In the past 24 hours (November 21 to November 22, 2025), many places in China have intensively released policies and achievements around industrial Internet, housing and construction, digital economy and AI applications; Overseas, it continues to speak out on the generative 3D world, AI opportunities in developing countries, the "AI bubble" controversy in the capital market, and AI governance in the public sector.
1. Wuhan released 103 AI application scenarios, and "AI + new industrialization" became the starting point
of the cityWuhan released 103 typical scenarios at the China 5G+ Industrial Internet Conference, covering manufacturing, urban governance and people's livelihood services, and used "scenario list + supply and demand docking" to drive the implementation of computing power and data elements.
2. The top ten AI cases in Shanghai's housing and construction industry were unveiled, and the "cloud engineering model" drove the intelligence of the construction industry
The2025 Shanghai Housing and Construction Science and Technology Conference released the top ten AI applications, of which the "cloud engineering model" and other solutions were implemented in construction site safety, BIM modeling and operation and maintenance management, promoting the upgrading of the whole chain of housing and construction to AI empowerment.
3. Beijing's Dongcheng District has introduced a new policy for the digital economy, increasing the deployment of "artificial intelligence +" and large models
Thenew policy of Dongcheng District provides tiered financial support for computing power applications, large model deployment, industry large models and benchmarking projects, and writes "artificial intelligence +" into the main line of urban digital economy development.
4. China's AI application "Lingguang" exceeded one million downloads in 4 days, and the competition at the application layer accelerated
Ant's full-modal AI assistant "Lingguang" exceeded one million downloads in 4 days after its launch, and together with applications such as Alibaba's "Qianwen", it shifted the focus of competition from model parameters to product experience and daily activity.
5. The IDEA conference emphasizes "from generation to execution", and the Greater Bay Area focuses on embodied intelligence and low-altitude economy
At the Shenzhen IDEA conference, Shen Xiangyang sorted out the evolution of AI from the five dimensions of algorithms, carriers, interactions, architectures and data, and called for making large models into operable systems in "execution-side" scenarios such as embodied intelligence and low-altitude economy.
6. World Bank released an AI report: Middle-income countries have become important users of GenAI but lack infrastructure
According to theWorld Bank's "Digital Progress and Trends" report, middle-income countries have contributed a lot of GenAI traffic and jobs, but high-income countries still control the vast majority of models and capital, and the report calls for narrowing the gap through open source and "small AI".
7. Meta launches WorldGen, which can generate interactive 3D worlds in minutes
TheMeta WorldGen system can generate 3D worlds with collision and navigation meshes from text prompts in minutes and export them directly to engines such as Unity, and is regarded as a "usable" generation tool for games and industrial digital twins.
8. The global market has resumed debate around the "AI bubble", and giants have issued bonds intensively for AI investment
In thecontext of large fluctuations in AI-related stocks, many reports have questioned whether valuations are out of fundamentals; At the same time, technology giants have intensively issued bonds to raise a new round of data center and chip investment, causing the market to worry about the long-term burden of AI capital expenditure.
9. Oklahoma has established a chief AI and technology officer, and the institutionalization of state-level AI governance has been promoted
Oklahoma has appointed a chief artificial intelligence and technology officer who is responsible for coordinating the state government's AI strategy, risk governance and application implementation, reflecting the institutionalized exploration of AI adoption in the public sector at the local level in the United States.
10. AI+drone is used to monitor turkey behavior on farms and verify a new path for intelligent animal husbandry
monitoringPenn State University used a small drone equipped with a computer vision model to identify a variety of turkey behaviors, verifying the feasibility of using "drone+AI" for low-manpower, quantifiable animal welfare monitoring on large farms.
Frequently Asked Questions (Q&A)
Q: How can the main line of global AI development be summarized in these 24 hours?
A: Domestic focus on the trinity of scenarios, policies and applications, while overseas discussions are held on the 3D world generation, AI opportunities in developing countries, and the balance between "AI bubble" and capital expenditure, which is a combination of "landing practice + global differentiation + capital repricing".
Q: Among these new measures in China, what are the most direct benefits to the industry?
A: The Wuhan list, Shanghai housing and construction cases and Dongcheng District policies all point to: use the government to take the lead in clarifying scenarios and financial support, so that it is easier for industries to make replicable AI projects in industries, housing and construction and other rigid needs.
Q: What metrics is currently focused on in the debate around the "AI bubble"?
A: The debate focuses on three points: whether the profit growth rate of the AI business can match the current valuation, whether the huge computing power investment and new debt will drag down cash flow, and how much room is there for the relevant stock price to pull back once the expectation is revised.
Q: From a developer and entrepreneur perspective, what kind of opportunities are more worth paying attention to now?
A: Compared with recreating a general large model, there are two types of directions that are more feasible: one is to dig deep into vertical industry processes like WorldGen and aquaculture monitoring, and the other is to combine the "small AI" mentioned by the World Bank to make tools that can run in local equipment and small teams and truly solve subdivision problems.
