The off-balance sheet commitments related to AI infrastructure by US Big Tech have been estimated at approximately $3 trillion. Analysts suggest that these commitments could increase future cash flow burdens as long-term contractual obligations accumulate. A comparison of the disclosure notes from nine companies, including Alphabet, Meta, Amazon, Microsoft, Oracle, NVIDIA, Broadcom, SpaceX, and AMD, revealed that the undisclosed lease commitments amount to $1.2 trillion, while purchase commitments total $1.9 trillion. During the same period, these companies' publicly disclosed capital expenditures were about $600 billion. There are cases where leases that have not yet begun or products/services that have not been delivered are not reflected as liabilities on the balance sheet. Meta's Hyperion data center has a structure with a four-year lease starting in 2029 and a maximum 20-year renewal option, with the initial lease commitment set at $12.3 billion. Alphabet's purchase commitments increased to $811 billion in the second quarter of 2026, and Meta's purchase commitments also rose to $349.31 billion. There are analyses suggesting that Big Tech's AI purchase commitments have exceeded $1.5 trillion, and this tally broadly illustrates the scope of off-balance sheet burdens. Oracle's remaining performance obligations for the fiscal year 2026 were $638 billion, while free cash flow was a deficit of $23.7 billion. Long-term commitments can serve as a proactive means of securing infrastructure when AI demand continues, but if demand is delayed beyond expectations, fixed costs and financial expenses may increase.
This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.









![[Alpha Analysis] Peter Thiel Allocates 72% of Portfolio to Energy... The Bottleneck of AI Investment Shifts](/public-static/7_ca1b7746d1.png?format=avif)












WEEX launches Season 6 of its Trade to Earn campaign from August 17 to September 6, 2026 — its most rewarding edition yet, stacking real-time WXT rebates up to 35%, a brand-new Treasure Map fragment hunt, a USDT trading leaderboard, and a referral bonus into a single campaign.







