In August 2026, the artificial intelligence industry faced a new stage of financial transparency, revealing the massive commitments of leading technology corporations. As it turned out, real investments in AI infrastructure are many times higher than the official figures reflected in financial statements. According to The Wall Street Journal, the total commitments of tech giants reached $3 trillion, which is five times their official capital expenditures for the previous fiscal year.
Beyond the Balance Sheet: Trillions in Hidden Commitments
Classic capital expenditures (CapEx) of leading American technology companies last year reached $600 billion. However, beyond this figure lies an even more impressive sum — about $3 trillion in commitments related to the development of AI infrastructure. This category includes future payments under contracts for the rental of computing power, debt obligations, and other financial instruments that are not reflected on the companies' balance sheets until the facilities begin operation.
For example, Alphabet (Google) increased its hidden commitments by more than eight times compared to last year, while official capital expenditures doubled. This indicates that companies are actively using complex financial schemes to finance large-scale projects without overloading their balance sheets.
Structure of Commitments and the Role of Lease Contracts
The eight largest US technology companies have commitments totaling $1.2 trillion that are not reflected in their balance sheets. These commitments include long-term contracts for the rental of computing power and loans stretching as far as 2054. For example, Alphabet has commitments totaling $811 billion, which cover not only the construction of data centers (DCs) but also expenses for energy and engineering infrastructure.
Meta, in turn, had data center lease commitments totaling $347 billion by mid-2026. These commitments allow companies to build facilities on "other people's money," using bonds and financial guarantees that do not require reflection on the balance sheet until the start of operations.
Risks and Consequences for Investors
Despite massive investments, companies are not afraid of an infrastructure surplus in the future. However, risks remain significant. If projects do not bring the expected profits, companies will have to pay debts for facilities that will not have a worthy application. This could lead to the need to take out new loans to pay off old ones, creating additional pressure on financial stability.
Already now, Alphabet and Amazon have gone "negative" on free cash flow, as their capital expenditures exceed operating income. This is an alarming signal for investors, who are beginning to fear the growth of borrowings and potential financial problems.
Contradictory Data
There are disagreements regarding the accuracy of the assessment of hidden commitments. Some analysts believe that $3 trillion is an inflated figure, as not all commitments may be realized in full. Other experts, on the contrary, point out that real commitments may be even higher, given the dynamic growth in demand for AI infrastructure.
Furthermore, there are questions about the methodology for calculating commitments. For example, some companies may use complex financial instruments that make it difficult to accurately assess their real commitments. This creates uncertainty for investors and regulators.
Conclusions and Prospects
The hidden commitments of tech giants in the AI sector have reached unprecedented scales, indicating a massive boom in this industry. However, risks remain significant, and companies must be prepared for possible financial difficulties. Investors should closely monitor the development of the situation and take potential risks into account when making decisions.
