US-China AI rivalry reaches critical flashpoint

Lead

China has dramatically narrowed its gap with the United States in top-tier artificial intelligence, with several reports indicating the lead has shrunk from years to just months — or even weeks. The rapid advance, driven by Chinese firms' focus on open-source models and algorithmic improvements, has prompted a new wave of policy responses in Washington, including proposed export controls and a push to promote American AI abroad.

Chinese AI models such as DeepSeek, Alibaba's Qwen, and Xiaomi's MiMo are now widely deployed across businesses, government services, and consumer devices, according to Deutsche Welle. The German broadcaster reports that China, facing US curbs on high-end chips, has prioritized "good enough" technology built on open-source models.

The South China Morning Post reports that the US is strengthening its AI posture towards China, with lawmakers advancing new export control bills and the Trump administration deepening cooperation with leading US AI firms. A senior Anthropic executive said the US holds a six-to-nine-month lead in frontier AI models over Chinese competitors.

Coverage comparison

Deutsche Welle's reporting emphasizes China's rapid progress and the closing gap, highlighting the deployment of Chinese models across sectors and the unveiling of Moonshot AI's Kimi K3, which the outlet says rivals one of Anthropic's most advanced models on key benchmarks. The broadcaster quotes technology specialist Alvin W. Graylin, who said the model "commoditizes frontier models," making high-end AI affordable for most customers.

The South China Morning Post's coverage focuses on the US policy response. The paper reports on Washington's new "AI Spark" initiative to promote American AI systems in Southeast Asia, which it describes as a move to counter China's growing influence in the region. The paper also covers the administration's deepening cooperation with AI firms and the congressional push for new export controls.

One notable difference in tone: Deutsche Welle's reporting is neutral and informative, while the South China Morning Post adopts a more cautious, concerned tone, using language like "unauthorized" and "adversarial" to describe Chinese AI practices.

Key claims

• China's gap with the US in top-tier AI has narrowed from up to five years to just a few months, as reported by Deutsche Welle. The South China Morning Post adds that the lead has shrunk from years to months, and in recent days, to weeks for the first time.
• Chinese AI models are being deployed across businesses, government services, and consumer devices, according to Deutsche Welle.
• Chinese startup Moonshot AI unveiled Kimi K3, a model that rivals Anthropic's Claude on key benchmarks, Deutsche Welle reports. In some tests, Kimi K3 matches or surpasses Claude with less computing power and at a fraction of the cost.
• The US is considering restricting open-weight Chinese models at home, according to the South China Morning Post, which reports that Washington faces three difficult questions: how tightly to control access to frontier US models, whether to restrict open-weight Chinese models, and how to drive global adoption of US AI.
• Washington has launched the AI Spark initiative to promote American AI systems in Southeast Asia, and is investing $2.5 billion in strategic investments in the region, the South China Morning Post reports.
• The AI race is also setting up a new rivalry over plans to take successful AI labs public, as reported by Deutsche Welle.
• A senior Anthropic executive, Tarun Chhabra, said the US holds a six-to-nine-month lead in frontier AI models over Chinese competitors, the South China Morning Post reports. Chhabra accused Chinese companies of using unauthorized "distillation" techniques to replicate American advances, saying the US could have been "a year to 18 months ahead" without this practice.

Perspectives

US Government

The Trump administration is deepening cooperation with leading US AI firms and considering new export controls to restrict China's access to advanced chips, as reported by the South China Morning Post. Treasury Secretary Scott Bessent has warned the US could sanction Chinese models over alleged intellectual property theft, while President Donald Trump has threatened to block US access to Chinese AI models, according to Deutsche Welle.

Anthropic

A senior executive at Anthropic, Tarun Chhabra, maintains that the US still holds a six-to-nine-month lead in frontier AI models, but accuses Chinese labs of using unauthorized "distillation" techniques to replicate American advances, calling the practice "adversarial" and saying it costs the US a year to 18 months of lead time.

Chinese AI Firms

Chinese tech firms are pursuing a different path from US players, prioritizing open-source models built on "good enough" technology, as Deutsche Welle reports. The deployment of these models across sectors suggests a strategy of rapid commercialization and "fast following" that has narrowed the technology gap.