China’s semiconductor industry is accelerating its drive for artificial intelligence (AI) self-sufficiency, with leading chip designers and cloud providers embracing DeepSeek’s cutting-edge models. From Moore Threads to Iluvatar Corex, the nation’s tech giants are integrating DeepSeek’s open-source AI solutions, marking a significant step in reducing reliance on foreign technology.
Moore Threads Technology, a prominent GPU design firm founded by former Nvidia China executive Zhang Jianzhong, recently announced its commitment to DeepSeek’s V3 and R1 models. In a WeChat post, the company pledged to leverage its proprietary KUAE GPU intelligent computing cluster to support the distributed deployment of DeepSeek’s AI systems. “We aim to set China’s AI ecosystem on fire using locally made GPUs,” the statement read, highlighting the models’ role in inspiring domestic AI development.

The momentum extends beyond Moore Threads. Huawei Technologies, a key player in China’s tech landscape, has partnered with AI infrastructure start-up SiliconFlow to make DeepSeek’s models accessible via its Ascend cloud service. Huawei claims its Ascend solution delivers computing performance comparable to global premium GPUs, further bolstering China’s AI capabilities.
Meanwhile, Gitee AI, a Shenzhen-based platform for AI developers, is offering DeepSeek-R1-based models powered by GPUs from Shanghai’s MetaX. The collaboration underscores China’s push for end-to-end domestic innovation, from chips and platforms to computing power and AI models. Gitee AI emphasized the significance of this integration, calling it a milestone in achieving technological self-reliance.
Iluvatar Corex, another Shanghai-based GPU start-up, echoed this sentiment, describing DeepSeek-R1 as a “bolt of lightning” that shattered technical barriers and revitalized China’s AI industry. The company is working with partners to ensure seamless access to DeepSeek’s models, further solidifying their adoption across the ecosystem.
China’s major cloud service providers, including Alibaba, Tencent, Huawei, and Baidu, have also joined the fray. During the Lunar New Year holiday, these companies worked tirelessly to integrate DeepSeek’s models into their platforms, responding to surging demand from businesses and consumers.
While Chinese AI chips still lag behind those produced by US giant Nvidia, DeepSeek’s success offers a glimmer of hope. By focusing on inference processors—lightweight, efficient chips designed for generative AI tools—China aims to narrow the gap with the US. Inference chips are easier to develop than training chips, making them a strategic focus for the country’s AI ambitions.
The collaboration between chip developers and software engineers is expected to enhance the performance of home-grown open-source models, reducing dependence on US technology. As China’s tech war with the US intensifies, DeepSeek’s rising prominence signals a pivotal shift toward AI self-sufficiency, positioning the nation as a formidable contender in the global AI race.
With DeepSeek at the forefront, China’s semiconductor industry is poised to redefine its role in the AI landscape, leveraging innovation and collaboration to achieve technological independence.