AI Business

Z.ai revenue jumps 400% as open-platform and API demand narrows losses

Chinese AI company Z.ai reported a 400% increase in first-half revenue, helped by demand for its open platform and API services, while total losses narrowed.

Published Updated
Z.aiZhipu AIAI Models

Z.ai, the Chinese artificial intelligence company known domestically as Zhipu AI, reported a sharp first-half revenue increase as demand for its open platform and API services accelerated. According to the South China Morning Post, the company said revenue for the first six months of 2026 rose 400% to 953.89 million yuan, or about $142 million, while total losses narrowed despite continued spending on model development and computing infrastructure.

The results offer a useful view into the economics of China’s generative AI market. Z.ai has built its business around large language models and developer-facing cloud services, a strategy that makes it less dependent on a single consumer app and more exposed to enterprise and platform demand. The company said annual recurring revenue had reached $1.6 billion by the end of August, a figure that suggests customers are committing to ongoing usage rather than one-off experimentation. SCMP also noted that MiniMax chief executive Yan Junjie said his company’s annual recurring revenue reached $800 million in August, underscoring how fast Chinese model providers are trying to turn usage into predictable revenue.

Z.ai’s loss profile remains closely watched. The company reported that its total loss narrowed 12.1% to 2.07 billion yuan in the first half, while adjusted net loss rose 12.1% to 1.96 billion yuan. The difference reflects the reality of the sector: revenue is growing quickly, but leading AI companies are still spending heavily to improve base-model performance, buy or rent compute, and keep pace with increasingly capable competitors. Z.ai said research and development expenses rose 33.6% to 2.13 billion yuan, more than twice its reported first-half revenue. That level of investment shows why investors are focused not only on sales growth, but also on whether model companies can eventually improve margins as inference efficiency, enterprise contracts and cloud utilization mature.

The company’s market performance also reflects a more complicated mood around AI assets. Z.ai shares closed up 9.63% at HK$1,195 ahead of the earnings release, but SCMP noted that the stock was still down about 60% from a June high of HK$2,980. That split reaction is familiar across the AI sector: investors reward evidence of demand, yet remain cautious about valuations that rose quickly during the model boom. Revenue growth can validate a business model, but it does not remove questions about capital intensity, pricing pressure, and the speed at which open-source or low-cost competitors can narrow technical gaps.

The broader significance is that China’s AI platform companies are moving from headline model releases toward revenue proof. Z.ai’s open-platform and API gains show that enterprises and developers are increasingly willing to embed domestic models into products, workflows and cloud services. This matters for China’s technology policy as well as for competition. Local model providers are expected to support industrial software, consumer applications and government-backed digital infrastructure while operating under national data, security and compute constraints. If companies like Z.ai can keep improving performance while growing recurring revenue, they may become durable infrastructure providers rather than only model labs.

For now, Z.ai’s report points to both momentum and tension. The company is selling more AI capacity at a much faster pace, but it is also spending aggressively to stay competitive. Its latest numbers will therefore be read less as a declaration of profitability and more as a signal that China’s model market is entering a new phase, where revenue quality, platform adoption and operating efficiency matter as much as benchmark performance.