Big Tech Firms Commit $1.09 Trillion in Future AI Data Center Leases

Big Tech’s aggressive pursuit of artificial intelligence computing power has quietly built an enormous financial commitment that sits largely off traditional balance sheets. According to company filings compiled by Reuters, five major technology companies have committed about $1.09 trillion in future payments under leases that have not yet begun, mostly targeting the data centers essential for running advanced AI infrastructure.

Reuters compilation of uncommenced lease payments

This massive pipeline reveals that a substantial portion of the industry’s ongoing spending spree has already been locked in legally, even though these obligations do not yet appear as debt-like lease liabilities on corporate balance sheets. Accounting rules dictate that signed leases are generally not recorded as liabilities until a facility is actually available for use. Until that operational milestone is reached, companies disclose future payments in the notes to their financial statements.

–>

The scale of these future commitments towers over what has already been formally recognized. According to company filings, the $1.09 trillion pipeline is nearly four times the roughly $285 billion of lease liabilities already recognized on the companies’ books. While rating agencies may already account for some of these disclosures, the sheer volume highlights how much of the infrastructure buildout remains outside reported fixed charges and leverage metrics.

Analysts emphasize that these figures cannot simply be added directly to corporate debt totals. Uncommenced lease commitments generally represent undiscounted payments spread over many years, whereas recognized lease liabilities reflect present value calculations. Even so, the distribution of this risk varies widely across the firms driving the artificial intelligence expansion.

–>

Oracle concentration risk and fiscal disclosures

Among the major players, Oracle carries the largest apparent concentration risk. The company disclosed $260 billion of uncommenced commitments, which is nearly seven times its $37.89 billion of recognized lease liabilities, according to company disclosures. These commitments are substantially tied to data centers expected to begin between fiscal 2027 and fiscal 2029, running for durations of 15 to 19 years.

Oracle has explicitly warned that the duration, renewal terms, and pricing of its data center leases may not align properly with customer contracts. This mismatch leaves the company exposed if customers ultimately fail to renew or cannot perform. At the end of May, Oracle’s borrowings equaled about 4.4 times trailing EBITDA, a figure that jumped to about 5.7 times once recognized operating and finance lease liabilities were included, based on a Reuters analysis of LSEG data and company filings. S&P Global Ratings noted that it incorporated Oracle’s $260 billion of uncommenced leases into its adjusted-debt forecast, projecting leverage to hover around 4.4 in fiscal 2027.

AI Data Center Bubble: The $1 Trillion Race That May Crash

Other tech giants reported even larger total pipelines, though with different portfolio compositions. Microsoft disclosed the largest overall pipeline at $329.1 billion, standing against $88.52 billion of recognized lease liabilities. Meta Platforms reported $278.99 billion in uncommenced operating and finance lease payments, and subsequently signed an additional $68 billion in data center leases in July, pushing the collective pipeline for the five companies to about $1.16 trillion.

Rounding out the group, Alphabet reported $85.2 billion in uncommenced leases, while Amazon disclosed $137.21 billion. Amazon’s figure carries a broader scope because its lease portfolio encompasses not only data centers but also warehouses, offices, aircraft, and vehicles.

–>

Amazon and Alphabet lease portfolio compositions

The financial architecture of the AI boom now rests on a foundational question of future demand. If the surge in artificial intelligence computing continues unabated, these facilities will provide the necessary bedrock for the next phase of cloud growth. However, if demand falters or fails to materialize at the expected scale, these corporations risk paying for vast amounts of costly, long-lived capacity that cannot be easily shed.

Keep reading


Discover more from Archyworldys

Subscribe to get the latest posts sent to your email.