Zhipu AI Reports 400% Revenue Jump as New Model Runs Entirely on Chinese Chips

Zhipu AI Achieves 400% First-Half Revenue Growth

Zhipu AI delivered remarkable financial performance in the first six months of 2026, reporting revenue of 953.9 million yuan ($142 million), representing a 400% surge compared to the same period last year1

1

4

4

. The Beijing-based startup's gross profit climbed 163.7% to 251.6 million yuan during this period4

4

. This AI-driven revenue growth signals the company's success in converting demand for its artificial intelligence models into tangible sales. Open platform and API revenue surged 27-fold to 825.2 million yuan, now accounting for 86.5% of total revenue, compared with just 15.2% a year earlier4

4

. The company's MaaS platform attracted more than 7.4 million enterprise and developer users as of the announcement4

4

. Despite the revenue surge, net loss narrowed 12.1% to 2.07 billion yuan, though adjusted net loss widened 12.1% to 1.96 billion yuan4

4

. R&D spending increased 33.6% to 2.13 billion yuan as the company continued investing in its large-language models4

4

.

GLM-5.3-Flash AI Model Runs Entirely on Chinese Chips

Zhipu AI launched GLM-5.3-Flash, a groundbreaking AI model that the company claims operates entirely on domestically produced Chinese chips2

2

3

3

. The low-cost version of Zhipu AI's flagship model ranks 10th on the Artificial Analysis Intelligence Index, ahead of DeepSeek V4 Pro Max2

2

. The company stated it deployed 100,000 China-made chips to handle all online requests for GLM-5.3-Flash, including during its preview period2

2

3

3

. While CNBC was unable to independently verify these chip claims and the company declined to share details on which chipmakers supplied the hardware, the announcement underscores China's push for tech self-sufficiency2

2

. Running an AI model requires less computing power than training one, but the achievement still marks a significant milestone given geopolitical constraints2

2

. China has ramped up domestic semiconductor capabilities in response to U.S. restrictions on advanced chip sales, with Huawei and other Chinese companies intensifying efforts to build alternatives to Nvidia's restricted chips2

2

.

Stock Surge Follows Ox Alpha Model Revelation

The Hong Kong-listed company experienced a dramatic stock surge, with shares climbing more than 8% in Thursday trading following the GLM-5.3-Flash announcement2

2

. Additional gains of 9.7% came after Zhipu AI confirmed to Bloomberg that it developed the mysterious Ox Alpha model, which had appeared anonymously on AI platform OpenRouter on August 205

5

. The Ox Alpha model quickly climbed to the top of OpenRouter's usage rankings, attracting attention for its performance in coding, reasoning and agentic tasks5

5

. The model supports text, image and video inputs and features a context window of approximately 1 million tokens5

5

. Zhipu AI plans to release the model's weights, allowing developers to build applications and services on top of it5

5

. Since its Hong Kong IPO in January, Zhipu AI shares have skyrocketed by more than 800%2

2

.

Competitive Landscape and Future Implications

Zhipu AI's achievements arrive amid intensifying competition in China's AI sector. Rival MiniMax reported a 283% surge in first-half revenue versus a year earlier, though its adjusted net loss more than doubled to $293 million during that period2

2

. MiniMax's shares climbed around 3% in Hong Kong trading, though they've only gained over 80% since its January IPO, significantly trailing Zhipu AI's performance2

2

. MiniMax's flagship M3 model ranks 18th on the Artificial Analysis Intelligence Index2

2

. The developments highlight rapid progress among Chinese AI developers, who are increasingly challenging established U.S. models with lower-cost and open-weight systems5

5

. Leading U.S. AI models remain officially unavailable in China, creating opportunities for domestic players to capture market share2

2

. The combination of restricted semiconductor access and limited access to Western AI models is accelerating China's development of homegrown alternatives, potentially reshaping the global AI landscape as Chinese companies demonstrate they can compete on performance while building entirely on domestic infrastructure.