Kopite7kimi has been rather silent regarding the recent rumors about the GeForce RTX 40 series, mainly because a significant portion of the RTX 40 lineup has already been launched. Nonetheless, with the RTX 4090 announcement’s one-year anniversary approaching in September, people are wondering if NVIDIA has any intentions of a mid-series refresh.
Despite NVIDIA already launching two Ti models (4070/4060), there are no signs of an RTX 4090Ti or RTX 4080Ti, which would typically be the logical choices for a refresh or specification upgrade. Although the RTX 4090 non-Ti doesn’t utilize the entire AD102 silicon, one could reasonably expect a higher-end model with more cores and possibly faster GDDR6X or GDDR7 memory. However, it seems that such plans are not currently being considered, at least not for the flagship model.
I'm afraid there won't be RTX 4090 Ti anymore. Some low-grade AD103 and AD106 chips will be another versions of RTX 4070 and 4060.
— kopite7kimi (@kopite7kimi) July 27, 2023
Kopite7kimi, a renowned and dependable source on upcoming NVIDIA hardware, has revealed that there are no intentions to release an RTX 4090 Ti. Furthermore, it’s been mentioned that NVIDIA will introduce new AD103/AD106 models for the RTX 4070/4060. However, these models aren’t entirely fresh SKUs; they seem to be geared towards optimizing chip sales and might primarily target the Chinese market.
At present, NVIDIA doesn’t find it necessary to release the RTX 4090 Ti since the RTX 4090 already holds the position as the most powerful GPU available, and there’s limited competition from AMD in this high-performance segment. While there have been rumors about potential refreshes or the addition of 3D V-Cache to AMD’s RDNA3 series, these speculations haven’t raised any significant concerns for NVIDIA. Unless there’s a notable market shift, the introduction of an RTX 4090 Ti is unlikely.
Kopite mentioned that a possible flagship GPU from the next generation could feature a 512-bit memory bus, which would mark a significant departure in NVIDIA’s gaming lineup, as this configuration hasn’t been used in a long time, like with the GTX 280 or certain dual-GPU models. If NVIDIA were to adopt GDDR7 memory with a maximum speed of 32 Gbps for this potential next-gen GPU, it would result in an impressive 2TB/s bandwidth, which is double the bandwidth offered by the RTX 4090.
Combined with multiple sources, I confirm the gaming flagship of Ada-next will have a 512-bit memory interface.
— kopite7kimi (@kopite7kimi) July 27, 2023
Nevertheless, it is improbable that NVIDIA will employ the fastest G7 memory right from the initial release, and the mere possibility of the GPU supporting a 512-bit bus does not guarantee that this configuration will be enabled for every variant. Undoubtedly, there are more questions than answers at this stage.
The next-generation consumer architecture from NVIDIA is not anticipated to be revealed until 2025. Considering the company’s usual two-year cadence for gaming architecture releases, there might be a chance of a mid-generation update in some capacity to fill the gap. However, specific details regarding this possibility remain uncertain at this time.