As China and the U.S. battle for AI dominance, AI agents are emerging as the next frontier in the race, with their ability to work autonomously toward a goal set to transform both enterprise and consumer markets. In China, the push for consumer-facing AI agents faces headwinds as some of the biggest tech players including Alibaba and Tencent restrict agent access to their proprietary platforms. But Beijing has a secret weapon that could turbocharge adoption and tilt the balance in its favour: far lower inference costs.
“Adoption of AI agents by enterprises and consumers alike may be faster in China than in other markets such as the US because the inferencing cost in China is roughly 90% below the inferencing cost in the US,” Barclays’ analysts said in a recent note. “As in all tech adoption curves, having lower costs tends to turbo charge adoption.”
Walled Garden Roadblocks Limit Consumer Reach
While Chinese big techs have rolled out their own agents, the likes of Alibaba Group Holdings Ltd ADR (NYSE:BABA) in e‑commerce, Tencent in payments, and Meituan in food delivery have kept tight control over access unless special partnerships and deals are made.
Without special partnerships, even the smartest AI agents cannot complete simple transactions in these ecosystems. The analysts said that for consumer-facing AI, this means applications remain limited to tasks like information retrieval rather than seamless order placement or bookings.
Enterprises Have the Edge Without Third‑Party Dependencies
Enterprises, however, face fewer barriers. Internal AI agents, tailored with proprietary data and workflows, do not need access to third‑party platforms. Almost 200 enterprises and government entities have already spent at least RMB 1.1 billion in the first half of 2025 working with DeepSeek. The analysts highlighted that adoption in sectors from finance and healthcare to education and supply chain management, where agents can boost productivity and efficiency.
Chinese large language models, or LLMs, have achieved competitive performance at a fraction of global prices, with most models launched in recent months priced at just about 10% of OpenAI, Google (NASDAQ:GOOGL), or Anthropic equivalents.
That pricing power, combined with free‑to‑use quotas for most domestic agents, creates a lower barrier for mass deployment. Barclays insists that as adoption accelerates, feedback loops will emerge, improving both foundational models and the agents themselves.
China’s Strategic Play for AI Dominance
The combination of rapid model iteration, falling inference costs, and growing enterprise adoption gives China a unique opportunity to scale AI agents faster than rivals, Barclays said. While walled gardens still limit consumer utility, the analysts suggested that the low‑cost advantage could enable agent technology to mature and spread more quickly in China than in the U.S.
Source :
https://www.investing.com/news/economy-news/chinas-ai-agents-hit-big-tech-barriers-but-low-costs-give-it-the-edge-in-ai-race-4183797
