Nvidia to Charge AI Startups by Usage, Aiming for $1B Recurring Revenue
NVDA•Nvidia will roll out usage-based GPU pricing for generative AI startups via cloud partnerships, charging per inference and training hour to create a recurring revenue stream. The program targets hundreds of early-stage AI firms and aims to add $500 million to $1 billion in annual revenue by fiscal 2027.
1. New Pricing Strategy for AI Startups
Nvidia will introduce usage-based GPU pricing for generative AI startups through partnerships with leading cloud platforms, replacing traditional hardware discount models. Clients will pay per inference and per training-hour fees, aligning costs directly with AI workload demands.
2. Financial Impact and Forecast
Executives project the new service could contribute an additional $500 million to $1 billion in annual recurring revenue by fiscal 2027, diversifying Nvidia’s revenue mix beyond one-time hardware sales. This predictable income stream is expected to smooth quarterly fluctuations tied to product cycles.
3. Market and Competitive Context
By locking in early-stage AI firms, Nvidia strengthens its ecosystem and raises barriers to entry for rivals like AMD and Intel. The move also positions Nvidia as a platform provider, potentially increasing long-term customer stickiness and market share.
