Skip to main content

We use cookies to enhance your experience. By continuing to visit this site you agree to our use of cookies. Cookie Policy

AI in ASIA
News

China's AI Models Outstrip US in Downloads Annually

Chinese AI models officially surpass US downloads on Hugging Face for the first time, marking a historic shift in global AI development.

Intelligence DeskIntelligence Deskโ€ขโ€ข4 min read

AI Snapshot

The TL;DR: what matters, fast.

Chinese AI models captured 17.1% of global downloads vs US 15.8% on Hugging Face

DeepSeek R-1's launch catalyzed China's dominance in open-source AI development

Alibaba's Qwen family surpassed Meta's Llama with 700 million downloads globally

China Rewrites the AI Playbook as Downloads Eclipse US for First Time

One year after DeepSeek R-1's launch, China has fundamentally altered the global open-source AI landscape. Chinese-developed models now dominate downloads on Hugging Face, officially surpassing those from the United States for the first time in history.

This milestone represents more than market share statistics. It signals a strategic shift that's reshaping how artificial intelligence develops globally, with implications stretching from Silicon Valley boardrooms to emerging markets across Africa and Asia.

Chinese AI
China's AI models have surpassed US downloads globally, marking a historic shift in the open-source landscape

Hugging Face, in collaboration with the Massachusetts Institute of Technology, recently published an analysis titled "The One-Year Anniversary of the DeepSeek Moment." The report revealed that Chinese open-source AI models captured 17.1% of global downloads between August 2024 and August 2025, edging out the US share of 15.8%.

Advertisement

Giants like Baidu, ByteDance, and Tencent have dramatically increased their open-source contributions. Baidu alone moved from zero releases to over 100 models on the platform within 12 months.

The DeepSeek Catalyst That Changed Everything

DeepSeek R-1's release in January 2025 dismantled three critical barriers that had previously limited open-source AI adoption:

  • Technical democratisation: It made advanced reasoning capabilities accessible as downloadable, fine-tunable assets rather than proprietary black boxes.
  • Licensing liberation: The permissive MIT licence allowed swift integration into production environments, accelerating real-world applications.
  • Psychological transformation: It shifted industry mindset from "can this be done?" to "how can we execute this effectively?"

The ripple effects are unmistakable. Even prominent American models like Deep Cogito v2.1, released in November 2025, now build upon Chinese foundations. This represents a fundamental reversal from traditional AI development patterns where Chinese AI models now lead global token rankings.

Alibaba Cloud's Qwen family has achieved over 700 million downloads, outstripping Meta's Llama to become the most popular open-source AI system globally. This success extends beyond download metrics into practical deployment across diverse markets.

By The Numbers

  • Chinese AI models grew from 1.2% to 30% of global usage in 12 months, a 2,400% increase
  • China released 1,509 large language models by July 2025, accounting for 40% of all global LLM releases
  • Alibaba's Qwen achieved 600 million downloads worldwide by 2025
  • China's core AI industry reached over 1.2 trillion yuan ($173.9 billion) in 2025
  • 515 million Chinese users adopted generative AI by June 2025, forming the world's largest AI user market

Western Leaders Sound the Alarm

The shift hasn't gone unnoticed in boardrooms across Silicon Valley and Washington. Microsoft President Brad Smith recently voiced concerns that US AI companies are falling behind their Chinese counterparts outside Western markets.

We have to recognise that right now, unlike a year ago, China has an open-source model, and increasingly more than one, that is competitive. Brad Smith, President, Microsoft

Smith specifically highlighted DeepSeek's rapid expansion in Africa and other emerging markets, where Western platforms often face affordability challenges. A Microsoft research report estimates DeepSeek holds an impressive 89% market share within China.

The competitive pressure has sparked strategic responses. The US launched the ATOM (American Truly Open Models) initiative in July 2025, backed by industry leaders including Hugging Face CEO Clement Delangue and OpenAI's chief strategy officer Jason Kwon. This initiative explicitly cites Chinese AI momentum as its primary motivation.

Li Lecheng, Minister of Industry and Information Technology, recently emphasised China's strategic position:

Chinese-developed AI models are gaining global traction as China ranked first worldwide in downloads of open-source AI models over the past year. Li Lecheng, Minister, Ministry of Industry and Information Technology, China
Timeline Chinese AI Milestone Global Impact
January 2025 DeepSeek R-1 launches with MIT licence Triggers open-source AI acceleration
July 2025 China releases 1,509 LLMs (40% global share) US launches ATOM initiative in response
August 2025 Chinese models hit 17.1% download share Surpasses US models for first time
November 2025 Qwen reaches 700M+ downloads Becomes world's most popular open-source AI

Investment Surge Follows Technical Breakthroughs

The "DeepSeek shock" has catalysed significant investment flows into Chinese AI startups. MiniMax raised $619 million in its Hong Kong IPO, with shares doubling on their first trading day. This success story reflects broader confidence in China's war of a hundred models.

Zhipu AI followed suit, raising $558 million and becoming the first major Chinese large language model developer to go public. These funding rounds signal institutional belief that Chinese AI development will continue challenging Western dominance.

While DeepSeek currently holds 4% of the global chatbot market share compared to ChatGPT's 68% and Google Gemini's 18% (according to Similarweb data), its trajectory suggests rapid growth potential. The company's forthcoming V4 model, anticipated in mid-February 2026, reportedly outperforms both Claude and GPT series in code generation based on internal benchmarks.

Recent developments like Moonshot AI's valuation surge and Tencent's T1 reasoning model launch demonstrate the breadth of Chinese AI innovation beyond DeepSeek's initial breakthrough.

Frequently Asked Questions

Why did Chinese AI models suddenly surpass US downloads?

DeepSeek R-1's January 2025 launch with permissive licensing triggered widespread adoption. Its technical capabilities combined with accessibility democratised advanced AI reasoning, creating momentum for other Chinese models to gain global traction.

What makes Chinese open-source models competitive with Western alternatives?

Chinese models like Qwen and DeepSeek offer comparable performance to Western counterparts but with more permissive licensing, lower costs, and better optimisation for non-English languages, making them attractive for global developers.

How are US companies responding to this competitive pressure?

The US launched the ATOM initiative to promote American open-source models, while companies like Microsoft have publicly acknowledged the competitive threat and adjusted their strategies for emerging markets where Chinese models dominate.

Which markets are driving Chinese AI model adoption?

Beyond China's domestic market, African and Asian emerging markets show strong adoption due to affordability factors. Chinese models also perform well in markets where English isn't the primary language for AI applications.

What's next for the global AI landscape?

DeepSeek's V4 model launching in February 2026 could further accelerate Chinese AI adoption. The competition is driving innovation globally, with both Chinese and Western companies investing heavily in open-source development and emerging market expansion.

The AIinASIA View: This milestone represents more than market dynamics; it's a fundamental shift in AI development philosophy. Chinese companies have embraced radical openness while Western firms remain largely proprietary. This divergence creates strategic advantages for Chinese models in price-sensitive markets and developer communities. We expect this trend to accelerate as how AI reasoning models actually think becomes more democratised. The real question isn't whether Chinese models will maintain their lead, but how quickly they'll expand it across different AI applications and geographical markets.

The implications extend far beyond download statistics. As Hugging Face concluded in their analysis, "The world is still reacting, sparking a new wave of open-source enthusiasm." This shift indicates a broader movement towards diverse AI development that moves beyond Western-centric models and embraces global innovation.

What's your prediction for how this shift will reshape the AI industry over the next two years? Drop your take in the comments below.

โ—‡

YOUR TAKE

We cover the story. You tell us what it means on the ground.

What did you think?

Share your thoughts

Join 3 readers in the discussion below

This is a developing story

We're tracking this across Asia-Pacific and may update with new developments, follow-ups and regional context.

Advertisement

Advertisement

This article is part of the China's AI Regulatory Model learning path.

Continue the path รขย†ย’
Loading comments...

Privacy Preferences

We and our partners share information on your use of this website to help improve your experience. For more information, or to opt out click the Do Not Sell My Information button below.