Hugging Face Recaps "DeepSeek Moment": How Has China's AI Changed the Global Open-Source Landscape Over the Past Year?

On January 20, Hugging Face, the world’s largest AI open-source community, published an in-depth article—"The First Anniversary of 'DeepSeek Moment'"—detailing how Chinese AI has reshaped the global open-source ecosystem over the past year.

The article points out that the release of DeepSeek R-1 in January 2025 became a crucial turning point for the industry, lowering the barriers for technology and application. It was not only a watershed for the development of AI in China but also sparked profound changes in the global open-source model, propelling Chinese models to rise in both downloads and influence worldwide.

In the past year, tech giants such as Baidu, Alibaba, Tencent, and startups like Moonshot have significantly increased their investment in open-source. The download volume of Chinese models on Hugging Face has now surpassed that of the US. Despite Western attempts to seek alternatives, many global startups and researchers are increasingly relying on Chinese-developed open-source models as their foundation; Chinese AI is now deeply embedded in the global supply chain.

“DeepSeek Moment”: Breaking Three Barriers

As the most important collaboration platform for global AI developers, Hugging Face’s article first reviews the market situation before the release of DeepSeek R-1. The article describes: "Before R1, China's AI industry mainly focused on closed-source models... For most companies, open-source was not the default choice." However, the emergence of R-1 changed this landscape.

The author of the article believes that the true significance of R-1 lies not in whether it was the strongest model of its time, but in its lowering of three key barriers.

The first is the technical barrier—R-1 transformed advanced reasoning capabilities into downloadable and easily fine-tuned engineering assets, “reasoning began to behave like a reusable module.”

The second is the adoption barrier—The MIT license allowed the model to quickly enter production environments, and community discussions shifted from “which model scores highest” to “how do we deploy it.”

Most importantly, the psychological barrier was eliminated. As the article stated: “When the question shifts from ‘Can we do this?’ to ‘How can we do this well?’, decisions at many companies changed.” Hugging Face commented in the article that this release “gave China’s AI development something extremely valuable: time,” proving that even with limited resources, rapid progress can be made through open-source and fast iteration.

And the release of DeepSeek R-1 has turned “open-source” from merely a tactical choice into the long-term strategy for Chinese tech companies. The article emphasizes that in the past year, China’s AI development model underwent a fundamental transformation, rapidly shifting from a closed-source early phase to an open-source, dominant phase.

Giants Enter & Strategic Reorganization

As open-source moved into the mainstream, Chinese tech companies' strategies have changed markedly. Hugging Face’s article points out that compared to 2024, the period after R-1’s release has witnessed a new landscape in China’s AI world: “Large technology companies lead, startups follow, and vertical industry companies are increasingly entering the field.”

Giants such as Baidu, ByteDance, and Tencent—as well as startups like Moonshot—have joined in, making the top-ranking models on Hugging Face no longer monopolized by American developers.

The article cites data to support this trend: Baidu’s releases on Hugging Face rose from zero in 2024 to over 100 in 2025; ByteDance and Tencent’s releases increased eight to ninefold. Additionally, Moonshot’s release of Kimi K2 is regarded as “another DeepSeek moment.”

The article analyzes that the focus of competition has shifted from individual model performance to ecosystem. Taking Zhipu AI’s GLM and Alibaba’s Qwen as examples, these companies not only released model weights but also built engineering systems and ecosystem interfaces. The article bluntly states: “At this stage, simply comparing raw model performance is no longer enough to win. Competition increasingly centers on ecosystem, application scenarios, and infrastructure.”

“Collaboration Under Constraints” & Market Dominance

Hugging Face further introduces a compelling perspective in the article: The rise of the collective of Chinese AI players is not driven by protocol coordination, but by constraints.

The article writes: “What looks like cooperative behavior is better understood as alignment under shared technical, economic, and regulatory pressures.” Under the joint stress of limited computing power and cost control, companies began competing along similar technical foundations and engineering routes. This homogeneity allows the ecosystem to possess self-replicating and expanding capabilities.

This strategy has produced significant results in market data. The article reveals: “Among newly created models (<1 year), downloads of Chinese models have surpassed those from any other country, including the United States.” Heatmap data from Hugging Face also shows that from February to July 2025, open-source releases by Chinese companies became considerably more active.

For the market, this means China’s AI industry has evolved from a pure model parameter competition into a contest of system-level engineering capability with more commercial landing potential.

Global Response: Coexistence of Dependency and Catch-up

The article ends by analyzing the global market’s response to the rise of Chinese AI. Although institutions in the US and France (such as Mistral) are accelerating the launch of open-source models to stay competitive, Chinese models' influence has already permeated to the foundational level.

Hugging Face discloses a key fact: “Startups and researchers worldwide using open-weight models often default to, or even rely on, models developed in China.” The article cites the US leading open-weight model released in November 2025, Deep Cogito v2.1, which is in fact a fine-tuned version of DeepSeek-V3.

Meanwhile, the article also mentions that DeepSeek is widely adopted in markets such as Southeast Asia and Africa, with its multilingual support and cost advantages being key factors for enterprise adoption. Facing this situation, the US launched the ATOM (American Truly Open Model) project, which explicitly cited the strength of DeepSeek and Chinese models as motivation, calling for coordinated efforts in open-weight model development.

The article concludes: “The world is still responding, igniting a new surge of open-source enthusiasm.” Looking ahead to 2026, Hugging Face expects more significant releases from both China and the US, with architectural trends and hardware choices becoming focal points of the next stage.

Risk Warning and DisclaimerThe market involves risk; investments need caution. This article does not constitute personal investment advice, nor does it consider the special investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article fit their particular circumstances. Investments made accordingly are at your own risk.