The US-China artificial intelligence race has shifted from a distant competition into a head-to-head sprint, with Chinese labs now releasing competitive models within days of American announcements. While the United States maintains significant advantages in advanced computing power and frontier model quality, China's strategy of building cheaper, open-source alternatives is reshaping how the entire industry operates and forcing American firms to move faster than ever before.

How Has DeepSeek Changed the Competitive Landscape?

The turning point arrived in January 2025, when DeepSeek's R1 reasoning model matched leading US systems on several benchmarks while costing a fraction of what American companies spent to build comparable systems. The market reaction was immediate and severe: roughly one trillion dollars in US tech stock value evaporated on January 27, 2025, with Nvidia alone losing about 600 billion dollars. President Donald Trump called the moment a "wakeup call" for American AI companies.

That shock has fundamentally altered how the industry operates. In late April 2026, OpenAI released GPT-5.5, and within about a day, DeepSeek released a preview of its open-source V4 model, which was built to reduce reliance on Nvidia hardware. Six weeks later, on June 13, Beijing-based Zhipu AI released GLM-5.2 under a permissive MIT license and priced its associated coding plan at roughly one-tenth of Anthropic's premium pricing tiers. Each Chinese launch now lands as a competitive prompt rather than a footnote, and US firms are increasingly timing their own moves around it.

Where Does America Still Hold the Advantage?

Despite the acceleration in Chinese releases, the United States retains commanding advantages in the areas that matter most for frontier artificial intelligence development. The Council on Foreign Relations estimated that the best US artificial intelligence chips are about five times more powerful than China's best by total processing performance, with that lead projected to widen substantially through 2027. The RAND Corporation estimated that the US has roughly ten times more compute capacity than China, representing around 77 percent of the global total compared to China's 12 percent.

The US also maintains an edge in frontier model quality. On the Code Arena leaderboard cited by the South China Morning Post, the only two non-Anthropic models in the top ten as of mid-June were Alibaba's Qwen3.7-Max and Zhipu's GLM-5.1, ranked eighth and ninth respectively. Trust represents another significant American advantage. A Public First survey of more than 18,000 people across 15 countries found that respondents rated US models as more trustworthy than Chinese ones even as many judged China to be ahead on capability. On a net-trust measure, the US scored plus 16 and China minus 8.

What Is China's Long-Term Strategy?

Rather than competing directly on the most advanced chips, China is pursuing a different path that could prove harder to contain. Li Cheng, founding director of the Centre on Contemporary China and the World at the University of Hong Kong, argued in a co-authored paper that Beijing could erode America's artificial intelligence "moat" over the next 10 to 20 years. Using Nvidia chief Jensen Huang's framework of the sector as layers spanning energy, chips, infrastructure, models and applications, the authors held that US leadership in chips is offset by Chinese strength in power generation and in pushing artificial intelligence into industry.

The US-China Economic and Security Review Commission described China's open-model strategy and its manufacturing base as "mutually reinforcing," creating a digital loop and a physical loop that compound each other. Cheap, open, widely adopted models feed an industrial economy that then deploys them at scale, which in turn justifies more model development. Huang himself has noted that close to half of the world's artificial intelligence researchers are based in China, a talent base that is difficult to sanction away.

How Are US Companies Responding to Chinese Competition?

American firms have begun hedging on the open-source question they once largely ceded to China. Nvidia released an open-weight model called Nemotron 3 in March 2026, a notable shift for a company whose advantage has rested on closed, premium hardware and software. Pricing across the entire industry has fallen sharply as Chinese open-weight models undercut closed US systems. The throughline is clear: Chinese releases keep resetting expectations on cost and accessibility, and US incumbents keep having to meet them.

• Accelerated Release Cycles: US companies are timing product announcements around Chinese releases rather than following their own development schedules, compressing the time between major model updates.
• Price Reductions: Industry-wide pricing has fallen sharply as Chinese open-source models undercut the premium pricing that US companies previously maintained for closed systems.
• Open-Source Shifts: Companies like Nvidia that once relied entirely on closed, proprietary models are now releasing open-weight alternatives to compete with Chinese offerings.

What Do Global Perceptions Reveal About the Race?

The competition is no longer being judged only in Washington and Beijing. In the Public First poll, respondents in 11 of 15 countries believed China was outpacing the US, including more than 40 percent in Canada, Britain and France. In Germany, only 23 percent thought the US was ahead. That perception is feeding a wider hedging instinct among middle powers.

Chatham House has argued that middle powers are turning to "sovereign artificial intelligence" strategies, drawing on capabilities from both superpowers rather than committing to one, while parts of Europe explore deeper ties with Chinese suppliers to power their own artificial intelligence ambitions. This shift reflects a fundamental change in how countries view the artificial intelligence competition: not as a binary choice between US and Chinese systems, but as an opportunity to build independent capabilities using tools from both sources.

What Happens Next in the Competition?

The next set of proof points will sharpen the comparison between the two systems. Independent benchmarks for GLM-5.2, which Zhipu has yet to publish, will provide clearer data on how Chinese models perform on standardized tests. Whether Huawei meets its 2026 volume targets for artificial intelligence chips will indicate whether China can scale manufacturing to match its software ambitions. Huawei is scaling fast, with Bloomberg reporting plans to produce about 600,000 Ascend 910C chips this year, roughly double last year's output, and Reuters reporting a targeted 60 percent jump in artificial intelligence-chip revenue to around 12 billion dollars.

The most consequential question may be whether tightening US export and access controls slow China down or, as several analysts now suspect, push more of the world toward the open Chinese alternatives. The picture that emerges is of two systems running on different clocks: the US optimizing for frontier capability and the compute that underpins it, where its lead is widest, and China optimizing for cost, openness and deployment, where its gains are spreading fastest and proving hardest to contain.