In the past 24 hours, from the International AI Scientist Conference to the "AI + Brain-Computer" Industry Promotion Conference, new achievements have been intensively released in China, and a number of Chinese large model platforms have entered the global authoritative report for the first time; Overseas, there have been fierce discussions around Gemini 3, the super-large computing power project "Stargate" and the US stock AI bubble, and the game between AI development and regulation and antitrust has become more prominent.
1. The International Conference of Artificial Intelligence Scientists released a number of achievements and annual lists
The2025 International Conference of Artificial Intelligence Scientists opened in Beijing, releasing a series of achievements in the direction of scientific research agents and AI-enabled basic science, and announcing the annual cutting-edge technology list and future breakthrough forecasts. Experts at the meeting emphasized that a new generation of large models and agents will reshape the scientific research paradigm and promote "AI for Science" from concept to infrastructure.
2. Beijing released "AI + brain-computer" innovation achievements and data platforms
At the Beijing "AI + Brain-Computer" Industry Innovation Work Promotion Conference, a number of brain-computer interface achievements were unveiled, including the "AI+Brain-Computer Data" platform, the multimodal basic large model of brain science and the visual reconstruction system. By opening up cross-institutional brain data and providing general models, Beijing is laying out brain-computer interfaces and artificial intelligence as a new round of cutting-edge industrial highlands.
3. Gartner released the first AI application development platform report, and Chinese manufacturers were collectively shortlisted
Gartner, a world-renowned consulting agency, released the 2025 "Magic Quadrant" report for AI application development platforms, with Google, Amazon, and Microsoft in the leader quadrant, and Volcano Engine, Alibaba Cloud, and Tencent Cloud entering the challenger quadrant. According to the report, China's large model platform is accelerating its catch-up in terms of developer ecology, computing power services and multi-scenario implementation, and is expected to form stronger competitiveness in the enterprise application market.
4. Beijing-Tianjin-Hebei promotes the "AI + Xinchuang" chain group ecology, aiming for a scale of 3 trillion
yuan Thenewly released regional plan proposes that by 2027, the scale of the Beijing-Tianjin-Hebei new generation of information technology application innovation clusters will be expanded to 3 trillion yuan, and the key directions include the integration of artificial intelligence, cloud computing and Xinchuang software and hardware. The local government is building a safe and controllable "AI base" through computing power centers, basic software localization and industry large model pilots.
5. The "Four Sisters of Shenzhen Innovation" showcase the AI trump card, two-wheel drive between finance and industry
At theCHTF and the Golden Expo, a number of Shenzhen technology and financial institutions focused on displaying the implementation of AI in chip design, industrial manufacturing and financial services, forming a linkage of "hardware + software + capital". Shenzhen is embedding AI capabilities into key links in the manufacturing and financial industries through industrial funds and demonstration projects, enhancing the city's voice in the global innovation chain.
6. Google released Gemini 3 and released a roadmap for "1,000 times performance"
Technologymedia reported that Google officially launched the Gemini 3 series model and claimed that it would achieve "1,000 times higher computing power performance at the same cost and energy consumption" in the next four to five years. This aggressive goal means that cloud vendors will continue to invest heavily in AI chips, self-developed accelerators, and data center efficiency, which will further raise the technical threshold for large model competition.
7. OpenAI, Nvidia and Oracle jointly built "Stargate" and encountered antitrust doubts
Anumber of foreign media disclosed that OpenAI, Nvidia, Oracle and other companies are working together to build a super-large computing power facility called "Stargate", aiming to provide the foundation for stronger general-purpose models in the future. Antitrust scholars warn that this alliance at the chip, cloud and model levels may conflict with the spirit of antitrust laws for more than 100 years, raising concerns about a "new trust in the AI era".
8. The discussion on the AI bubble in the US stock market has shifted from "whether it exists" to "how it will evolve" A
number of comments at home and abroad have pointed out that there is a significant divergence between the valuation and actual profitability of the AI sector in the US stock market, and the characteristics of the AI bubble are becoming clearer. Some studies believe that the bubble may not burst immediately, but it is highly concentrated in a few giants, and the market capitalization is overly dependent on stories and forward expectations, making interest rate changes and earnings stall possible tipping points.
9. AI monitoring software spreads in remote work, causing privacy and labor disputes
Accordingto technology media reports, more and more companies are deploying "boss software" with AI analysis functions in remote and hybrid work teams to monitor employee keyboards, cameras and application usage in real time. Proponents claim it helps improve efficiency and detect anomalous behavior, but privacy groups and unions argue that the practice exacerbates workplace anxiety and erodes employee autonomy.
10. The game industry embraces generative AI, and Ubisoft says the change is "comparable to 2D to 3D"
The management of a well-known game manufacturer said that the company has fully integrated generative AI in programming, level design and NPC dialogue, believing that the development efficiency and experience improvement it brings is comparable to the industry's transformation from 2D to 3D. At the same time, manufacturers also emphasize that the main narrative line and values are still controlled by artificial creation to alleviate players' worries about "AI writing plots".
Frequently Asked Questions (Q&A)
Q: In the past 24 hours, how can the main line of domestic AI development be summarized?
A: On the one hand, there are concentrated breakthroughs in scientific research and cutting-edge technologies, from the International AI Scientist Conference to the "AI + Brain-Computer" platform, reflecting the transformation to "AI for Science" and brain-intelligence integration; on the other hand, it is the systematic layout of large model platforms, information innovation ecology and urban industrial clusters at the industrial and regional levels, laying the foundation for subsequent application landing and computing power infrastructure.
Q: What do these overseas moves mean for Chinese AI companies and developers?
A: Projects such as Gemini 3 and "Stargate" show that global cloud vendors and model companies are still accelerating the stacking of computing power and model scale, and the threshold continues to rise. For domestic teams, it is more difficult to be rigid directly on the "universal maximum model", and the more realistic path is to rely on local data and scenarios to do industry models, tool chains and platform services, and actively pay attention to the new regulations on privacy protection and anti-monopoly overseas.
Q: What does the latest debate around the AI bubble remind investors and entrepreneurship?
A: The current market is increasingly accepting the judgment that "a local bubble already exists", and the focus of discussion has shifted to who can survive the volatility and whether the cash flow and moat are solid enough. Whether it is the primary or secondary market, the risk of projects that rely solely on "storytelling" and long-term imagination space is rising, and more attention is paid to technical barriers, verifiable business models and real cost reduction and efficiency improvement capabilities.
Q: Judging from this news, which AI directions are more worthy of the attention of ordinary people and practitioners in the future?
A: Several signals are relatively clear: first, long-term tracks such as scientific research intelligence and brain-computer interfaces, which are suitable for people with scientific research backgrounds; second, AI application development platforms, information and innovation ecosystems and regional industrial clusters will continue to release the demand for engineers and solution providers; Third, content industries such as games, education, and media are accelerating the testing of generative AI, providing creators with new tools and forcing everyone to improve their ability to "collaborate with AI".
