The Nvidia GeForce RTX 4070 Ti has had a wild story, as it was supposed to be launched as a variant of the RTX 4080 with graphics memory reduced to 12 gigabytes, with Nvidia trying to offer customers a less powerful graphics chip than the regular RTX 4080 with 16 GB of memory.
The GeForce RTX 4070 Ti comes again in various board partner variants
The feedback and confusion surrounding this branding jumble was negative or large, which is why the RTX 4080 12GB was “unlaunched” and initially disappeared back into the manufacturer’s drawer. Now the virtually unchanged card returns with a more appropriate name, with a slight price drop as mentioned.
Board partners en masse at Bo(a)rd
Technically, nothing changes, so the number of graphics cores is halved compared to the RTX 4090, as before. Nvidia uses the “Ada Lovelace” architecture and specifies a power consumption of 285 watts, with power being supplied via two 8-pin connectors. As always, the memory is GDDR6X RAM, which is connected with a 192-bit wide interface. The RTX 4070 Ti has 7680 CUDA cores and works with a base clock of 2.31 gigahertz and a boost clock of 2.61 GHz. Among other things, Nvidia advertises that the new GPU can achieve significantly higher frame rates than an RTX 3090 Ti from the previous generation thanks to DLSS3, but the manufacturer does not mention that the GPU used for comparison does not offer any DLSS3 support.
Ultimately, the performance should be around 80 to 100 percent of the RTX 3090 Ti, i.e. deliver more than decent results for most games at 1440p resolution. As can be seen from the various leaks over the past few days, the RTX 4070 Ti will be available from tomorrow, with variants from Nvidia’s board partners such as ASUS, Colorful, Gainward, Galaxy, Gigabyte, Inno3D, KFA2, MSI, Palit, PNY and Zotac are expected. There is no Founders Edition from Nvidia this time.
See also:
