By many accounts, China’s AI industry is closing the gap with the U.S. American export controls have supercharged Chinese domestic innovation, and Chinese AI chatbot models such as DeepSeek now equal or exceed their American counterparts across many performance metrics. As a result, the race for AI dominance is becoming increasingly about not just AI innovation but also global AI diffusion. Recent media pieces highlight China’s position in this domain and show how China’s expanding ecosystem of open-source tools, infrastructure projects, and normative frameworks is enabling it to embed its AI systems—and their underlying values—into global technological networks.
A key lever for China to shape global development and norms around AI is its open-source approach. Ben Dubow wrote for The Diplomat this week, “U.S. AI models still control over 70 percent of the market, but a collaborative, open source approach has enabled Chinese labs to punch far above their weight.” He noted that a growing global community of programmers now works to adapt and improve Chinese models for specific uses, which may accelerate their development. A recent example is the release this month of Alibaba-backed Moonshot AI’s Kimi K2 model, which was enthusiastically received by global developers and recognised by a U.S. ranking as one of the world’s top open-source AI models. At MERICS, Wendy Chang, Rebecca Arcesati, and Antonia Hmaidi published a report this week describing how China’s strong presence in open-source ecosystems helps offset its hardware limitations and accelerate the domestic and global proliferation of its large language models (LLMs), such as DeepSeek:
In models and applications, China is closing in on the US. China is heavily embedded in global open-source communities. Coupled with a protected home market, this has spawned large language model (LLM) developers like DeepSeek. Hardware challenges still hinder wider deployment, but local adoption of LLMs is high, and China’s AI industry is pivoting toward specialized applications.
[…] Results of a test on the Chinese platforms offering DeepSeek illustrate the challenges companies face in supporting top models. The test was conducted by SuperCLUE on 18 Chinese platforms that offered DeepSeek-R1 in February 2025, including both free and paid offerings. The results were poor, showing long response times for questions and high rates of cut-off responses or none at all. The best performers were US firms Perplexity and together.ai, a testament to the importance of access to cutting-edge chips in deploying LLMs effectively. In a follow-up test in March, US platforms disappeared from the chart.
The best performing Chinese host of DeepSeek is VolcEngine, the cloud service from ByteDance. A relative newcomer after the big three – Alibaba, Huawei and Tencent – it is trying to catch up by doubling down on AI offerings. Overall, China’s cloud market is much smaller than that of the US, due in part to lower adoption of public cloud solutions by enterprises. While cloud service providers want to expand by investing in AI infrastructure, enterprises are also embracing private solutions that allow them quick access. The growing popularity of “all-in-one” machines offering DeepSeek models reflects this. In the end, widespread adoption and easy access to LLMs are what will propel the dissemination of AI technology into innovations across different fields. It remains to be seen if China’s divergent approach will yield better results. [Source]
Chinese AI systems have also been integrated into existing infrastructure, embedded within digital platforms that shape global connectivity. This strategy helps reach potential consumers, even those who may be skeptical of Chinese tech. On this point, a global survey this month found that 71 percent of U.S. respondents and 87 percent of E.U. respondents expressed willingness to adopt Chinese LLMs, but 58 percent of the former and 59 percent of the latter said they would only use such AI models hosted on non-Chinese infrastructure. In an essay this week for Global China Pulse, Weidi Zheng described how, rather than challenging Western tech dominance head-on, China expands its global AI footprint more subtly by embedding in existing platforms and tailoring its ecosystem to local contexts:
This essay argues that China’s global information system involves a triangulation of China, the United States, and recipient countries. Much like ‘vines’ that spread by climbing (see Figure 2), Chinese tech companies have built a complex ecosystem comprising digital infrastructures, intermediary platforms, and sectional apps. Crucially, China’s digital expansion relies heavily on an information system dominated by Google, Amazon, Facebook, Apple, and Microsoft (GAFAM), actively interacting with it, and embedding itself deeply into the digital geographies shaped by the social and technological conditions of recipient countries.
[…] China’s global information system reflects the country’s semi-peripheral status: while China has developed its own global platform ecosystem, it remains reliant on the infrastructural power of US big tech at the centre and tends to expand into areas with less competition at the periphery (Li 2021). The platform ecosystem is such a relational, ecological system—a product of global neoliberal capitalism. Chinese tech companies are embedded within, rather than challengers to, this global system.
[…] In China’s domestic market, tech giants such as Alibaba, Tencent, Huawei, and Baidu now dominate the AI ecosystem. These firms benefit from enormous data reserves, digital platforms, and infrastructure, and—like their US counterparts—tend to exclude smaller players. The emergence of DeepSeek offers an alternative to Silicon Valley’s ‘bigger-is-better’ model, which has been marketed as ‘democratisation of technology’. However, it is important to note that DeepSeek remains embedded within the broader platform ecosystems of US big tech due to its lack of infrastructure and applications. To access global markets, DeepSeek has already been integrated into the technology stacks of Microsoft, Nvidia, and Amazon—once again underscoring China’s reliance on the infrastructural power of US tech firms when it ventures beyond its domestic market. In sum, the advent of AI reinforces existing global power structures, further concentrating resources and profits in the hands of the dominant technology players. The initial optimism surrounding AI’s potential to disrupt global power hierarchies now appears increasingly overstated. [Source]
This AI localization strategy is pushing ahead. As highlighted in Lingua Sinica’s “China Chatbot” newsletter, Alibaba Cloud announced this month “the launch of a third data center in Malaysia and a second in the Philippines, following […] others it built this year in South Korea, Mexico, and Thailand, bringing the total number to 33.” Alibaba’s CEO Wu Yongming said the company would triple this number over the next three years and invest $53 billion to expand its network. The company also announced it would launch an AI Global Competency Center in Singapore, where it hopes to partner with 120 universities to train 100,000 students in the use of AI tools.
Ultimately, China’s AI diffusion is seen by both Chinese and American experts as part of a push to reshape the global digital order and contest U.S. technological supremacy. Yasemin Yam wrote this month for Global Voices about how the international spread of Chinese AI models such as DeepSeek, which has built-in political filters that adhere to Chinese censorship norms, reveals how AI diffusion might reshape global discourse to suit state-aligned narratives: “The broader concern is what it means when millions worldwide start depending on AI systems deliberately designed to reflect and reinforce Chinese government perspectives.” At Foreign Affairs, Colin H. Kahl and Jim Mitre expressed concerns about the various facets of U.S.-China AI competition beyond innovation, such as infrastructure, integration across society, and standard setting: “Taking these additional AI races into account makes the United States’ position look precarious.” Meanwhile, Thomas des Garets Geddes and Jordan Schneider at Sinification summarized a recent article by Liu Shaoshan, a leading figure in China’s embodied AI research and a state-designated “high-end overseas talent,” who highlighted diffusion as a key part of China’s path to global AI dominance:
US tariffs and export controls heighten global uncertainty but create a strategic opening for China’s AI industry to expand internationally and reshape “the global technological order”.
Global adoption of US or Chinese technology — not domestic technological prowess alone — is becoming the key battleground for great-power status in AI.
America’s success with TCP/IP’s global rise in the 1980s shows centrally-led government policies, open-source, mandatory standards and talent “exports” can turn national tech into the global default.
Rogers’ diffusion model suggests four steps for China: woo “innovators” with cutting-edge tech, attract “early adopters” through open-source, secure an “early majority” by setting international standards, and reach late adopters through Chinese talent going global.
Thus, China’s first objective should be to match US-level capabilities so that its AI-related technologies are credible and attractive to global “innovators” and “early adopters”. [Source]
