Home
Zhipu AI launches 1GW domestic AI computing center, acquires Zhongke Jiahe to boost AI infrastructure

ZhiPu AI (Z.ai) has built a 1GW-level domestic AI computing data center, powered entirely by domestically produced AI chips. The company also recently acquired XCore Sigma, a domestic AI heterogeneous computing software firm, strengthening its AI infrastructure from computing resources to foundational software.
XCore Sigma traces its roots to the Compiler Laboratory of the Institute of Computing Technology, Chinese Academy of Sciences. The company specializes in heterogeneous computing software stacks, compiler optimization, runtime systems, and AI inference infrastructure, and is widely recognized as a leading AI Infra technology team in China. This acquisition bolsters ZhiPu's ability to adapt to domestic chips, optimize computing resource scheduling, and streamline model deployment.
Industry observers note that these two actions align with the "computing power supply" and "computing power release" stages of the AI value chain. The 1GW-level domestic computing center delivers stable computing resources for large-scale model training, while XCore Sigma's software technology boosts the efficiency of various AI chips—improving hardware utilization and cutting inference costs via compiler, runtime, and inference engine optimizations.
As large-model competition shifts toward infrastructure and engineering prowess, AI companies are moving beyond model size alone to building comprehensive technical systems spanning chips, computing power, software stacks, and application deployment. The market is now closely watching ZhiPu's next-generation foundational models. Some analysts suggest that with large-scale domestic computing resources, a mature AI Infra system, and its accumulated post-training expertise, ZhiPu's future models will likely scale toward larger parameters and higher intelligence, while enhancing inference efficiency and real-world deployment.
The construction of this computing center and the integration of foundational software capabilities signal that domestic large-model firms are accelerating the creation of an autonomous and controllable AI infrastructure ecosystem. This reflects a broader shift in AI industry competition—from model capabilities alone to full-stack technological system competition.
Related article
Meta Removes AI Photo Editing Feature Following User Backlash
Meta, the social media giant, is once again embroiled in a public debate regarding the delicate balance between artificial intelligence and user privacy. According to TechCrunch, Meta’s Superintelligence Labs introduced a new AI image generator, Muse
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo
OpenAI, Anthropic Vie for Market Share Despite Revenue Shortfalls
Despite recent reports suggesting OpenAI missed revenue targets, creating pressure on tech stocks this Tuesday, private AI lab investors remain resilient. Seasoned backers have confirmed they will not reduce investment despite negative media coverage
Related Special Topic Recommendations
Comments (0)
0/500

ZhiPu AI (Z.ai) has built a 1GW-level domestic AI computing data center, powered entirely by domestically produced AI chips. The company also recently acquired XCore Sigma, a domestic AI heterogeneous computing software firm, strengthening its AI infrastructure from computing resources to foundational software.
XCore Sigma traces its roots to the Compiler Laboratory of the Institute of Computing Technology, Chinese Academy of Sciences. The company specializes in heterogeneous computing software stacks, compiler optimization, runtime systems, and AI inference infrastructure, and is widely recognized as a leading AI Infra technology team in China. This acquisition bolsters ZhiPu's ability to adapt to domestic chips, optimize computing resource scheduling, and streamline model deployment.
Industry observers note that these two actions align with the "computing power supply" and "computing power release" stages of the AI value chain. The 1GW-level domestic computing center delivers stable computing resources for large-scale model training, while XCore Sigma's software technology boosts the efficiency of various AI chips—improving hardware utilization and cutting inference costs via compiler, runtime, and inference engine optimizations.
As large-model competition shifts toward infrastructure and engineering prowess, AI companies are moving beyond model size alone to building comprehensive technical systems spanning chips, computing power, software stacks, and application deployment. The market is now closely watching ZhiPu's next-generation foundational models. Some analysts suggest that with large-scale domestic computing resources, a mature AI Infra system, and its accumulated post-training expertise, ZhiPu's future models will likely scale toward larger parameters and higher intelligence, while enhancing inference efficiency and real-world deployment.
The construction of this computing center and the integration of foundational software capabilities signal that domestic large-model firms are accelerating the creation of an autonomous and controllable AI infrastructure ecosystem. This reflects a broader shift in AI industry competition—from model capabilities alone to full-stack technological system competition.
Meta Removes AI Photo Editing Feature Following User Backlash
Meta, the social media giant, is once again embroiled in a public debate regarding the delicate balance between artificial intelligence and user privacy. According to TechCrunch, Meta’s Superintelligence Labs introduced a new AI image generator, Muse
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo
OpenAI, Anthropic Vie for Market Share Despite Revenue Shortfalls
Despite recent reports suggesting OpenAI missed revenue targets, creating pressure on tech stocks this Tuesday, private AI lab investors remain resilient. Seasoned backers have confirmed they will not reduce investment despite negative media coverage











