Compute is becoming a tradable commodity as AI models run on GPUs, servers, cloud capacity, and vast electricity, with future access to AI computing power potentially financialized through contracts for GPU-hours, AI inference capacity, data center power, and reserved cloud processing.

Investors may soon value compute infrastructure like scarce resources, favoring owners of data centers, chip manufacturing, power, and network infrastructure over pure software developers.

Notable players shaping this shift include Nvidia as a GPU leader, Broadcom in AI networking, Vertiv for data center cooling and power, and energy firms like Constellation Energy, Vistra, and NextEra Energy, alongside Digital Realty Trust as a data center REIT.

Demand remains tight as Nvidia faces chip supply constraints and hyperscalers such as Microsoft, Amazon, Alphabet, and Meta pour capital into AI infrastructure, creating a scarcity-driven pricing dynamic.

Energy demand is a rising driver, with projections that data centers could account for up to about 8% of U.S. electricity use by 2030, placing power providers and related infrastructure at the core of the AI value chain.

Industry veteran leaders like Larry Fink argue that AI infrastructure shortages across compute power, chips, memory, and electricity could spark a new trillion-dollar asset class built around futures on compute, similar to commodity futures.

The takeaway: the AI revolution could redirect investment away from software toward the underpinning infrastructure, potentially yielding a tradable market for future access to computing power and transforming compute into a financial asset comparable to a digital oil field.