NVIDIA Targets AI Software Subscription Revenue With $2,000 per GPU Pricing
NVIDIA CEO Jensen Huang announced an AI software subscription pricing model of approximately $2,000 per GPU annually to support running AI agents on AI factories. This strategy highlights a major corporate push toward recurring software and infrastructure service revenues alongside hardware sales. This development signals NVIDIA's ongoing evolution from a pure hardware chipmaker into a comprehensive full-stack provider of enterprise AI infrastructure and software solutions. Establishing predictable recurring software revenue could significantly strengthen the company's financial model and deepen its ecosystem lock-in. NVIDIA has not yet specified the exact software scope included in the $2,000 annual price tag or how it aligns with existing offerings like the NVIDIA AI Enterprise suite, which lists at $4,500 per GPU annually. Meanwhile, the company is rolling out complementary infrastructure architectures like OpenShell and Sentry to handle agent security and management at the hardware level.
## BACKGROUND
NVIDIA is widely known as the dominant manufacturer of graphics processing units (GPUs) that power modern artificial intelligence training and workloads. Beyond hardware, the company develops proprietary software stacks and enterprise platforms to help corporations efficiently deploy and manage accelerated computing workloads.