• Goldman Sachs and JPMorgan are exploring ways to trade or finance the cost of computing power, signaling a new frontier in Wall Street's bet on AI infrastructure.
  • The move reflects a broader shift toward monetizing compute capacity amid surging demand for data centers and AI hardware.
  • Both banks are taking divergent approaches: Goldman has scaled back speculative quantum projects, while JPMorgan pursues a more application-focused strategy.

Wall Street's New Frontier: Compute Power as an Asset

Goldman Sachs and JPMorgan Chase are evaluating strategies to trade on the cost of computing power, according to people familiar with the matter. The initiative marks a significant expansion of Wall Street's engagement with AI infrastructure, moving beyond traditional financing into direct monetization of compute capacity.

“This is about treating computing power as a tradeable commodity, much like energy or bandwidth,” a senior banker at one of the firms said, speaking on condition of anonymity. The banks are exploring instruments tied to data-center costs, GPU pricing, and energy consumption, according to the people.

Diverging Strategies

Goldman Sachs has quietly scaled back some of its quantum computing investments, focusing instead on near-term opportunities in AI-driven compute financing. In contrast, JPMorgan has maintained a more aggressive posture, with dedicated teams working on scalable applications of advanced computing for trading and risk management, the people said. The divergence reflects broader strategic bets on the maturity of different technologies.

“JPMorgan sees a clear path to monetizing compute power today, while Goldman is more cautious about the hype cycle,” said an industry analyst who tracks AI infrastructure investments.

The exploration comes as both banks report solid earnings, underpinning continued investment in technology. Goldman's TMT team has been actively advising on data-center deals, while JPMorgan has expanded its in-house AI capabilities.

Market Implications

The cost of computing power has become a critical input for AI-driven businesses, with data-center build-outs driving demand for specialized hardware and energy. Banks are increasingly financing these assets, potentially reshaping project finance in the tech infrastructure space. Analysts note that the shift could lead to new financial products, such as compute-capacity derivatives or securitized data-center assets.

“If successful, this could create a new asset class,” said a partner at a private credit firm that has partnered with banks on data-center deals. “But it's early days, and the economics are still being tested.”

Regulatory and environmental considerations loom large. Data centers consume vast amounts of energy, drawing scrutiny from policymakers and communities. Banks will need to navigate sustainability concerns as they structure deals.

Broader Context

The trend fits within a longer arc of Wall Street betting on technology-driven productivity gains. Parallel developments include other banks forming dedicated teams to invest in AI infrastructure. The financing of compute power is also drawing interest from asset managers and private credit funds, who see it as a way to gain exposure to the AI boom without taking on equity risk.

Efforts to reach spokespeople for Goldman Sachs and JPMorgan for comment were not immediately successful.

Correction: An earlier version of this article misstated the timeline of Goldman's quantum pullback. The scaling back occurred over the past two quarters, not the past year.