[SMM Computing Power Express] The industry standard intermediary fee for the 4090 series models defaults to an additional 300 yuan per unit per month.

Published: Aug 11, 2026 18:27
SMM has learned that a pricing convention has also been established for intermediary channels in computing power leasing for entry-level models—an intermediary fee of about 300 yuan per machine per month, which is generally applicable to the 4090 series (e.g., RTX 4090). As a channel facilitation cost, it is added on top of the base rental. SMM believes that the intermediary fee for entry-level models is significantly lower than that for high-end models, reflecting ample supply of low and mid-end computing power and sufficient channel competition.

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
[SMM Computing Power News] The default brokerage fee for the H series models in the industry is an additional 1,000 yuan per unit per month.
2 hours ago
[SMM Computing Power News] The default brokerage fee for the H series models in the industry is an additional 1,000 yuan per unit per month.
Read More
[SMM Computing Power News] The default brokerage fee for the H series models in the industry is an additional 1,000 yuan per unit per month.
[SMM Computing Power News] The default brokerage fee for the H series models in the industry is an additional 1,000 yuan per unit per month.
SMM has learned that a default industry practice has gradually emerged in the intermediary channel for computing power leasing—a monthly intermediary fee of approximately 1,000 yuan per machine, a rule that is generally applied to H-series models (e.g., H100/H200). The intermediary fee, as a channel matchmaking cost, is added on top of the base rent. SMM believes that the solidification of intermediary fees for high-end H-series models will raise lessees' overall computing costs and also reflects the channel scarcity premium for top-tier models.
2 hours ago
[SMM Computing Power Daily] Relocatable H100 Long-Term Contracts Released, A100 80G Whole Machine Channel Quotation 1.34 Million
2 hours ago
[SMM Computing Power Daily] Relocatable H100 Long-Term Contracts Released, A100 80G Whole Machine Channel Quotation 1.34 Million
Read More
[SMM Computing Power Daily] Relocatable H100 Long-Term Contracts Released, A100 80G Whole Machine Channel Quotation 1.34 Million
[SMM Computing Power Daily] Relocatable H100 Long-Term Contracts Released, A100 80G Whole Machine Channel Quotation 1.34 Million
Abstract: August 11, a cluster released 8 relocatable H100 servers, with a monthly rent of 72,000 yuan per unit and a three-year minimum lease; A100 80G server channel tax-inclusive quotation of 1.34 million yuan per unit, valid for only 3 days.
2 hours ago
[SMM Computing Power Express] Two IB-networked A100 80G nodes for rent in South China, at 35,000 yuan per month with a one-year minimum lease.
Aug 10, 2026 17:06
[SMM Computing Power Express] Two IB-networked A100 80G nodes for rent in South China, at 35,000 yuan per month with a one-year minimum lease.
Read More
[SMM Computing Power Express] Two IB-networked A100 80G nodes for rent in South China, at 35,000 yuan per month with a one-year minimum lease.
[SMM Computing Power Express] Two IB-networked A100 80G nodes for rent in South China, at 35,000 yuan per month with a one-year minimum lease.
SMM learned that a computing cluster in South China released the rental resources of 2 IB-networked A100 80G servers, with a monthly rent of 35,000 yuan per unit and a minimum one-year lease (long-term agreement). These resources are aimed at high-bandwidth interconnection needs such as large model training, and IB networking can support low-latency communication for multi-machine distributed training. SMM believes that the release of a small batch of high-end GPUs in the form of a long-term agreement reflects the resilience of training computing power demand in South China.
Aug 10, 2026 17:06