In the past 24 hours (December 10 to December 11, 2025), China's AI industry has accelerated from the "pilot period" to the "full application period" around the open source of large models, the scale of AI applications and "AI + finance"; Overseas, it continues to increase model security, data compliance and industry supervision, and at the same time pushes the "AI Creators" to the position of the person of the year, highlighting the special significance of this year in the global technology map.
1. The "2025 AI Top Ten Trends Report" was released: open source AI enters "China Time"
At the "MEET2026 Intelligent Future Conference", the "2025 AI Top Ten Trends Report" was officially released, proposing ten major trends from computing power infrastructure, chip AI to open source ecology, believing that AI is moving from the "tool era" to the "partner era". The report specifically pointed out that China has formed systematic advantages in large models, domestic computing power and open source ecology, and open source large models have become an important technical base for global developers.
2. "AI+" applications are spread across thousands of industries: the scale of China's AI industry has exceeded 900 billion yuan
CCTV reports show that 611 generative AI services have been filed in our country, of which 309 will be newly filed as of early November 2025, more than the whole of last year. The number of AI applications has reached hundreds of scenarios, the scale of mobile terminal users has exceeded 700 million, the number of new users of artificial intelligence comprehensive assistant applications has exceeded 70 million, the total scale of the artificial intelligence industry has exceeded 900 billion yuan, and the number of enterprises has exceeded 6,000, indicating that China's AI is shifting from technological breakthroughs to large-scale industrialization.
3. Alibaba Tongyi Qianwen has become a global "open source base", and Meta's new model has been exposed to distilling Qwen
It has been reported that Meta has chosen to distill third-party models such as Alibaba Tongyi Qwen in the research and development of the new generation of "Avocado" models, and has gradually shifted from open source to a more closed business strategy. According to the article, Qwen has replaced Llama as one of the most active open source large models in the world, and the number of derivative models and downloads have been overtaken, and institutions and enterprises from Europe, America, Southeast Asia, the Middle East and other countries regard it as the preferred base for building AI applications.
4. Shanghai 2025 Financial Development Conference: The industry discusses "AI reconstructing the financial value chain"
At the 2025 Financial Development Conference held in Shanghai, the roundtable forum focused on "How AI Reconstructs the Financial Services Value Chain", and representatives of fund companies, foreign banks, insurance and financial technology companies discussed the practical path of "AI + finance". The participants believed that AI has changed from an optional to a "compulsory course" for financial institutions, deeply embedded in risk management, investment research, operations and customer service, but it also needs to establish industry standards in data governance and algorithm transparency.
5. Regarding the possible restriction of NVIDIA H200 chips, the Ministry of Foreign Affairs responded to maintain its principled position
In response to foreign media reports that "China is considering restricting access to NVIDIA H200 chips", a spokesperson for the Ministry of Foreign Affairs said at a regular press conference that relevant issues have been responded to before, and it is recommended to learn about the specific situation from the competent authorities, and has not positively confirmed or denied the news. This response is seen as maintaining strategic flexibility between key computing resources and industrial security, while leaving room for follow-up policies.
6. OpenAI warns that the network security risk of next-generation models is "high" and will establish a frontier risk committee
OpenAI issued a new statement saying that its upcoming cutting-edge model may have the ability to develop zero-day remote attacks or assist in complex cyber intrusion operations, posing a "high risk" to cybersecurity. The company said it will simultaneously strengthen defensive capacity building, including code audits, vulnerability patching tools, access control, traffic monitoring and other security measures, and announced the establishment of a "frontier risk committee" to invite front-line security experts to participate in assessment and governance.
7. The EU investigates Google's AI training data sources, focusing on web scraping and monopoly issues
Thereare reports that EU regulators have launched an investigation into Google to see if its large-scale scraping of web content to improve the capabilities of large models violates data protection and competition rules. The investigation focuses on whether training data sources are legally authorized and whether they are using search market advantages to improperly expand their AI business, which is seen as the latest move in Europe to further tighten the regulation of platform giants in the era of generative AI.
8. The U.S. House of Representatives held a hearing to examine the opportunities and risks of AI in financial services
TheU.S. House of Representatives Financial Services Committee held a hearing on "The Application of Artificial Intelligence in Financial Services and Housing" to discuss the innovative value and compliance challenges of AI in risk control, anti-fraud and customer service. A number of lawmakers and regulators emphasized the need to adopt a "technology-neutral" and "risk-oriented" regulatory framework to prevent discriminatory algorithms, systemic risks and data abuse while avoiding stifling fintech innovation.
9. The U.S. Department of Health and Human Services released the overall AI strategy, and the FDA launched the "Agent AI" platform
Some legal and industry analysts pointed out that the U.S. Department of Health and Human Services officially released a department-level AI strategy to build a unified roadmap from governance and risk management, infrastructure, talent and scientific research. Its subsidiary FDA has taken the lead in launching a full-staff "agent AI" platform to assist in regulatory affairs, product review, and aftermarket monitoring, emphasizing "human oversight", "optional use" and openness and transparency to balance efficiency and public safety.
10. Time Person of the Year selection "AI Founders", symbolizing the global consensus of the first year of AI
"Timemagazine" announced that it jointly named the "Founders of Artificial Intelligence" as the 2025 Person of the Year, saying that this year is a turning point for "thinking machines" to fully enter public life and make the world both amazed and worried. The reason for the selection pointed out that from basic models to application platforms, AI is rewriting innovation paradigms, industrial patterns and social cognition, and also pushing safety, ethics and governance issues to unprecedented heights.
Frequently Asked Questions (Q&A)
Q: What is the main line of the global AI field in the past 24 hours?
A: The main line can be summarized as "comprehensive industrialization in China, strengthening security and supervision overseas": China continues to accelerate in open source large models, industrial scale and industry landing, while overseas puts model security, data compliance and governance mechanisms in a more core position through hearings, regulatory investigations and industry strategies.
Q: What are the latest highlights of China's AI development?
A: First, the number of registered large models and AI applications has grown rapidly, with an industrial scale of more than 900 billion yuan and more than 6,000 enterprises, indicating that the application has entered an explosive period; second, open source models such as Tongyi Qianwen occupy an increasingly important base position in the global developer ecosystem; Third, key industries such as finance have begun to systematically discuss the practical path and regulatory needs of "AI reconstructing the value chain".
Q: What impact will the new overseas trends around AI security and compliance have on the industry landscape?
A: OpenAI publicly characterizes the cybersecurity risk of cutting-edge models as "high", the European Union investigates Google's data scraping behavior, and the U.S. Congress and federal agencies formulate rules around the use of AI.
Q: What are some short-term opportunities for developers and enterprises?
A: First, the secondary development and industry fine-tuning around the head open source large model, especially in Chinese and multilingual scenarios; second, seize the application landing window in vertical fields such as "AI + finance" and "AI + medical care", and combine industry knowledge with large model capabilities; The third is to lay out "AI infrastructure services" such as model security, data governance, and compliance assessment to provide tool-based solutions for the increasingly stringent regulatory environment.
