The software operations of Nvidia, which as of now generate $1 billion annually, are set to become a major part of the company’s business.
Nvidia may boost its software business by introducing a proprietary enterprise software licensing strategy where one GPU may cost $2000 annually. CEO Jensen Huang unveiled the figure at GTC Berlin.
The software operations of Nvidia, which as of now generate $1 billion annually, are set to become a major part of the company’s business. According to Huang, while much of Nvidia’s software is open, proprietary models will be offered via enterprise licensing.
The chipmaking giant also unveiled a reference architecture based on OpenShell open-source runtime software at the German event. Since many enterprises lack the internal IT and software departments to maintain such systems, this software is crucial for developing AI factories outside large cloud providers, noted Huang.
Describing Nvidia as a “full stack end-to-end, complete AI factory platform”, Huang emphasized the importance of software to activate the components in AI factories beyond its CUDA (Computer Unified Device Architecture) offerings.
The proposed fee hasn’t been confirmed by Nvidia, nor the extent to which the license would interact with existing production lines is clear.
