compute obligations

Contractual agreements to purchase a specified amount of computing power or cloud infrastructure over a set period.

Compute obligations are legally binding commitments made by technology companies—most notably artificial intelligence developers and software firms—to purchase a specified amount of data center capacity, specialized processing hardware (such as GPUs or TPUs), or cloud infrastructure over a multi-year horizon. These agreements are typically signed with cloud hyperscalers or specialized infrastructure providers to guarantee access to the massive computing power required to train and run complex software models. These commitments are tracked in the footnotes of financial statements, often under off-balance-sheet commitments, purchase obligations, or remaining performance obligations. Because high-performance computing hardware is in high demand and short supply, companies must sign long-term, non-cancelable contracts to secure their pipelines, making these obligations a major fixed-cost component of their future operating profiles. For investors, compute obligations represent a significant fixed-cost overhang and a key indicator of capital intensity. If customer demand for a company's software or AI services falls short of expectations, the company remains contractually obligated to pay for the pre-negotiated capacity. This mismatch can severely depress gross margins, accelerate cash burn, and restrict financial flexibility. Conversely, rising compute obligations can signal management's confidence in future revenue growth, provided they can successfully monetize the capacity.

Company A signs a 5-year contract with a cloud provider committing to spend $100 million annually on GPU server capacity. In Year 1, Company A only generates $60 million in AI-related revenue. Because the $100 million compute obligation is fixed and non-cancelable, the company's infrastructure costs exceed its product revenue, resulting in a gross loss of $40 million ($60M revenue - $100M compute cost) before accounting for other operating expenses.