The Hidden Cost of Big Tech's AI Investment

The escalating investment in artificial intelligence (AI) by the major tech players has unveiled an economically complex and challenging landscape. According to recent analysis from The Wall Street Journal, nine of the world’s largest technology firms—Alphabet, Meta, Amazon, Microsoft, Oracle, and Nvidia—have off-balance-sheet commitments related to AI infrastructure totaling approximately US$ 3 trillion. This figure is significantly higher than the capital expenditure (capex) reported by these groups over the past year and triples their current outstanding lease and long-term borrowing obligations. The report further highlights uncommenced lease obligations, such as Meta’s Hyperion data center project in Louisiana, which total US$ 1.2 trillion, four times more than a year ago. Chip and hardware purchase commitments across the companies amount to nearly US$ 1.9 trillion, with Alphabet’s alone reaching $811 billion by June, doubling from the previous quarter's report. These commitments remain off-balance-sheet until contracts are executed or products delivered, making it difficult for investors to assess the real financial exposure. With recent reports of negative free cash flow by both Alphabet and Amazon, such substantial obligations may increase risk if AI demand fails to meet expectations. These findings could intensify scrutiny on individual balance sheets heading into upcoming earnings, particularly for companies with the fastest-growing purchase and lease commitments relative to reported capex.