China's Low-Cost AI Agents Poised to Outpace US Rivals

In the ongoing race for AI supremacy between China and the US, a new frontier has emerged: AI agents. These autonomous AI systems are set to transform both enterprise and consumer markets. However, China's push for consumer-facing AI agents faces some headwinds.

As of August 2025, the major tech players in China, including Alibaba and Tencent, have restricted access to their proprietary platforms for AI agents. This means that even the most advanced AI agents cannot complete simple transactions within these ecosystems without special partnerships and deals.

Despite these walled gardens, China has a secret weapon that could turbocharge the adoption of AI agents: far lower inference costs. Barclays' analysts estimate that the inferencing cost in China is roughly 90% below the cost in the US. This pricing advantage, combined with free-to-use quotas for most domestic agents, creates a lower barrier for mass deployment.

The impact is already being felt. In the first half of 2025, nearly 200 enterprises and government entities in China have spent at least RMB 1.1 billion (around $160 million) working with DeepSeek, a leading AI agent platform. Sectors ranging from finance and healthcare to education and supply chain management are embracing these AI agents to boost productivity and efficiency.

Furthermore, Chinese large language models (LLMs) have achieved competitive performance at a fraction of global prices, with most models launched in recent months costing just 10% of their OpenAI, Google, or Anthropic equivalents. This pricing power, combined with the low inference costs, gives China a unique opportunity to scale AI agents faster than its rivals.

While consumer-facing applications remain limited due to the walled gardens, the analysts suggest that the low-cost advantage could enable agent technology to mature and spread more quickly in China than in the US.