[SMM Computing Power Midday Review] H800 highest price in Beijing-Tianjin-Hebei rises, and scattered resource subletting intensifies supply tightness.

Published: Jul 31, 2026 11:54

Save

The highest monthly rent for H800 in the Beijing-Tianjin-Hebei region rose to 75,500 yuan (up 3.42%), and the highest per-card-hour rate rose to 13.11 yuan (up 3.39%), reflecting tight spot supply of Nvidia. Scattered resources are bulk-rented by platforms/operators and then subleased for profit, driving up listing prices; domestic computing power still needs to build market confidence as it enters the market, and is unlikely to replace Nvidia's high-end in the short term.

[Market Overview] This morning, the highest price in the H800 Beijing-Tianjin-Hebei region saw a notable rise. The SMM-H800-80G-Complete Unit-Beijing-Tianjin-Hebei-Monthly Subscription highest monthly rent increased from 73,000 yuan to 75,500 yuan, equivalent to a highest card-hour rate of 13.11 yuan. The uplift in high-end prices reflected tight spot supply of NVIDIA high-end computing power in the region, while demand remained elevated.

[H800 Beijing-Tianjin-Hebei: Highest Price Rose, Price-Ceiling Anchor Lifted] The SMM-H800-80G-Complete Unit-Beijing-Tianjin-Hebei-Monthly Subscription highest monthly rent increased from 73,000 yuan to 75,500 yuan, up 2,500 yuan, a gain of 3.42%; the highest card-hour rate rose from 12.67 yuan to 13.11 yuan, up 0.43 yuan, a gain of 3.39%. The core driver of this rise was: undersupply of NVIDIA spot supply in the market. A large volume of scattered computing-power resources was being leased in advance by major platforms and operators, then, after unified scheduling, subleased to earn the spread; markups in the circulation process directly pushed up the listed highest price. The rise in the highest price rather than the lowest price indicates that the shortage was concentrated in compliant spot supply that could be delivered immediately, with bargaining power tilting toward schedulers holding the resources.

[SMM View] Overall demand for computing power stayed high, and the structural imbalance on the supply side was difficult to ease in the short term. Two points warrant attention: first, scattered resources, after being “master-leased—subleased” by platforms and operators, flowed out with added markups. While this improved scheduling efficiency, it also amplified end-user prices; insufficient NVIDIA spot supply was the fundamental constraint.

Second, domestic computing power was entering the market in succession, but developers’ and consumers’ confidence in its large-model training efficiency and ecosystem compatibility still needed time to build, making it difficult in the short term to form a substantive substitute for NVIDIA high-end prices. The market called for a more transparent spot supply and pricing mechanism to compress unnecessary intermediary markups.

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

15 hours ago

AI Demand and Infrastructure Projects Drive Fiber Optic Price Surge, Top Players Secure Long-Term Orders

Benefiting from the explosion of demand from AI intelligent computing centres, the "East Data, West Computing" initiative, and transoceanic submarine cables, the upstream optical fibre preform expansion cycle takes 18-24 months, leaving industry capacity supply tight, and fibre optic prices have been continuously raised significantly in 2026. Hollow-core fibre and ultra-low-loss G.654E high-end fibre are rigid demand for computing power backbones, and top-tier players have locked in orders through long-term agreements until 2028.

15 hours ago

NDRC: China's AI Innovation Accelerates, Boosting Economic Growth with Breakthroughs in Computing, Models, and Data

The NDRC held a press conference for July. A spokesperson of the NDRC stated that in H1, China’s pace of independent AI innovation accelerated, achieving rapid breakthroughs across multiple links of the entire industry chain. First, the computing power foundation was consolidated; the first fully homegrown 100,000-card AI supercluster was officially put into operation, and as of month-end June, the national intelligent computing power scale reached 2.8 times that of the same period last year. Second, model capabilities leaped forward; local companies such as DeepSeek and Moonshot AI successively released multiple open-source large models with trillion-level parameters, total global downloads of domestic large models surpassed 10 billion times, and domestic models were rapidly adapted to domestic computing power chips. Third, factor mobility was vibrant; over 120,000 high-quality datasets were built, and a virtuous cycle of mutual reinforcement among computing power, models, and data took shape at a faster pace. Driven by coordinated efforts across all areas, AI became a surging source of economic growth momentum, and related sectors maintained high growth of over 30%.

15 hours ago

Chinese AI Models Surpass US, Hold 63.5% Global Market Share by July 26

OpenRouter previously disclosed that the global market share of Chinese models had overall surpassed that of US models by early June. As of July 26, Chinese models held approximately 63.5% of the market over the past 28 days, while US models held only 35.5%, accounting for two-thirds of the global market, with the gap further widening.

15 hours ago