THE CRUNCH
Nvidia chief executive Jensen Huang has put the cost of licensing the company's AI software at roughly $2,000 per GPU per year, describing it as the fee to "run all the software that Nvidia provides that runs agents on these AI factories." He made the comments at GTC Berlin on October 22, where the company framed software licensing as an increasingly major part of its business alongside chips, networking and AI-agent security.
The arithmetic behind such a model is substantial. At $2,000 per GPU annually, five million covered licences would generate $10 billion a year, or $50 billion over five years if customers kept renewing. That scenario, outlined by FourWeekMBA, assumes a large paid base that Nvidia has not confirmed: the company has not said how many GPUs the fee would apply to, whether it is a standard price, or which products the licence would include.
The figure invites comparison with Nvidia's existing enterprise software. Its AI Enterprise platform is listed at $4,500 per GPU for a one-year subscription. The $2,000 figure is lower, but Nvidia has not confirmed that the two prices cover the same products or terms, so it should not be treated as a confirmed price cut.
The pricing discussion sits within a broader push to secure AI agents in enterprise settings. On October 21, Nvidia announced its OpenShell Reference Architecture, combining Vera Rubin computing infrastructure, Vera CPUs and BlueField-4 data processing units running Nvidia Sentry, which monitors agent activity and can quarantine an agent that crosses set boundaries. Anthropic is integrating OpenShell with Claude Managed Agents, and Salesforce has built a Slack Admin Bridge for reviewing agent access requests.