Businesses Shift to Cheaper Chinese AI Models Amid Soaring US Costs
Global businesses are increasingly opting for less expensive Chinese open-weight artificial intelligence models as the cost of premium, closed-source US alternatives like OpenAI's GPT and Anthropic's Claude continues to rise. These Chinese models are reported to offer performance levels close to the current frontier of AI capabilities. A significant surge in usage has been observed on Vercel, a cloud platform for AI web development based in San Francisco. Specifically, the daily token volume for Zhipu's GLM-5.2 model has increased fiftyfold since mid-June. This model operates at approximately one-fifth the cost of Anthropic's Claude Opus 4.8, highlighting a substantial cost advantage. Another Chinese model, DeepSeek's V4 Flash, is also gaining traction in this evolving market landscape. This trend indicates a growing demand for cost-effective AI solutions that can still deliver high performance, challenging the dominance of established US-based AI providers.
The escalating costs associated with proprietary, closed-source AI models from US companies may be creating a market opening for competitors offering more affordable, open-weight alternatives. This shift suggests that for many businesses, the economic viability of AI deployment is becoming as critical as raw performance. The emergence of Chinese models at a fraction of the cost, while maintaining near-frontier capabilities, could reshape the global AI landscape. Companies must weigh the potential benefits of cost savings and model flexibility against factors such as data security, regulatory compliance, and long-term vendor support when considering these alternative solutions. The competitive pressure from these lower-cost options may also incentivize US-based providers to re-evaluate their pricing structures and open-source strategies in the coming years.
AI-generated to prompt reflection — not editorial opinion, not advice, not a statement of fact. How this works.