Nvidia has reportedly informed its largest customers that server prices containing its AI chips will increase by more than 15% in many cases. The warning comes as memory costs continue to surge, impacting the broader AI infrastructure market. The company's latest-generation chips, including the Vera Rubin and Grace Blackwell models, are expected to be affected by these hikes. The price changes are anticipated to take effect on systems shipped early next year, according to available reports.

The price increases are attributed to the rising costs of high-bandwidth memory (HBM) and other critical components used in AI servers. These components are essential for the performance of Nvidia's AI chips, which power a significant portion of the global AI infrastructure. The company's second-quarter results are expected to be released on August 26, and the financial impact of these rising costs may be reflected in those reports. Industry analysts suggest that the price hikes could influence the purchasing decisions of major tech firms relying on Nvidia's hardware.

The 15% increase is expected to affect a range of AI server models, particularly those incorporating the latest Vera Rubin and Grace Blackwell chips. These systems are designed for high-performance computing and are in demand by companies developing large-scale AI applications. The cost of HBM, which is crucial for these systems, has been increasing due to supply chain constraints and rising demand. This trend is likely to continue into 2027, with the first shipments of the affected systems expected in early next year.

The price hikes could lead to increased costs for businesses relying on Nvidia's AI servers, potentially affecting their budgets and project timelines. Companies may seek alternatives or negotiate better terms with Nvidia to mitigate the impact. The move may also influence the broader market, as competitors could adjust their pricing strategies in response. Additionally, the increased costs may prompt discussions around vendor lock-in and the long-term governance of AI infrastructure, as firms evaluate their reliance on a single supplier.

The situation remains fluid, with more details expected as Nvidia releases its second-quarter results on August 26. The company has not yet commented on the reports, and Reuters has not confirmed the claims. As the AI server market evolves, the implications of these price increases will likely become clearer in the coming months.