure above. It shows the fixed investment in Information Technology Equipment and Software as a percentage of GDP (ends in 2026Q1). The last two quarters of data saw a surge of about 0.4% of GDP.

This is certainly impressive, although the impact on GDP may be less than what the above figure suggests due to the associated rise in imported silicon chips — buying foreign equipment to invest locally improves the capital stock, but does not represent a rise in domestic production. (The exporting country is producing the goods.) This relates to one of the perennial online economics debates: do imports subtract from GDP? In addition to the statement “imports subtract from GDP” being mathematically correct, the cancellation of domestic spending does matter: there is a financial cost associated with buying foreign goods, and that financial cost can displace spending that would have been made on domestic production.

For more information, the article by Hannah Rubinton and Bontu Ankit Patro of the Saint Louis Federal Reserve gives a more detailed breakdown of the effects of AI spending on growth. From a quick survey of publicly available research, there are arguments that the AI boom added 1% to American GDP growth, which is consistent with the linked paper (and the chart above). This helped insulate the U.S. economy from tariff shenanigans.

I have no real expertise nor interest in forecasting the future of the AI industry relative to other sectors of the economy. My interests are on the macro side, and my feeling has been that this has been a sectoral boom. This matters a lot to the people, firms, and investors involved, but the number of people involved are a limited slice of the population. This is different than the 2000s housing boom — there are a lot of homeowners, and construction as well as real estate are large employers. At the same time, the 2000s boom was a global boom built around wacky financial innovations. Even the technology boom of the late 1990s may not be directly comparable — there was a wider global investment boom that resulted from reversing the corporate retrenchment in the 1980s (as well as the beginning of the housing bubble).

It is straightforward to argue that the current growth rate of AI investment cannot be sustained, but that does not say very much (and is useless for equity investing). Investment levels could be sustained at “high” levels for longer than most sceptics would predict. Economic commentators tend to love forecasting popping bubble and the dire effects thereof, but my feelings on the sectoral nature of the boom also lead me to have limited worries about any potential retrenchment. My main concern is that a “popping of the AI bubble” would likely coincide with equity market weakness, and so “animal spirits” more broadly would be at risk.

The main risk to the global economy is the ongoing cut off of energy flows, which is hitting Asia and Africa first. There is a direct analogy to the Asian Crisis in 1997, with the American tech-driven economy shrugging it off — until financial market blowback hit with the “LTCM Crisis” in 1998 (a few months after I entered the financial industry). American tech investors have not fundamentally changed and they are not going to care about what happens to the rest of the world (or even the rest of the American economy). As such, there is no reason to expect that the financial market counterpart of the AI boom will lead global growth trends in a downturn, instead, weakness elsewhere would finally crack tech sector optimism.