The rumor has been circulating for some time and a new rumor seems to confirm AMD’s intention to abandon the high-end of the market with the next generation of graphics cards.
According to what emerged, in fact, the Radeon RX 8000 based on RDNA 4 architecture could use the old 18 Gbps GDDR6 memories instead of the fastest available solutions. Here are all the details.
No GDDR7 for AMD?
The rumor comes from yet another tweet by Kepler, an insider with a long and positive history in the field of accurate predictions. According to the leaker, the next generation of AMD GPUs it will be accompanied by 18 Gbps GDDR6 memories for the entire rangetherefore without any transition to the new and faster Samsung GDDR6W or to the GDDR7 standard.
If this decision were confirmed it would underline, once again, AMD’s intentions for the new cards based on RDNA 4 architecture: the strategy of the red house seems in fact focus everything on the price/performance ratioattacking the market on the low and medium range, without touching the high segment, dominated by NVIDIA’s competitors.
The rumors that have been circulating so far speak of two chips, Navi 48 and Navi 44, the first with similar performance to the current Radeon RX 7900 and the second which will replace the Radeon RX 7800 XT. In short, maintaining GDDR6 memories could help AMD control prices.
Cost issue?
AMD’s choice, in addition to economic issues, could directly bring into play the RDNA 4 architecture: the 18 Gbps GDDR6 memories would in fact be enough to provide a good level of bandwidth and could be an ideal solution for the new Infinity Cache , capable of eliminating any “bottlenecks”.
On the opposite front, NVIDIA seems to be focusing on the faster GDDR7, ready to launch the RTX 50 range based on Blackwell architecture
