Today, NVIDIA officially released the GeForce RTX 4090 DGPU for gamers, aiming to bring flagship gaming performance to Chinese players. Its performance is very similar to the original RTX 4090, but it reduces the AI computing power. The launch of this graphics card is very special because the standard version of RTX 4090 has been restricted in markets such as China due to relevant regulations in the United States.
Based on the spec information we have, the NVIDIA GeForce RTX 4090D is expected to use a scaled-down version of the AdaLovelace GPU on the RTX 4090. Dubbed the AD102-250-A1 SKU, this variant offers a lower core count to stay within the TPP (Total Processing Performance) requirements set by the US government and make it compliant with the latest export rules.
NVIDIA GeForceRTX4090DGPU has 14592 CUDA cores, a 2520MHz boost clock, 24GBGDDR6X memory with a clock frequency of 21Gbps, a 384-bit bus interface and a 425W TGP. All graphics cards will get three display outputs, including three DP1.4a and one HDMI2.1 port. None of the custom models support overclocking. Almost all major Chinese manufacturers such as MSI, GALAX, ZOTAC, Inno3D, etc. have released the same custom design as the RTX4090 model:
The U.S. Department of Commerce and Secretary Gina Raimondo have recently taken a carrot and stick approach to China's AI chip restrictions, saying that Nvidia should sell chips to China because these chips are used in commercial applications such as games. In terms of pricing, the upcoming gaming GPU will be priced similarly to the previous RTX 4090 model, with a suggested retail price of 12,999 yuan, which is equivalent to nearly $1,800.