AI's never-before-seen capital grab

The AI industry is experiencing an unprecedented capital surge, with an estimated $10.3 trillion in infrastructure investment projected through 2032. This investment accounts for 3.6% of global GDP each year, surpassing the financial outlays for past transformative projects like railroads, highways, electric grids, and telecommunications.

The scale of the AI boom has made financing challenging for the largest tech firms, resulting in trillions of dollars in demand for external capital. This has placed significant strain on bond markets, banks, and private credit, adding to the existing debt competition at a time when borrowing costs are already rising. The increased borrowing costs, however, have not dampened the industry's appetite for growth, as evidenced by Meta's strategy of financing its $30 billion Hyperion data center mostly through external investors, despite paying higher interest rates.

The flood of capital intended for AI is also putting pressure on other borrowing entities, including the U.S. government, individuals seeking loans, and pension funds. This competition for funds could potentially lead to higher interest rates across the economy. Stijn Van Nieuwerburgh, a Columbia Business School professor who authored a paper on the financing of AI infrastructure, notes that the industry's investment appetite remains robust even in the face of higher rates.

While the influx of outside money is fueling the AI boom, it also introduces potential risks to the financial system. These include complexities in tracking risk distribution and the possibility of an oversupply of AI infrastructure, leading to a potential collapse in prices - a scenario reminiscent of past real estate cycles. Van Nieuwerburgh warns that the AI investment boom could soon outpace actual demand, potentially leading to negative consequences for the broader economy.