Nvidia discusses $2,000-per-GPU annual fee for software that 'runs agents on AI factories' — details unclear about potential overlaps, firm already charges more than twice that amount for its AI Enterprise software platform - Ireland News Beep

Add us as a preferred source on Google

Newsletter

Subscribe to our newsletter

Nvidia CEO Jensen Huang has put the annual cost of licensing the company’s AI software at around $2,000 per GPU to ‘run all the software that Nvidia provides that runs agents on these AI factories,’ according to a recent video of the comments (below), a pricing model that could generate $50 billion over five years under a proposed scenario involving five million paid licenses outlined by FourWeekMBA.

Huang discussed the pricing at GTC Berlin on October 22, where the company outlined its broader push to build an AI infrastructure business spanning chips, software and AI-agent security. “Software licensing will become increasingly a major part of our company’s business. The way that we license will be…for example, it’s about $2,000 per GPU per year to run all the software that Nvidia provides that runs agents on these AI factories,” said Huang.

The numbers add up quickly. At $2,000 per GPU each year, Nvidia would bring in $10 billion annually if five million GPUs were covered by paid licenses. Over five years, that would amount to $50 billion, provided customers kept renewing their subscriptions. The catch, however, is that Nvidia has not disclosed how many GPUs could be covered by the proposed $2,000 fee, or whether that’s its standard price. It’s also unclear which software the license would cover, or if there is overlap between products.

You may like

On October 21, Nvidia announced its OpenShell Reference Architecture, a system that combines its Vera Rubin computing infrastructure, Vera CPUs, and BlueField-4 data processing units (DPUs) running Nvidia Sentry. OpenShell provides a secure runtime for AI agents, establishing boundaries around their actions and access to tools and resources. Sentry monitors agent activity independently and can quarantine an agent that attempts to cross those boundaries, according to Nvidia. The architecture is designed to bring security controls into the underlying infrastructure rather than relying solely on safeguards built into an AI model or application.

Other companies are also incorporating these controls into their products. Anthropic is integrating OpenShell and BlueField hardware with Claude Managed Agents to give businesses additional control over what their AI agents can access inside execution sandboxes. Salesforce has developed a Slack Admin Bridge for OpenShell that lets administrators review agent access requests, approve or reject permissions, and monitor activity directly in Slack.

The proposed software pricing fits Huang’s wider argument that Nvidia is becoming more than a supplier of GPUs. The company has expanded into networking, systems, enterprise software, and infrastructure designed to run AI workloads. Speaking on the BG2Pod podcast recently, Huang said, “One of these days, people might not categorize us as a chip company anymore.” He pointed to Nvidia’s presence across industries including healthcare, automotive, telecommunications and life sciences, where the company provides broader technology stacks rather than chips alone.

Nvidia discusses $2,000-per-GPU annual fee for software that ‘runs agents on AI factories’ — details unclear about potential overlaps, firm already charges more than twice that amount for its AI Enterprise software platform

Tags: