Shares of Z.ai surged by more than 12 percent in Hong Kong on Thursday following the release of its latest artificial intelligence model, GLM-5.3-Flash. Designed specifically to operate on domestic Chinese hardware, the new model represents a leaner, lower-cost alternative to the company’s flagship offering. According to Z.ai, the system is powered by 100,000 domestically produced chips, though the company declined to specify which manufacturers provided the hardware and these claims have not been independently verified.
The market reaction follows an explosive debut for the model, which initially operated under the code name Ox Alpha during a one week preview window. During those first few days, it generated over 11 trillion tokens on OpenRouter, marking a record opening for the platform and quickly ascending to the top spot among coding models. With pricing set at roughly one tenth the cost of similar services, Z.ai aims to disrupt the market through sheer affordability and efficiency.
Industry experts suggest that Z.ai likely utilized a mix of processors from various local vendors, potentially including Huawei Ascend chips. This move aligns with a broader strategic shift within China to tighten integration between homegrown software and hardware stacks to bypass U.S. export controls on high end Nvidia semiconductors. By building a dedicated inference engine that reportedly triples serving performance over previous baselines, Z.ai claims it has reached efficiency levels comparable to industry standard Nvidia GPUs.
This technological pivot arrives as Beijing intensifies efforts to reduce reliance on foreign silicon while domestic firms accelerate their own competing offerings. For investors, the momentum appears unstoppable so far; Z.ai is scheduled to report its first half results this coming Monday with its stock already having climbed more than 800 percent since its public listing in January.