NVIDIA has recently announced the new GeForce RTX 5090D v2, specifically designed for the Chinese market. This update to the “Dragon” edition is a response to the evolving U.S. export restrictions on high-performance graphics cards. While NVIDIA has not confirmed the implications of the U.S. Department of Commerce’s latest regulations on the RTX 5090D, it is noteworthy that this version was initially crafted to comply with these restrictions.
Based on preliminary reviews, the RTX 5090D performs almost identically to its predecessors in both gaming scenarios and synthetic benchmarks. Gamers have reported that overclocking capabilities remain robust, making it challenging to discern any performance differences between models. However, the availability of the RTX 5090D has been dwindling in the Chinese market, with most remaining options tied to prebuilt PCs that are increasingly hard to find.
The anticipated RTX 5090D v2, previously referred to as the RTX 5090DD, aims to address the scarcity of the original model. According to reliable leaker MEGAsizeGPU, the new version will be distinctly recognizable when compared to its predecessor. Despite the less-than-ideal naming convention, which raises questions about having three cards in the 5090 series, NVIDIA is expected to clarify this by maintaining the GeForce RTX 5090 D model label and appending “v2” for differentiation.
The official launch of the RTX 5090D v2 is projected for August 2025. Although this version will feature a reduced memory capacity, pricing is not anticipated to fluctuate significantly. HKEPC has corroborated this information, indicating that the new model will come equipped with 24GB of GDDR7 memory and a 384-bit memory bus, while retaining the full GPU core count of 21,760 and a 575W TDP.
For gamers, the memory reduction might not pose significant challenges, as very few current games demand more than 24GB of memory. However, this brings up an important question regarding the rationale behind branding this model as a 5090. Some industry experts suggest that NVIDIA could have opted to launch it as the RTX 5080 Ti, a move that would have aligned better with market expectations and provided the company with greater flexibility in global sales strategies.
Ultimately, it will be up to consumers in China and surrounding markets to make purchasing decisions based on the value they perceive in the RTX 5090D v2. Currently, the highest-end model available in the Blackwell lineup is the RTX 5080, with a SUPER variant expected to debut several months down the line.
The upcoming NVIDIA GeForce RTX 50 Series is generating significant buzz in the gaming community. Here’s a quick overview of the rumored models:
RTX 5090: 21,760 CUDA Cores, 32GB G7, 512-bit, 575W - ReleasedRTX 5090 D: 21,760 CUDA Cores, 32GB G7, 512-bit, 575W - ReleasedRTX 5090 D v2: 21,760 CUDA Cores, 24GB G7, 384-bit, 575W - Coming August 2025RTX 5080 SUPER: 10,752 CUDA Cores, 24GB G7, 256-bit, 415W - RumoredRTX 5080: 10,752 CUDA Cores, 16GB G7, 256-bit, 360W - ReleasedRTX 5070 Ti SUPER: 8,960 CUDA Cores, 24GB G7, 192-bit, 350W - RumoredRTX 5070 Ti: 8,960 CUDA Cores, 16GB G7, 256-bit, 300W - ReleasedRTX 5070 SUPER: 6,400 CUDA Cores, 18GB G7, 192-bit, 275W - RumoredRTX 5070: 6,144 CUDA Cores, 12GB G7, 192-bit, 250W - ReleasedRTX 5060 Ti: 4,608 CUDA Cores, 8/16GB G7, 128-bit, 180W - ReleasedRTX 5060: 3,840 CUDA Cores, 8GB G7, 128-bit, 145W - ReleasedRTX 5050: 2,560 CUDA Cores, 8GB G6, 128-bit, 135W - ReleasedAs we look forward to the launch of the RTX 5090D v2 and the broader RTX 50 Series, gamers and tech enthusiasts alike are eager to see how these advancements will shape the landscape of high-performance graphics cards.
