A Wall Street Journal analysis revealed that nine major tech companies, including Alphabet, Amazon, Meta, Microsoft, Oracle, Nvidia, Broadcom, AMD, and SpaceX, have committed to approximately $3 trillion in future spending that largely does not appear on their balance sheets. This off-balance sheet leverage is attributed to two main categories: about $1.2 trillion in uncommenced leases for data centers and roughly $1.9 trillion in purchase commitments for items like chips and servers. This $3 trillion figure is significantly higher than the approximately $600 billion these companies spent on traditional capital expenditures over the past year and roughly triple their currently reported outstanding leases and long-term borrowings.
This hidden leverage stems from accounting rules that allow companies to keep these obligations off their balance sheets until the money is actually spent or the service delivered. For instance, Meta's Hyperion Data Center project, a $27 billion venture, is built and owned by a joint venture, with Meta providing a 20-year lease payment guarantee that isn't recorded as a liability because it's deemed not probable to be enforced. Similarly, purchase commitments for chips and power for data centers, like Alphabet's $707 billion in long-term supply agreements as of June 30, 2026, are often non-cancelable but remain off the balance sheet until delivery.
Analysts and investors are concerned because this practice understates actual leverage and future liquidity needs. While companies are betting on continued rapid growth in AI demand to justify these commitments, they are generally on the hook for these expenses regardless of profitability. For example, Alphabet's contractual obligations jumped from $322 billion to $811 billion in a single quarter, a 152% increase, mainly for chips and power for data centers. Meta's purchase commitments also rose from $238 billion to $349.3 billion over three months, largely for cloud capacity and technical infrastructure. The lack of uniform and transparent disclosure across companies makes it challenging for analysts to fully assess this financial risk.