Alibaba Cloud said it launched its Lingjun Zhenwu M890 supernode instance on Aug. 12, with the first batch available in the Ulanqab region, as it expands cloud-based AI computing for enterprise customers.

The company said the service lets enterprise customers activate 64-card, high-speed interconnect computing units through the cloud instead of building their own data centers.

Alibaba Cloud said the instance supports inference for mixture-of-experts, or MoE, large models with up to 10 trillion parameters.

It also said the instance is the first supernode-format computing platform in China to run a large model with more than 2 trillion parameters.

The company added that Kimi K3 and Qwen3.8-Max have already provided external services through the instance.

🔗 Source: Yicai

Recent Alibaba Cloud developments