At the 2026 World Artificial Intelligence Conference, Alibaba Cloud officially released the Lingjun Zhenwu M890 supernode instance, providing a 64-card computing power unit with high-speed interconnection and ready to use out of the box in the form of a supernode. This is the first time that Alibaba Cloud has provided AI computing power services in the form of supernodes through a public cloud. The Lingjun Zhenwu M890 supernode instance supports FP8/FP4 low precision computing. Through the ICNSwitch1.0 chip, the scale-up interconnection scale was raised from 16 color to 64 cards, and the inter-card interconnection was raised to 800Gb/s. A supernode instance can carry the inference of a large 10-trillion parameter-level MoE model. Recently, the Lingjun Super Node instance is now open for testing in the Ulanqab region.

At the 2026 World Artificial Intelligence Conference, Alibaba Cloud officially released the Lingjun Zhenwu M890 supernode instance, providing a 64-card computing power unit with high-speed interconnection and ready to use out of the box in the form of a supernode. This is the first time that Alibaba Cloud has provided AI computing power services in the form of supernodes through a public cloud. The Lingjun Zhenwu M890 supernode instance supports FP8/FP4 low precision computing. Through the ICNSwitch1.0 chip, the scale-up interconnection scale was raised from 16 color to 64 cards, and the inter-card interconnection was raised to 800Gb/s. A supernode instance can carry the inference of a large 10-trillion parameter-level MoE model. Recently, the Lingjun Super Node instance is now open for testing in the Ulanqab region.