In the past 24 hours (April 8, 2026 to April 9, 2026), the AI industry has continued to update at a high frequency. The domestic focus is on the implementation of regulatory rules, organizational adjustments of leading companies, upgrading of open source models and changes in computing power prices; Abroad, the main line of the industry is still "model capability improvement + infrastructure expansion + commercialization realization".
1. Ten departments issued new regulations on ethical review of artificial intelligence technology
The Ministry of Industry and Information Technology, the Cyberspace Administration of China and other ten departments jointly issued the "Measures for Ethical Review and Service of Artificial Intelligence Technology (Trial)". The new regulations further clarify the requirements for ethical review, expert review, risk control, and liability traceability for AI projects. For enterprises, this means that the compliance threshold for AI R&D and implementation is being refined, and high-risk scenarios will place more emphasis on pre-review in the future.
2. Alibaba established a group technical committee to upgrade the Tongyi system
Alibaba announced organizational adjustments around AI, established a new group technical committee, and upgraded Tongyi Laboratory to Tongyi Large Model Division. This adjustment puts the model, cloud infrastructure and inference platform in a more core position, releasing a signal that Alibaba will further concentrate its resources to fight the tough battle of AI. The competition of large factories is shifting from product launch to organizational and resource mobilization.
3. Zhipu released and open-source GLM-5.1, emphasizing long-term independent implementation
Zhipu officially launched GLM-5.1 as a new generation of flagship open source model. This model focuses on programming and long-distance task processing, emphasizing the ability to perform continuously and autonomously for longer periods of time in a single task. For the domestic developer ecosystem, the continued expansion of open source high-capability models will help accelerate the implementation of agents and engineering applications.
4. Tencent Cloud raised the price of AI computing power-related services
Tencent Cloud announced an increase in the regular prices of AI computing power, container services, and EMR-related products. Previously, Alibaba Cloud and Baidu Intelligent Cloud have successively adjusted their prices, and now Tencent has followed up, indicating that the tight supply of domestic AI computing power and rising hardware costs have been transmitted to the cloud server. The industry has moved from "price competition" to "resource pricing", which will also force enterprises to pay more attention to reasoning efficiency and cost control.
5. Meta released Muse Spark, turning to a new round of model competition
Meta unveiled Muse Spark, the first model of its superintelligence team after a restructuring. The model positions native multimodal reasoning and takes a closed-source path, indicating that Meta is shifting from an open-source narrative to a greater emphasis on product control and commercial fulfillment. For the competitive landscape of overseas large models, this means that Meta is trying to regain its leading position.
6. Intel and Google expand AI chip cooperation
Intel and Google announced an expanded collaboration focused on promoting CPU and IPU collaboration for AI inference and deployment. As the industry shifts from training boom to large-scale inference, the importance of general-purpose CPUs has risen again. This change also shows that the competition for AI computing power is no longer just a battle for GPUs, but a battle for the efficiency of the whole system.
7. Meta and CoreWeave have added large-scale cloud cooperation
Meta and CoreWeave have reached a new long-term partnership worth $21 billion. The cooperation will further support Meta's large-scale AI training and inference needs, and also highlight the continuous rise of the position of professional AI cloud service providers in the industrial chain. The deep binding between large model companies and computing power service providers is becoming the norm.
8. Amazon disclosed that AWS's annual AI revenue has exceeded $15 billion
Amazon said that AWS's annualized AI revenue has exceeded $15 billion. This data shows that the AI business of cloud vendors is gradually entering the "cashing period" from the "investment period". For the market, the demand for enterprise-level AI is still expanding, especially around model hosting, inference services and development platforms.
9. Anthropic launched the AI network security project Glasswing
Anthropic announced the promotion of the network security project Glasswing and joined forces with several technology companies to test the application of stronger models in defense security scenarios. The project emphasizes leveraging advanced models to discover software vulnerabilities and enhance defenses while controlling high-risk capability spillovers. AI security has moved from principle discussion to actual industrial collaboration.
10. OpenAI was exposed to accelerate the exploration of advertising business models
The latest news shows that OpenAI is showing investors a larger advertising revenue expectation. If the subsequent progress goes well, advertising could become an important source of revenue beyond subscriptions and APIs. This shows that leading AI companies are looking for a more stable and scalable commercialization path with high computing power investment.
Frequently Asked Questions (Q&A)
Q: What is the most prominent main line in the AI industry in the past 24 hours?
A: The most prominent main line is the simultaneous advancement of computing power, model and commercialization, and industry competition has entered the stage of "paying equal attention to capability upgrading and cost management".
Q: What is the most noteworthy change in domestic news?
A: First, the ethical review rules are further refined, second, leading companies are accelerating the restructuring of AI organizations, and third, the price adjustment of cloud computing power reflects that resource scarcity is intensifying.
Q: What signals have been released by foreign dynamics?
A: Overseas giants are betting on new models, chip collaboration and cloud infrastructure at the same time, indicating that the focus of competition is shifting from single-point model capabilities to full-stack system capabilities.
Q: What are the enlightenments for developers and startups?
A: On the one hand, the ability of open source models continues to improve, and the innovation threshold is decreasing; On the other hand, computing power and compliance costs are rising, and refined product positioning will be more important.
