As highlighted below, environmental financing deals are not the only sign of the AI boom wall street journal. Off their balance sheets, the tech giants are willing to spend many times more than their officially announced capital expenditures, which in their own right are sizable. According to the publication’s authors, the total amount of such liabilities has increased to $3 trillion.
Image source: Intel, Adobe Stock
By comparison, classic capital expenditures reflected in the financial statements of leading U.S. technology companies rose to $600 billion in the previous fiscal year. At the same time, the implicit liabilities of U.S. tech giants related to the development of artificial intelligence infrastructure increased approximately threefold in monetary terms at that time. Such financial resources include upcoming payments under computing power lease agreements and various types of loan obligations.
Taking the US holding company Alphabet (Google) as an example, such implicit liabilities have increased more than eight times compared with last year, while the official capital expenditure amount has almost doubled. Now, U.S. tech giants are investing heavily in building artificial intelligence infrastructure and aren’t too worried about future overabundance. These facilities sometimes have complex ownership structures; they are often owned by other companies and, from the perspective of the end operator, were built with “other people’s” money. To implement the project, bonds are issued and financial guarantees are provided by companies such as Meta✴ Until the facility being built is operational and rent is paid, it should not appear on the balance sheet. By the middle of this year, the same yuan✴ Has $347 billion in data center lease obligations.
Eight major U.S. technology companies currently have such liabilities totaling $1.2 trillion, although the companies do not show this amount on their balance sheets. Payments on long-term contracts and loans totaled $1.9 trillion, bringing the total to more than $3 trillion. Of this amount, $811 billion came from Alphabet; this amount also includes costs related to the implementation of projects in the fields of energy and data center construction engineering infrastructure. Payment terms under the contract extend to 2054, so payments do not need to be made within a short period of time. On the other hand, just three months ago, Alphabet’s debt did not exceed $332 billion.
Sometimes such obligations involve purchasing stock in other companies. Nvidia alone is preparing to allocate up to $27 billion to these needs between April this year and the end of January 2027. Alphabet and Amazon’s free cash flow has turned negative as their capital expenditures have begun to outpace operating profits, and that’s before taking into account off-balance sheet obligations. If everything doesn’t go as planned, the company will have to pay down debt on building facilities that won’t find a valuable and profitable use. In many cases, you will have to take out new debt to pay off old debt. Anxious investors are increasingly worried about the growth of such borrowing.
If you find an error, select it with your mouse and press CTRL+ENTER.
