In recent discussions, attention has turned to memory shortages and the resulting changes in the graphics accelerator market. WCCF has sought to clarify many questions. Various sources close to Nvidia and graphics card manufacturers have been interviewed. It turns out that the situation is not as dire as previously reported: indeed, there is a memory shortage (impacting all types of memory), but Nvidia is still producing the entire GeForce RTX 50 lineup.
Nvidia commented to Wccftech, stating that the demand for the GeForce RTX 50 remains high, despite rising prices, and that all models in the lineup continue to ship to partners. According to the company, especially strong sales occurred in the fourth quarter and the holiday sales season (“Black Friday” and other promotions).
The Memory Market’s Broad Impact
Regarding the memory market, the issue is not the shortage of any single type of memory. Constraints have affected the entire DRAM market: DDR memory is massively directed to data centers and PCs; HBM is almost entirely “absorbed” by the AI and HPC sectors; GDDR6 and GDDR7 for graphics cards are produced but with lower priority. Key memory manufacturers-Samsung, SK Hynix, and Micron-have shifted their capacities toward the AI segment, where orders are scheduled for years ahead. As a result, GDDR memory is becoming more expensive, directly affecting the cost of graphics cards-both new RTX 50 and older models on GDDR6.
According to sources, Nvidia has held off on price increases longer than other manufacturers and has not passed the higher memory costs directly onto partners and end consumers. Furthermore, the actual increase in memory prices was lower than some online leaks claimed. Nevertheless, amid high demand and limited supply volumes, the final effect still results in shortages and inflated retail prices.
Dispelling Rumors
One of the loudest rumors was that Nvidia allegedly stopped supplying AIC partners with “GPU + memory” kits, forcing them to purchase DRAM independently. Industry sources confirm: this is untrue. Nvidia continues to supply partners with complete chipsets and memory without changes in the working scheme.
Separately, rumors about the “death” of the RTX 5070 Ti, RTX 5080, RTX 5090, and other models have been debunked. According to Wccftech today: all RTX 50 models remain in production; configurations with 8, 12, 16, and 32 GB of video memory continue to be delivered; no model has been officially marked as EOL (end of life); no VRAM quantity has priority or, conversely, reduced status. As noted by sources, Nvidia “supports all market segments,” and disruptions in card availability are solely related to the overall memory market situation.
