Woofun AI reports that Zhiyu is evaluating the design of proprietary AI chips following a surge in demand for its GLM series models. The company has established preliminary contact with several domestic chip design firms to explore collaboration on custom AI processors.

This strategic move aims to achieve software-hardware co-optimization, enhance computing power utilization, and decrease dependence on external GPU suppliers.

Concurrently, DeepSeek is advancing its own inference-focused chip project, mirroring global trends seen with OpenAI and Anthropic.