California Antitrust Push Targets Finance Amid NVIDIA Retreat

The antitrust question is moving up the stack: What California’s new push and NVIDIA’s retreat mean for finance. California-based NVIDIA recently paused parts of its AI Compute Partnership Program, a financing initiative that helped smaller AI cloud companies buy NVIDIA chips and build data centers. The program addressed a key problem where these companies need billions of dollars in infrastructure before they have enough customers to easily secure traditional financing. NVIDIA’s proposal allowed the chipmaker to not only provide the necessary capital but also to align its financial returns with the performance of the client’s business. By sharing in the revenue those customers generated and potentially taking back unused computing capacity, NVIDIA reduced the risk of default for the startups while ensuring it retained a claim on the valuable hardware assets. This commercial structure created a cooperation where the financier and the supplier were effectively operating as the same entity. However, this tight integration raised concerns about whether NVIDIA was overstepping by controlling both the supply of the critical component and the financing required to utilize it. Internal warnings circulated among employees that such a full structure could invite antitrust scrutiny from regulators. Meanwhile, prospective partners objected to specific restrictions that limited their freedom to manage their compute assets independently. Consequently, NVIDIA paused some of those deals, although its broader effort to finance AI infrastructure remains active.

The firm has been moving toward a broader institutional-capital model, joining Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR in an initiative aimed at mobilizing more than $500 billion of third-party capital for AI infrastructure. This transition suggests a calculated shift in strategy where NVIDIA is reducing direct exposure to individual financing risks while partnering with giants of the asset management world. It is becoming increasingly difficult to separate ecosystem building from market dominance. In previous industrial sectors, similar entanglements—such as a utility company controlling both generation assets and transmission networks—have blurred the lines between operational necessity and restrictive control. Future regulatory challenges are likely to focus less on the traditional black-and-white issue of direct price-fixing and more on the complex structural integrity of the entire supply chain.

Related: BILL Takes Lead as SMB Financial Backbone

Legislative response in California

What happened with NVIDIA points to a much bigger question about how market power works in increasingly interconnected industries. At which point does helping an ecosystem grow start using your position to shape how that ecosystem works is the question California lawmakers are now considering. California’s antitrust law has traditionally focused on coordination between companies, such as competitors agreeing to fix prices or divide markets. Assembly Bill 1776, known as the COMPETE Act, is a proposed California law that would expand the state’s antitrust framework, the Cartwright Act. This legislation represents a significant departure from past approaches by broadening the law to reach certain single-firm conduct. In practice, this change means regulators will look more closely at how powerful companies leverage their existing market position rather than just looking for agreements between rivals. Lawmakers are examining scenarios where a single entity holds too much influence, such as a bank that owns the primary customer relationship while simultaneously competing against the fintech companies it distributes. Similarly, the focus is turning toward whether it creates an unfair advantage for NVIDIA to finance the cloud providers that are purchasing Nvidia’s chips, effectively controlling the entry points into the market.