When Zhipu AI and MiniMax listed on the Hong Kong Stock Exchange within weeks of each other in January 2026, they became the first pure-play Chinese generative AI companies to go public, a milestone that marked the sector’s transition from a venture-backed research phase to the scrutiny of public markets. This week, both companies delivered their inaugural earnings reports, and the contrast between them could hardly be sharper.
Zhipu AI, the Beijing-based company behind the GLM series of large language models, reported revenue growth that exceeded analyst expectations by a considerable margin. According to the South China Morning Post, the company’s top line more than doubled compared with the same period a year earlier, driven primarily by its enterprise AI services business. Zhipu has pursued a deliberate strategy of targeting large corporate and government clients with customized AI deployments — a model that generates lower headline user numbers than consumer apps but produces more predictable, higher-margin revenue. The market’s reaction was emphatic: shares surged more than 35% in the session following the announcement, adding billions to the company’s market capitalization.
MiniMax told a different story. The Shanghai-based company, which has built a reputation for its multimodal capabilities and its Hailuo AI video generation product, reported strong growth in active users and a significant expansion of its international footprint. But losses widened beyond what analysts had projected, as the company continued to invest aggressively in model research, infrastructure, and the marketing costs associated with consumer product launches. Management framed the spending as necessary to establish a durable position in what they described as a winner-take-most market for foundation model capabilities.
The divergence between the two companies reflects a genuine strategic debate playing out across China’s AI industry. The enterprise-first model that Zhipu has pursued has the advantage of generating revenue relatively early in a company’s development, large clients pay for customized deployments, integration work, and ongoing support, providing a revenue base that does not depend on achieving mass consumer adoption. The risk is that enterprise AI contracts can be slow to scale and are vulnerable to competition from the AI divisions of large technology platforms, which can bundle AI capabilities with existing enterprise software relationships.
The consumer and capability-first model that MiniMax represents bets on a different outcome: that the companies that build the most capable and widely used AI products will eventually find monetization pathways, just as social media and search companies did in earlier technology cycles. The risk is that the path to profitability is long, capital-intensive, and uncertain, particularly in a market where Baidu, Alibaba, and Tencent are competing for the same users, leveraging existing platform relationships.
Both companies operate in a market that is simultaneously intensely competitive and heavily state-supported. China’s government has made the development of a domestic generative AI industry a strategic priority, and a range of subsidies, procurement preferences, and regulatory accommodations have flowed to domestic players. But state support does not eliminate the need to build sustainable businesses, and the public markets are now providing a real-time verdict on which strategies are working.
The broader significance of these first earnings reports extends beyond the two companies themselves. Zhipu and MiniMax are at the forefront of what is expected to be a wave of Chinese AI company listings in Hong Kong and, potentially, on mainland exchanges. Their performance will shape investor appetite for the sector and influence the valuations at which subsequent companies come to market. A strong showing from Zhipu, in particular, sends a signal that there is genuine commercial demand for Chinese AI products, not just government-mandated adoption but organic enterprise spending driven by demonstrated value.
For international observers trying to assess the maturity of China’s AI industry, these earnings reports are among the most useful data points to emerge in some time. They confirm that at least some Chinese AI companies are building real businesses, not just impressive technology demonstrations. They also confirm that the path to profitability in generative AI is not straightforward — a lesson that applies equally to companies in the United States and Europe.