Chinese artificial intelligence developers released 16 AI models in a single month, highlighting the rapid pace of development in the country’s technology sector. DeepSeek and Xiaomi were among the companies leading the latest wave of launches, even as some major US-based AI companies have urged developers to place greater emphasis on safety measures. The activity underscores the widening differences between Chinese AI firms and some US frontier laboratories over how quickly advanced systems should be developed.

Anthropic CEO Dario Amodei has argued that AI capabilities are progressing faster than the safeguards designed to control associated risks. He has called on frontier AI companies to adopt a more measured approach to development. However, Chinese AI researchers and industry observers have questioned the argument, with some expressing concern that slowing development could allow leading US companies to maintain an advantage in increasingly competitive AI markets.

The debate is also closely linked to the open-weight approach adopted by several Chinese AI developers. Companies such as DeepSeek have gained attention by releasing models that developers can download, customise and operate themselves, rather than relying entirely on closed AI services. A DeepSeek engineer involved in the development of its V4.1 models reportedly rejected calls for a slowdown, arguing that there was little confidence that companies such as Anthropic or OpenAI would keep advanced AI systems accessible and affordable.

Also Read: Why Did Anthropic Bring A Hindu Monk Into Its AI Ethics Discussions?

Competition is also being fuelled by significant investment. Bloomberg reported that Moonshot AI had completed a funding round at a valuation of around $50 billion, while DeepSeek was reportedly nearing a $12 billion fundraising round. The rapid flow of capital into Chinese AI companies reflects expectations that demand for advanced models will continue to grow, while developers compete to improve performance and expand access to increasingly capable systems.

At the same time, tensions between Chinese AI developers and US technology companies have intensified over alleged misuse of AI systems. Anthropic accused DeepSeek, Moonshot AI and MiniMax in February of conducting model-distillation campaigns involving thousands of accounts and millions of exchanges with its Claude system. OpenAI and US officials have made similar allegations involving Chinese AI companies. Model distillation refers to the process of using the outputs of one AI system to help train another.

Security concerns have added another dimension to the debate. Forbes reported last month that a Chinese-speaking hacker used AI agents, including DeepSeek and Kimi models as well as an older version of Claude, in attacks targeting as many as 100 organisations. Anthropic and OpenAI have cited broader AI safety risks while arguing for greater caution in development. The contrasting positions suggest that the global race for more capable AI models is likely to continue alongside increasingly intense debate over safeguards, security and responsible deployment.

Also Read: Trump Plans White House Dinner With Anthropic CEO Dario Amodei Amid AI Tensions