Chinese firms buy Nvidia gaming and workstation chips
Chinese technology companies reportedly bought more than 1 million Nvidia gaming and workstation chips this year; the processors were not originally designed for advanced AI systems.

Key takeaways · 4
- 01
The reported purchases exceed 1 million chips and include the RTX Pro 6000D, RTX Pro 5000 and RTX 5090.
- 02
The chips are suited to simpler AI work, including processing prompts before a model generates a response.
- 03
China’s expected 2026 AI chip supply is 3.5 million units, while estimated demand is more than 5 million.
- 04
The proposed RTX Pro 5500 uses GDDR7 rather than the high-bandwidth memory found in server-class AI accelerators.
Reported Nvidia chip purchases
TrendForce reports that Chinese technology companies purchased more than 1 million Nvidia gaming and workstation chips this year.[1] The reported models include the RTX Pro 6000D, RTX Pro 5000 and RTX 5090.[1] These processors were not originally designed for advanced AI systems, but they can handle simpler AI workloads.[1]
Useful for prompts, weaker on many tasks
The Nvidia processors generally perform below advanced Chinese AI chips such as Huawei’s flagship Ascend 910C across most computing tasks.[1] They can, however, process the prefill stage of inference: the point when a model processes a user’s input before decoding and generating a response.[1] For teams assessing these chips, that distinction matters: reported suitability for a simpler workload does not establish that they can replace more capable accelerators across an AI system.[1]
Domestic supply is growing, but demand is higher
China’s domestic AI chip production more than doubled this year, with Huawei, Alibaba, Baidu and Cambricon among the companies driving the increase.[1] TrendForce expects China’s AI chip supply to reach 3.5 million units in 2026, including about 1 million manufactured overseas, while estimating total demand at well over 5 million units.[1] Huawei Rotating Chairman Eric Xu said in September that the company could not produce enough AI computing hardware to meet domestic demand; the shortage led Huawei to prioritize Chinese customers over overseas expansion.[1]
A proposed model brings timing and trade-offs
Nvidia is reportedly preparing the RTX Pro 5500, expected to outperform its existing China offerings and better support basic AI computing.[1] The chip is designed primarily for engineering and graphics work, uses GDDR7 instead of the high-bandwidth memory found in server-class AI accelerators, and was reported to cost about RMB 85,000–90,000 in China.[1] The Ministry of Industry and Information Technology requested intended purchase volumes and usage plans, and may approve domestic purchases, including by ByteDance and Alibaba.[1] ByteDance was reportedly considering about 1 million units; Nvidia reportedly planned quarterly shipments of about 500,000 starting in late December, with orders due by September 30 to secure supply.[1]
The reports point to a practical distinction for AI infrastructure planning: hardware that can handle prompts or simpler workloads may help address capacity needs, but it is not equivalent to advanced accelerators across computing tasks. Teams evaluating purchases should weigh workload fit, memory type, price and reported delivery plans rather than treating the chip count alone as usable capacity.
Why it matters
Test yourself on this story — 2 questions.
Create a free account to take the quiz, earn XP, and get a daily session built for your industry.
Take the quizHow this developed
15 October 2026
Chinese firms buy Nvidia gaming and workstation chips