China is pursuing a distinct path in the global AI race, prioritizing state control and widespread domestic adoption over the rapid development of superior intelligence that drives U.S. strategy. While both nations agree that AI power dictates global standards and future economic and military dominance, their methods diverge sharply. Chinese officials are focused on integrating AI into every level of the economy, a goal outlined in the 15th Five-Year Plan released in March. The plan mentions AI 52 times, five times more than the previous five-year plan, and dedicates a full chapter to computing power. The stated objective is to integrate AI across 90% of the economy by 2030.
This strategic focus stems from both perceived threats and structural weaknesses. Chen Yixin, China's spy chief, warned in public this month that AI could be used for "cognitive warfare against China," threatening political, institutional, and ideological security. The regime is less concerned with speculative risks like rogue AI and more focused on preventing AI from turning against the state. Domestic challenges also drive this approach. China's population has shrunk for four consecutive years, with a large over-60 demographic that fuels an obsession with robotics. Youth unemployment stands at 19%, roughly double the U.S. rate, posing a stability risk if the population gains powerful tools. Additionally, China faces a housing crisis and inferior chip production, with Huawei lagging significantly behind Nvidia despite state support.
Despite these vulnerabilities, China holds structural advantages in the physical infrastructure required for AI scaling. BloombergNEF estimates that China will add nearly six times as much generating capacity as the U.S. over the next five years, addressing the power bottleneck facing American buildouts. The International Energy Agency tracks 20 essential minerals for AI hardware, and China dominates all but one of them, often by large margins. Export controls have been applied to half of these materials to slow competitors. In robotics, more than half of all industrial robots installed worldwide in 2024 went to China. The country also leads in cheap, widespread adoption, with five of the six most popular models on the OpenRouter marketplace coming from Chinese firms.
U.S. experts generally assess that Chinese models trail American top-tier AI by about half a year but are significantly cheaper to use. Key models include Kimi K3 from Moonshot, which excels at long multi-step tasks; Qwen3.8-Max from Alibaba, which tops some rankings; GLM-5.3 from Zhipu, a coding specialist whose predecessor scored within a point of Anthropic's Claude Opus 4.8 at a fraction of the cost; and DeepSeek V4, which is MIT-licensed and among the cheapest options. U.S. officials have accused Chinese labs of "distillation," covertly training on answers from American models, and have threatened sanctions without yet implementing them.
Public sentiment in China reflects this integration. Eighty percent of Chinese people use AI weekly compared to 54 percent of Americans. Eighty-four percent say AI excites them, the highest rate of any country, while 87 percent trust it, compared to 32 percent in the U.S. However, ChinaTalk argues that this optimism is overstated and masks rising worry about job losses.