
NVIDIA is set to launch a new export-compliant GPU, a rebranded version of the GeForce RTX 5090, named the “GeForce RTX 5090 D V2.”This development comes as the company aims to meet stringent regulations affecting sales in China.
NVIDIA GeForce RTX 5090 D V2: A Compliant GPU Tailored for Chinese Gamers
The US-China trade landscape has imposed significant challenges for gamers, particularly those with desktop setups. As pricing peaks in the US, Chinese gamers encounter limitations on access to premium graphics cards. This situation has evolved since 2023 and extends into 2025. Acknowledging these complications, NVIDIA is introducing export-compliant GPU models that cater to both the high-performance computing (HPC) and gaming sectors. However, the dynamic nature of export controls means that availability can fluctuate unexpectedly, leaving potential buyers waiting for alternatives.
The saga of the GeForce RTX 5090 D reflects these complexities. NVIDIA initially developed the “GeForce RTX 5090 D”to align with US export guidelines, but as new regulations took effect, it became ineligible for sale in China. This prompted NVIDIA to initiate plans for a second, China-exclusive variant of the RTX 5090.
Recently, insights into this revised model have emerged, clarifying that NVIDIA has shifted from earlier naming conventions. Contrary to prior speculation, the card will not be labeled “GeForce RTX 5090 DD.”Sources such as MEGAsizeGPU confirm that the official name is the GeForce RTX 5090 D V2, raising questions about the company’s naming strategy for these updated offerings.

According to reports from HKEPC, the launch of the NVIDIA GeForce RTX 5090 D V2 is scheduled for August 2025. This means Chinese gamers will have to wait an additional month to access the new flagship model.
The upcoming GPU is expected to utilize the GB202-240 die, a downgrade from the GB202-250 in the 5090 D and the GB202-300 in the standard 5090. This iteration will also include a revised printed circuit board (PCB) design, the PG145 WeU 40, which accounts for changes in memory and GPU layout.
【沒有叫作 DD ❌】最新的消息傳出,NVIDIA 最終不會採用 GeForce RTX 5090 DD 型號作為命名方式,而是沿用 GeForce RTX 5090 D 型號,並加上「v2」作為識別。目前計劃於 2025 年 8 月上市。雖然記憶體容量減少,但售價預計不會有太大變化。 由於美國修改了對中國的晶片出口限制,GeForce RTX 5090 D… pic.twitter.com/t0okwVCMba
— HKEPC (@hkepcmedia) July 2, 2025
Notably, the GeForce RTX 5090 D V2 will retain the same number of cores—21, 760—but will feature a reduced memory capacity of 24 GB GDDR7, compared to the previous 32 GB configuration. The memory bus will remain 384 bits, and while the total board power (TBP) is expected to remain at 575W, further details on clock speeds are still pending. Despite the cut in VRAM, the gaming performance is projected to be similar, barring any bottlenecks created by reduced memory capacity.
The consistent pricing in China, despite the decrease in VRAM, is disappointing news for local consumers who previously enjoyed a larger memory capacity. This change may also impact entry-level AI systems and businesses relying on significant VRAM for tasks involving large language models (LLMs).
Specifications of the NVIDIA GeForce RTX 5090 D V2:
GPU Name | RTX 5090 D V2 | RTX 5090 D | RTX 5090 |
---|---|---|---|
Process Node | TSMC 4N | TSMC 4N | TSMC 4N |
GPU WeU | GB202-240 | GB202-250 | GB202-300 |
GPU Cores | 21, 760 | 21, 760 | 21, 760 |
Core Clock (Boost) | TBD | 2407 MHz | 2407 MHz |
VRAM Capacity | 24 GB GDDR7 | 32 GB GDDR7 | 32 GB GDDR7 |
Memory Bus | 384-bit | 512-bit | 512-bit |
Bandwidth | 1.79 TB/s | 1.79 TB/s | 1.79 TB/s |
TBP | 575W | 575W | 575W |
For further details, you can refer to the original source here.
Leave a Reply