NVIDIA has teased the RTX 3090 Ti in a CES presentation. The graphics card is expected to make use of the entirety of the GA102 GPU's 10752 CUDA cores. NVIDIA has added 24GB of GDDR6X memory clocked at 21 Gbps Ti. With this memory connected to a 384-bit memory bus, the highest theoretical bandwidth is 1 TB/s is almost 8% percent faster than the regular GeForce RTX 3090.

The Founders Edition RTX 3090 Ti has a triple-slot configuration that is pretty much identical to that of the original model. Several customized variants with a recommended 1000W power supply have already been reported; this indicates that the product line is not meant for systems with mid-range specs. The founder card will be powered by a single 16-pin power connector that will be capable of delivering up to 450W of output power. The card is rumored to get a base clock of 1560 MHz and a boost clock of 1860 MHz, which is faster than the regular model. This represents a 12 percent and a 10 percent boost over the non-Ti RTX 3090, respectively, over its predecessor. In terms of single-precision compute capacity, the card has a maximum of 40 TFLOPS of processing power.

The company has not yet confirmed the model's suggested retail price (MSRP). NVIDIA did not share specs yet, so the below is unofficial just yet.