[SMM Computing Power Flash] South China ROG 5090 tax-inclusive price drops to 53,000, down 4.3% WoW

Published: Oct 29, 2026 10:03 (GMT+8)

SMM learned that 15 ROG Edition RTX 5090 units were offered for sale in South China at a tax-inclusive price of 53,000 yuan per unit. Last week, channel distributors in the same region quoted the same model at 49,000 yuan per unit excluding tax and 55,370 yuan including tax. This week's tax-inclusive price fell by 2,370 yuan from that level, a decline of 4.3%. Converted at a 13% tax rate, this week's quotation corresponds to approximately 46,900 yuan per unit excluding tax, down 2,100 yuan from last week's 49,000 yuan. Previously, on October 10, the same series in the same region was quoted at 47,000 yuan excluding tax and 51,000 yuan after tax.

For details, contact SMM Cao Zheng at 021-51595793.

Data Source Statement: Except for publicly available information, all other data are processed by SMM based on publicly available information, market communication, and relying on SMM's internal database model. They are for reference only and do not constitute decision-making recommendations.

For any inquiries or for more information, please contact: lemonzhao@smm.cn
For more information on how to access our research reports, please contact:service.en@smm.cn
Related News
Kexiang to Invest $2B in AI Server PCB Substrate Project in Huizhou, Aiming for 46,000 m² mSAP and 1.2M m² PCB Capacity
Oct 30, 2026 19:27 (GMT+8)
Kexiang to Invest $2B in AI Server PCB Substrate Project in Huizhou, Aiming for 46,000 m² mSAP and 1.2M m² PCB Capacity
Read More
Kexiang to Invest $2B in AI Server PCB Substrate Project in Huizhou, Aiming for 46,000 m² mSAP and 1.2M m² PCB Capacity
Kexiang to Invest $2B in AI Server PCB Substrate Project in Huizhou, Aiming for 46,000 m² mSAP and 1.2M m² PCB Capacity
Kexiang Co., Ltd. (300903.SZ) announced that the company plans to invest in the construction of a high-end PCB substrate project for AI computing power in the Xiangshuihe Industrial Park, Daya Bay, Huizhou, with a total investment of approximately 2 billion yuan. The project is planned to have a construction period of 2 years, and upon completion, it will add capacity of 46,000 m² of mSAP substrates per year and 1.2 million m² of PCB substrates for AI servers per year.
Oct 30, 2026 19:27 (GMT+8)
[SMM Computing Power News] 910B4 relocatable, reduced to 18,000, available for rent in early November
Oct 30, 2026 18:12 (GMT+8)
[SMM Computing Power News] 910B4 relocatable, reduced to 18,000, available for rent in early November
Read More
[SMM Computing Power News] 910B4 relocatable, reduced to 18,000, available for rent in early November
[SMM Computing Power News] 910B4 relocatable, reduced to 18,000, available for rent in early November
SMM has learned that five relocatable 910B4 units have been released to the market, with a monthly rental of 18,000 yuan per unit, available for lease in early November. Compared with recent quotes of 19,000–20,000 yuan for the same form factor, this deal is 1,000 to 2,000 yuan lower, a reduction of 5.3%–10%. Compared with online resources at 14,500–15,000 yuan, the relocatable units carry a premium of 3,000 to 3,500 yuan, or about 21%–24%, notably narrower than the previous price spread of 5,000 yuan, or 33%. This model is mainly intended for inference scenarios.
Oct 30, 2026 18:12 (GMT+8)
Nvidia Unveils DSX POD: Modular AI Factory in a Shipping Container at GTC 2026
Oct 30, 2026 15:26 (GMT+8)
Nvidia Unveils DSX POD: Modular AI Factory in a Shipping Container at GTC 2026
Read More
Nvidia Unveils DSX POD: Modular AI Factory in a Shipping Container at GTC 2026
Nvidia Unveils DSX POD: Modular AI Factory in a Shipping Container at GTC 2026
Nvidia unveiled DSX POD, a modular AI factory reference design, at its GTC conference in Berlin on October 21, 2026. The design integrates compute, networking, rack-level power and liquid cooling into a standard shipping container. Each container carries seven Vera Rubin NVL72 racks with 504 Rubin GPUs in total, plus about 20 kilometers of copper cabling and 17 kilometers of scale-out networking installed and tested at the factory; a unit weighs roughly 28 metric tonnes and packs about 10 million components. Where site power, cooling and network connections are ready, on-site deployment time falls from six months to as few as 14 days. The system runs on coolant up to 45 degrees Celsius with dry cooling, which Nvidia estimates saves about 54 million liters of water per day for a 2-gigawatt deployment. Seventeen partners including Dell, HPE, Lenovo, Bull, Schneider Electric, Vertiv and Armada will begin worldwide deliveries in early 2027, moving inter-rack copper wiring from field construction into factory prefabrication.
Oct 30, 2026 15:26 (GMT+8)