Nvidia has announced a significant shift in how it partners with AI cloud providers, unveiling a new revenue-sharing and credit-support model for deploying large-scale AI factories. The announcement marks a strategic pivot for the chip giant as the industry transitions from pure model training to always-on token production.
The new model targets a broad range of customers including startups, enterprises, model builders, and regional AI players. Rather than requiring massive upfront capital investments in Nvidia hardware, the revenue-sharing approach allows partners to deploy AI factories with lower financial barriers while Nvidia earns a portion of the operational revenue.
This announcement reflects the broader industry trend toward inference-heavy workloads. As AI models become more widely deployed in production applications, the demand for continuous inference capacity has grown substantially. Cloud providers are increasingly building dedicated AI inference infrastructure, and Nvidia’s new model is designed to capture more of this growing market.
The move also positions Nvidia to compete more aggressively with cloud giants like Amazon Web Services, Google Cloud, and Microsoft Azure, which have all been developing their own AI inference capabilities. By offering a revenue-sharing model, Nvidia can enable smaller players to compete with the hyperscalers while maintaining its position as the dominant hardware provider.
Industry analysts note that the revenue-sharing approach could accelerate AI infrastructure deployment across regions, particularly in markets where capital constraints have limited AI adoption. The model may also benefit regional cloud providers and startups that have struggled to secure the massive GPU clusters needed for large-scale AI deployment.
The announcement follows Nvidia’s continued expansion beyond its traditional GPU business into complete AI infrastructure solutions, including networking, software, and now financing models that make AI deployment more accessible to a broader range of customers.