JD Cloud to Deploy 100,000 Moore Threads GPUs for AI Computing
JD Cloud has unveiled a plan to construct a large-scale computing cluster equipped with 100,000 general-purpose graphics processing units (GPUs) manufactured by Chinese chipmaker Moore Threads. The announcement was made at JD.com's 2026 Global Technology Explorers Conference, where the company outlined its ambition to support artificial intelligence workloads including large-model training, inference, and embodied AI applications. The computing capacity is expected to be offered to businesses across various industries, positioning JD Cloud as a key player in the domestic AI infrastructure market.
According to JD Cloud, the project aims to establish a robust computing foundation built entirely on Chinese hardware, which it claims will facilitate the commercialization of physical AI. The company emphasized that this deployment represents the first instance of a major domestic AI cloud provider integrating 100,000 Chinese GPUs into a single core computing cluster. The initiative is seen as a strategic move to reduce reliance on foreign chip technology and bolster China's self-sufficiency in advanced computing resources, particularly amid ongoing export restrictions.
The cluster is designed to handle the intensive computational demands of training sophisticated neural networks, as well as the rapid inference required for real-time AI applications. JD Cloud highlighted the potential for this infrastructure to accelerate innovation in sectors such as autonomous driving, robotics, and smart manufacturing, where embodied AI—AI that interacts with physical environments—plays a crucial role. By offering these resources on a cloud platform, JD Cloud aims to democratize access to high-performance computing for smaller enterprises that lack the capital to build their own facilities.
Industry analysts note that the move aligns with China's broader national strategy to achieve technological independence in critical sectors. Moore Threads, founded in 2020, has emerged as a leading domestic GPU designer, with products targeting both data center and consumer markets. The 100,000-unit order from JD Cloud would represent a significant scale-up for the company, which has been working to expand its production capacity and software ecosystem.
While specific financial details and timelines were not disclosed, JD Cloud stated that the project would proceed in phases, with the first cluster expected to come online within the next few years. The company also announced plans to invest in supporting software development tools and frameworks to ensure seamless integration with popular AI frameworks like PyTorch and TensorFlow. This move is likely to attract developers seeking alternatives to NVIDIA's CUDA platform, which currently dominates the AI computing landscape.
Observers suggest that this ambitious deployment could face challenges, including potential bottlenecks in chip manufacturing yields and the need to optimize software for Moore Threads' architecture. However, JD Cloud's commitment signals growing confidence in domestic technology. The project is also expected to contribute to the development of China's digital economy, as more enterprises leverage AI to improve efficiencies and create new products.
In summary, JD Cloud's 100,000-GPU cluster powered by Moore Threads represents a landmark initiative that underscores the rapid advancement of China's semiconductor industry and its integration into cloud services. As the project unfolds, it will be closely watched by both domestic and international stakeholders for its impact on the global AI supply chain.