Nvidia has reportedly begun increasing the cost of its GPU kits, including both GDDR6 and GDDR7 memory, for its add-in board (AIB) partners. This development follows a period of surprising price stability in the graphics card market and could signal future price hikes for consumers.
The price adjustments are said to affect a range of GeForce products, including potential RTX 50-series cards that will utilize GDDR7 memory and some RTX 40-series cards still using GDDR6. Information obtained by Benchlife suggests that pricing for the RTX 5090 and its China variant was adjusted in May, with this latest move impacting other GeForce models. The RTX 5050 is also cited as a potential candidate for price increases.
These increases mean that AIB partners, the companies that manufacture and sell non-Founders Edition graphics cards, will now face higher costs for the integrated GPU and memory kits they purchase from Nvidia. This situation is largely attributed to the ongoing global memory shortage, exacerbated by the high demand driven by the artificial intelligence boom.
Sources also indicate that the elevated cost of 3GB GDDR7 memory may have led Nvidia to put its 50-series Super cards on hold. These were anticipated for release later this year or early next year, but the prohibitive cost of the required memory has apparently put these plans in limbo.
While graphics card prices have remained relatively stable over the past few months, this reported price increase from Nvidia could alter that trend. The exact timeline for when these increased costs might be reflected in consumer pricing remains uncertain. AIB partners may have existing stock to sell through or could absorb some of the cost to maintain competitive prices. However, the slim profit margins often experienced by these partners, a factor previously cited by companies like EVGA for exiting the GPU market, could make absorbing these costs difficult.
Companies further down the supply chain, such as AIB partners, often operate with lean inventories to avoid the financial risks associated with holding excess stock. This approach, common in 'just-in-time' manufacturing and sales, means that price increases from suppliers are frequently passed on to customers. Framework, for example, recently announced price adjustments due to a supplier significantly raising memory costs beyond their predictions.
Nvidia, with its larger contracts and greater financial resources, is likely better positioned to manage these rising costs. However, if the reports of increased GPU kit prices are accurate, even Nvidia and its partners face challenges in the current market. The trend of increasing memory costs due to AI demand poses a significant headwind for the PC hardware market.