While the AI industry fixated on Anthropic's latest frontier models and Washington's regulatory battles, a quiet but profound shift has been occurring in the development landscape. Chinese open-weight models accounted for 41% of downloads on Hugging Face this spring, surpassing US models. On OpenRouter, the top six most popular models are all open models from Chinese firms including Tencent, Xiaomi, DeepSeek, MiniMax, and Z.ai, while Anthropic's Claude Opus 4.7 trails in seventh place. Data from Vercel shows that open-weight models handled nearly a third of AI requests on the platform in June, absorbing the bulk of infrastructure-heavy AI workloads, with closed models operating as high-cost premium layers.
Chinese open models conquer Hugging Face
The meteoric rise of open models is no passing trend. Hugging Face, the platform hosting almost three million public models and one million public datasets, sees a new repository created every seven seconds. CEO Clem Delangue has emphasized that enterprises increasingly prefer ownership over renting. "If you are an AI company or a technology company, you do not want to outsource your core capabilities to a black box API that you do not control," he said. This shift is underscored by the fact that half of all Fortune 500 companies use Hugging Face to deploy their own private or open source models.
Sponsored Protocol
Enterprises choose ownership over renting
The pivot to open models is driven by cost and control considerations. Delangue noted that after seeing the bills from scaling closed frontier models, companies are re-evaluating the benefits of owning their own AI. Satya Nadella, Microsoft CEO, recently warned against single-provider lock-in, arguing that data control should be paramount. "If learning flows in only one direction, economic value converges toward the owners of the learning infrastructure rather than the creators of the knowledge itself," Nadella said, advocating for distributing learning infrastructure to every firm.
Sponsored Protocol
The rise of open models coincides with powerful releases from Chinese labs. Z.ai, based in Beijing, recently launched GLM-5.2, an open-weight model that excels at agentic coding and competes with Anthropic's latest models in identifying security vulnerabilities. This demonstrates that open models are not just cheap alternatives but true innovation drivers.
The safety debate around open models
The growth of open models intensifies the debate over their availability. Dario Amodei, CEO of Anthropic, argues that releasing increasingly capable open models could be dangerous because once out, they are hard to control. Conversely, Delangue contends that the biggest risk is concentration of power. "You do not make it safe by keeping it behind closed doors for just a few players," he said, adding that transparency allows defenders to "patch cybersecurity risks that open source models already exploit."
Sponsored Protocol
The debate is far from settled, but one thing is clear: the AI race is no longer only at the frontier. Companies seeking to remain competitive must weigh the costs and benefits of open models, as evidenced by the choices of major enterprises. Tools like 1Password tracking AI spending in real time or Canva Code 2.0 democratizing AI website creation illustrate the evolving ecosystem. For further reading, refer to the Wikipedia page on Hugging Face.
