Chinese AI startup Z.ai (formerly Zhipu AI) is considering developing its own AI chip as soaring demand for its GLM family of large language models is adding pressure on computing capacity, according to a report by The Information.
The company is reportedly looking beyond off-the-shelf processors to secure long-term access to computing power amid growing demand for AI training and inference.
According to The Information, Z.ai has begun weighing the feasibility of designing a proprietary chip, although the discussions remain at an early stage and no final decision has been made.
AIM could not independently verify the report.
The report said interest in the project has been fueled by the rapid adoption of Z.ai’s GLM models. The increase in usage has intensified the need for reliable computing infrastructure as competition among Chinese AI developers accelerates.
Any custom chip initiative would represent a significant strategic expansion for the Beijing-based company. Developing AI processors typically requires years of engineering work, substantial capital investment, and manufacturing partnerships, with only a handful of AI companies globally tasting success.
The reported deliberations also come as Chinese AI companies continue to face restrictions on access to the most advanced US-made AI chips, prompting many firms to diversify their hardware strategies through domestic suppliers and custom silicon.
Currently, NVIDIA’s H20 is the only advanced AI GPU that can be legally exported to China, and even then, shipments require US export licenses.
The H20 was specifically designed to comply with US export controls, but last April, it was subjected to an indefinite licensing requirement before the US later agreed to approve licenses for eligible exports.
Z.ai’s rival, DeepSeek, is also developing its first custom AI chip for inference, according to a recent Reuters report. The project began about a year ago and remains in its early stages. DeepSeek has been holding discussions with chip design firms, foundries, and memory suppliers while privately expanding its chip engineering team.
Z.ai has become one of China’s leading AI startups, with its GLM models gaining traction among developers and enterprises. The company has continued releasing increasingly capable open-source models while competing with domestic rivals and international AI labs.
Its latest model, GLM 5.2, performs competitively against closed models such as Claude Opus 4.8 and GPT-5.5 on coding benchmarks, including Terminal-Bench, SWE-Bench, and agentic evaluations. It often narrows the performance gap while operating at a fraction of the cost.
Your reaction