Chinese artificial intelligence company Z.ai is on track to become the country’s first independent AI developer to reach $1 billion in annualized revenue, according to Bloomberg. The company reportedly achieved its full-year sales target by July, with that month’s revenue pace translating to approximately $1 billion over 12 months.
The figure represents an annualized run rate based on recent sales rather than $1 billion of revenue already recorded during 2026. Maintaining that level will depend on continued enterprise demand, customer retention and usage of Z.ai’s models.
Z.ai’s annual recurring revenue reportedly increased approximately 15-fold between January and July. The company moved from a $100 million annualized run rate to $1 billion in about five months, compared with approximately 15 months for Anthropic to complete the same progression.
The growth substantially exceeds earlier industry expectations for the company. In April, Visible Alpha consensus estimates projected that Z.ai would generate approximately HK$3.2 billion, or $409 million, in total revenue during 2026.
Z.ai, formerly known as Zhipu AI, develops the GLM family of large language models and related enterprise AI products. Its revenue comes from cloud-based services accessed through subscriptions or usage-based pricing and customized deployments installed within customers’ own data centers.
Cloud services have emerged as a major growth engine as companies seek AI products that can be deployed more quickly and with lower upfront infrastructure costs. Earlier estimates projected that Z.ai’s cloud revenue would rise from $24 million in 2025 to approximately $258 million in 2026 and account for about 63% of its total sales.
The company has placed a particular emphasis on coding, complex reasoning and autonomous agent workloads. Z.ai began concentrating additional resources on coding capabilities in early 2025 and has since released updated flagship models at a rapid pace.
Its latest model, GLM-5.2, has performed competitively against several leading U.S. systems on public coding and agent benchmarks. The 750 billion-parameter model supports a one-million-token context window and is designed to complete long-duration, multistep assignments.
GLM-5.2 was ranked fourth on Artificial Analysis’ intelligence leaderboard and second on Code Arena’s front-end coding benchmark following its launch. Reuters reported that the model operated at roughly one-sixth of the cost of leading closed U.S. frontier systems.
Z.ai releases its leading models with open weights, allowing businesses and developers to download, modify and deploy them on their preferred infrastructure. This approach could appeal to organizations seeking greater control over data, costs and reliance on external AI providers.
The company has also adapted the GLM-5 series to operate across domestic Chinese semiconductor infrastructure, including Huawei Ascend clusters. This work has become increasingly important as U.S. restrictions limit China’s access to the most advanced NVIDIA chips.
Despite intense price competition within China’s AI market, Z.ai has increased prices for several products as usage has expanded. Its GLM application programming interface prices rose by a cumulative 83% during the first quarter, while usage increased by approximately 400%.
The company’s commercial momentum is being driven partly by demand from enterprises, developers and public-sector customers. Coding tools can generate recurring revenue through subscriptions and token consumption as employees incorporate AI into software development workflows.
Z.ai went public in Hong Kong in January 2026 and has announced plans to pursue an additional listing in Shanghai. Its shares rose more than 2,000% during the first six months after the Hong Kong debut, pushing its market capitalization above $128 billion in June.
The company intends to use capital from its listings to support its pursuit of artificial general intelligence and continued model development. Its next major model, GLM-5.5, is expected to follow as Z.ai competes with Chinese developers including DeepSeek, Moonshot AI and MiniMax.
Reaching a $1 billion annualized revenue pace would provide one of the clearest signs that Chinese foundation-model companies can build large commercial businesses alongside their technical advances. The milestone also increases competitive pressure on OpenAI and Anthropic as enterprises compare model performance, deployment control and cost.
