Nvidia Launches Revenue-Share Program for AI Startup Infrastructure

Nvidia's new infrastructure play

Nvidia has launched a program that gives AI startups access to high-performance computing through a revenue-sharing arrangement with cloud providers, the company announced Wednesday. Under the initiative, cloud infrastructure companies will deploy Nvidia hardware and offer services to AI firms, with Nvidia earning revenue from both equipment sales and a portion of the providers' future income.

The approach addresses a persistent challenge in the AI sector: emerging companies often lack the capital to purchase or lease the expensive GPU clusters required for training and deploying large models. By embedding itself in the cloud provider layer and accepting deferred compensation, Nvidia effectively finances infrastructure buildouts while expanding the installed base of its processors.

Early deployments target scale

Cloud providers Sharon AI and Firmus are among the first participants in the program, both building infrastructure on Nvidia's DSX data center platform. Sharon AI has committed to deploying up to 40,000 Grace Blackwell GB300 GPUs, while Firmus is constructing what it describes as a DSX AI factory campus in Batam, Indonesia.

The scale of these initial commitments signals that the program targets substantial infrastructure projects rather than pilot deployments. The Grace Blackwell architecture represents Nvidia's latest generation of AI-optimized processors, combining ARM-based CPUs with high-bandwidth GPU accelerators.

Why it matters

This revenue-sharing model represents a strategic shift in how semiconductor companies monetize AI infrastructure. Rather than simply selling chips, Nvidia is positioning itself to capture ongoing value from the applications and services built on its hardware. For startups, the arrangement potentially lowers upfront costs, though it also creates long-term financial obligations tied to their cloud spending. The structure could accelerate AI adoption by reducing capital barriers, but it also deepens the ecosystem's dependence on a single hardware vendor.

Investment strategy draws scrutiny

The new program extends Nvidia's broader pattern of AI ecosystem investments. According to reports cited by Benzinga, the company committed over $40 billion to AI-related investments in early 2025, including a $30 billion stake in OpenAI and significant investments in Corning and IREN.

This investment approach has generated debate among market observers. Investor Michael Burry has criticized what he characterizes as circular financing, suggesting that capital flows primarily among interconnected firms rather than creating genuinely independent demand. Following CNBC's Jim Cramer praising Nvidia's position, Burry dismissed the chipmaker's apparent strength as "all Fugazi," implying that billions in GPU deployments might be obscured through accounting complexity.

Conversely, Futurum Group CEO Daniel Newman has defended the strategy as sound business practice. "AI Bubble bears will call it circular financing. I call it prudent investing," Newman stated, arguing that the approach enables Nvidia to scale operations and increase cash generation.

The tension between these perspectives reflects broader uncertainty about AI infrastructure economics: whether current buildouts represent sustainable demand or speculative overinvestment remains an open question as the technology matures.

These details were first reported by Benzinga and published on Yahoo Finance.