China’s AI Industry Transitioning from Models to Agents
A state report reveals that China’s AI industry is evolving, shifting from competing on large models and raw computing power to deploying and selling AI agents. By 2029, the China Telecom Research Institute predicts that inference will dominate China’s compute market, accounting for 80% of the country’s computing-power market, surpassing training demand.
This transition is accompanied by significant investment: Chinese technology companies are projected to spend approximately $89 billion on AI this year, representing over a tenth of the country’s total investment.
In contrast, Europe is building gigafactories to meet the same demand at a different pace. The EU aims to establish up to seven gigafactories by mid-2028 on a €30 billion budget, with roughly €1 billion secured thus far. While this initiative lags behind China’s pace, it represents a substantial investment in AI infrastructure.
The report from the China Telecom Research Institute highlights the economic shift towards inference, where companies optimizing models rather than building them aim to capture significant margins. Meanwhile, Europe is still in the early stages of deciding the location and construction timeline for these factories.