The date for Nvidia's next major GPU dump is set. It is January 3 next year and will be the keynote address at the GeForce Special Event at the CES Technology Jamboree in Las Vegas (opens in new tab).

Nvidia only describes the event as detailing "the latest GeForce technology, products, and software," but nothing more specific than that. However, we do expect two major announcements at the event.

On the desktop side, it is very likely that we will finally see what Nvidia did with the cancelled RTX 4080 12GB model (opens in new tab). It will likely be rebranded to RTX 4070 Ti and announced on the 3rd at its CES event.

Given that the RTX 4080 12GB board was already in mass production when Nvidia changed its mind, we don't expect much in the way of specification changes in the transition to the RTX 4070 Ti brand.

That means 7,680 shaders (CUDA cores in Nvidia terminology), 240 texture units, 80 ROPs, and 12GB of memory running on a 196-bit bus. Roughly speaking, these numbers are about 40-45% of the flagship GPU, the RTX 4090 (open in new tab). it is possible that Nvidia will tweak some of these specs, supposedly the RTX 4070 Ti, through video BIOS updates, but expect dramatic deviations Not necessarily.

Instead, the big unknown is the price: the RTX 4080 12GB is priced at $899, and the assumption is that the RTX 4070 Ti will be much cheaper. But we will know how much cheaper on January 3.

The other big announcement at the CES event is expected to be a mobile version of Nvidia's RTX 40 series GPUs, known as Ada Lovelace. At a minimum, RTX 4080 and 4070 branded mobile GPUs will be announced, with mobile RTX 4090 and 4060 mobile chips likely to be added.

Based on the best information currently available, the mobile lineup would look like this The top-of-the-line RTX 4090 will use the same AD103 GPU as the desktop 4080 card, with 9,728 shaders, 304 texture units, and 112 ROPs, all running on a 256-bit memory bus. The clock speed will be approximately 1.4 GHz, significantly lower than the desktop 4080.

The RTX 4080 and 4080 Ti mobile step down to the AD104 GPU seen briefly on the cancelled 4080 12GB desktop board, with the Ti model offering nearly the same hardware features as the 4080 12GB desktop but with lower clock speeds. Subtract roughly 15% from the number of shaders, textures, and ROPs and you get a non-Ti model. The smaller AD104 chip should clock a bit higher for mobile implementations and may reach 2 GHz for higher performance implementations. However, the usual low-power Max-Q variants for thin-and-light will inevitably be much slower.

As for the RTX 4070 mobile, it will likely be based on the yet-to-be-announced AD106 chip. Specs will probably be around 4,608 shaders, 144 textures, 48 ROPs, plus a 128-bit memory bus. Finally, the RTX 4060 may use another unannounced chip, the AD107, with even more reduced specs, likely around 2,560 to 3,072 shaders, 80 to 96 texture units, and 32 ROPs. Fortunately, the AD107 is currently expected to stick to a 128-bit memory bus, rather than stepping down further to a 96-bit bus.

That's all we have to say about what to expect on January 3. Let us know below if this sounds exciting or just frustrating for the latest GPU announcements.