According to well a known leaker NVIDIA is reportedly no longer planning on releasing a GeForce RTX 4090 Ti and instead will focus on the rest of its ADA lineup. For some time there have been rumors about a new flagship graphics card to replace the GeForce RTX 4090 and despite images of a massive heatsink that began surfacing in January, it now seems NVIDIA is choosing to finish rolling out the rest of the RTX 40 series stack while looking ahead to its next generation of GPUs. It’s rumored that NVIDIA will utilize some AD103 and AD106 GPUs to produce more RTX 4070 and RTX 4060 variants. This seems probable given that NVIDIA has typically released multiple versions of x60 and x70 tiers in past generations as each series launch nears its end.
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
As VideoCardz points out, there is no real need for team green to offer a new flagship right now since competition for the current one is virtually nonexistent. While AMD’s recent RDNA 3-based GPUs have made strides in competing on a rasterization level with cards below the RTX 4090 they still haven’t posed a real threat to it. There are rumors that AMD might integrate its 3D V-cache technology with RNDA3 for a refresh but until that happens and it’s proven to be a threat to RTX 4090 sales, NVIDIA would not have much to gain in releasing an RTX 4090 Ti. Meanwhile, another rumor has surfaced that the next-gen flagship card could feature a 512-bit memory bus.
Combined with multiple sources, I confirm the gaming flagship of Ada-next will have a 512-bit memory interface.
— kopite7kimi (@kopite7kimi) July 27, 2023
Next-gen Memory Upgrades
If it’s true that NVIDIA is reportedly no longer interested in the RTX 4090 Ti and that the next-gen king of the mountain will use a 512-bit bus, it will be the first time in over a decade since the GPU manufacturer has used such a memory architecture for a consumer card. The last time it did so was with the GeForce GTX 280 (per VideCardz). It’s still too early to tell what type of memory will be paired with the next-gen GPUs as NVIDIA has now used GDDR6X for its top-of-the-stack cards in both the RTX 30 and RTX 40 series.
GDDR7 is on the horizon with both Samsung and Micron having said they plan to release their respective offerings in the first half of 2024. GDDR7 is said to potentially offer up to 2x the memory bandwidth as that found on the RTX 4090 but it doesn’t mean that NVIDIA would use the fastest configuration at launch since it could save that for another flagship release or refresh down the road. NVIDIA is expected to launch its successor to the RTX 40 series sometime in 2025. Meanwhile, September 2023 will be the one-year anniversary of the launch of the RTX 4090 so there’s still plenty of time for a refresh if the need arises. The release of a Titan Ada remains a possibility as well.