AI companies are increasingly purchasing standard retail versions of the NVIDIA GeForce RTX 5090 graphics card for use in AI server clusters. Photos from HKEPC show bulk quantities of boxed consumer-grade cards being unpacked at server assembly sites, rather than professional-grade RTX PRO 6000 or data center accelerators.
The surge in demand from AI firms has significantly strained the supply of the flagship GPU. While the official launch price for the RTX 5090 is $1,999, retail prices in some channels have reportedly climbed above $5,000. This trend mirrors past market disruptions caused by cryptocurrency mining operations.
Industry analysts suggest that NVIDIA is unlikely to implement restrictions on AI capabilities for the standard RTX 5090, despite having previously developed the RTX 5090D with reduced AI performance for specific export markets. AI business remains a primary growth engine for the company, making it hesitant to limit the utility of its consumer hardware.
The shift in usage patterns is transforming the RTX 5090 from a product aimed at gaming enthusiasts and content creators into a component of AI infrastructure. As AI companies continue to prioritize these cards for training and inference tasks, retail availability for individual consumers remains severely constrained.
