Nvidia
It’s still not what most people would consider “affordable,” but if you want to get into Nvidia’s GeForce RTX 4000 series for less than $1,000, you’ll have a chance in a couple of days. The company’s GeForce RTX 4070 Ti will launch on January 5 starting at $799, $400 less than the MSRP of the RTX 4080 and $100 less than the originally planned 12GB version of the 4080.
As regulatory filings previously indicated, the RTX 4070 Ti is identical to the canceled 12GB version of the RTX 4080, which Nvidia “unlaunched” via a terse press release after widespread complaints about its branding and positioning. The 4070 Ti includes 12GB of GDDR6X memory on a 192-bit memory bus, plus 7,680 of Nvidia’s CUDA cores, and it can consume up to 285 W of power (though Nvidia says average power draw when gaming is closer to 226 W).
Comparing spec sheets, the 4070 Ti on paper is a little less than half of an RTX 4090, which includes 24GB of RAM on a 384-bit memory bus and 16,384 CUDA cores. It’s a bit closer to the RTX 4080, which has 16GB of memory on a 256-bit bus and 9,728 CUDA cores.
Like the RTX 4090 and 4080, the RTX 4070 Ti is based on Nvidia’s Lovelace architecture and supports Nvidia’s new DLSS 3 features for AI-accelerated resolution and frame rate boosting, plus encoding support for the AV1 video codec. Nvidia’s press release positions the card as “a great max settings 1440p gaming GPU,” though if (as Nvidia claims) the card is faster than an RTX 3090 Ti, it should be a decent 4K performer, too, especially with DLSS enabled.
Changing the 12GB RTX 4080’s branding midstream pushed the launch back from November to January, as Nvidia and its partners needed to trash and redo previously manufactured packaging and other materials. While we at least have a $100 price drop to show for the extra wait, Nvidia’s other 4000-series GPUs have been difficult-to-impossible to find at their officially announced retail prices, even as higher-priced versions from Nvidia’s partners have sat on shelves.