We may earn a commission when you buy through links in our articles. Learn more.

Nvidia’s RTX 4070 Ti Super cards are reportedly getting 4090 GPUs

Inno3D, MSI, and Colorful are all apparently making new GeForce RTX 4070 Ti Super cards based on the AD102 chip used in the RTX 4090.

MSI Nvidia GeForce RTX 4070 Ti Super with 4090 Founders Edition graphics card

Several Nvidia GeForce RTX 4070 Ti Super cards are reportedly being made with a different GPU from previous cards, with graphics card manufacturers apparently replacing the standard RTX 4070 Ti chip with the larger chip used in the flagship RTX 4090. Before you get excited, though, this sadly doesn’t mean these new RTX 4070 Ti cards are going to run at RTX 4090 speeds.

Released at the start of 2024, the RTX 4070 Ti Super updated the previous RTX 4070 Ti in several key ways. Nvidia increased the VRAM from 12GB to 16GB, widened the memory interface, and added 768 more CUDA cores as well, making it a more powerful gaming graphics card in every way. This wasn’t enough to propel the GPU onto our best graphics card guide, but it was definitely an improvement.

However, another change between the two GPUs was Nvidia moving from the AD104 chip on the original RTX 4070 Ti to the larger AD103 chip on the 4070 Ti Super. AD103 is also used in the RTX 4080 Super, with Nvidia simply disabling some of the parts on the chip for the RTX 4070 Ti Super. This time, however, it looks as though several graphics card makers are using the mighty AD102 GPU, as used in the RTX 4090, in their new RTX 4070 Ti Super cards.

One of these companies is Colorful, which Videocardz reports has confirmed the existence of an RTX 4070 Ti Super V2 card based on the AD102 GPU. Similarly, a recent post by tech leaker wxnod drew attention to a specs sheet for a new MSI RTX 4070 Ti Super 16G Ventus 3X card, cryptically asking “AD102-225-A1?” in the post, before posting a GPU-Z screenshot of another Inno3D RTX 4070 Super Ti card that the app couldn’t recognize, other than that it was an A1 revision, and posing the same question.

None of this has been confirmed by Nvidia, but it wouldn’t be the first time the company has swapped chips between different GPU models. One example is the GeForce RTX 4070, which is available in versions based on both the AD103 and AD104 chips. In all these cases, the alternative GPUs have the same core specification in terms of CUDA cores, RT cores, clock speeds, and so on, but there are other differences between them when it comes to PCBs, coolers, and power output.

AD102 is a massive chip, with a die size of 609 mm², compared with just 379 mm² for an AD103 die, and while Nvidia can disable enough parts on it to make it have the same core spec, it still won’t be exactly the same chip. One bonus for graphics card makers, though, is that they can use existing RTX 4090 PCB and cooler designs for their new RTX 4070 Ti Super cards, and just drop the new GPUs in place.

Why would Nvidia make a new RTX 4070 Ti Super chip? One possible reason is that, during the chip-making process, some chips will make the grade for the RTX 4090, but others won’t. However, some of the latter may well still function as lower-end chips with some of the parts disabled; if Nvidia has a stack of RTX 4090 GPUs that aren’t up to scratch, then this is a good opportunity to sell them off as RTX 4070 Ti Super chips.

In previous years, modders have even found ways to re-enable some of these disabled parts on high-end GPUs with a BIOS flash, and effectively get a high-end GPU for a cheaper price if it works, though this comes with a fair amount of risk.

If you’re looking to upgrade your graphics card, our current favorite right now is the GeForce RTX 4070 Super, which offers great all-round performance for a sensible price.