24-Hour AI News: OpenAI Relaxes Content Controls, Google Invests $15 Billion in Indian AI Hub

Over the past 24 hours, overseas, OpenAI's content policy and data center investment have accelerated, and Intel has released a timeline for its next-generation AI chips. Domestically, there has been a focus on industrial policy and computing network upgrades, with Shanghai releasing its smart terminal action plan and Huawei launching its AI WAN solution. AI applications in the consumer and biomedical sectors have also been active. The number and importance of AI applications are generally balanced both domestically and internationally.

1. OpenAI announced that it will open mature content to adult users starting in December and provide personalized conversation options

1. OpenAI confirmed that starting from December, mature content and a more flexible conversation style will be open to adult users who have completed age verification.
2. Simultaneously launch more fine-grained "tone and personality" options, emphasizing improving usability while ensuring safety.
3. Minors will not be affected, and the platform will continue to strengthen classification through age identification and protection strategies.

2. Intel announces new data center-class GPU "Crescent Island," focusing on cost-effective inference scenarios

1. Disclosure on October 14: The new GPU is planned to be launched next year, emphasizing "performance per dollar" and energy efficiency, and prioritizing inference and general AI workloads.
2. Official announcements indicate that new products will be released every year to keep up with the industry’s annual iteration rhythm.
3. Against the backdrop of tight supply of key components such as HBM, Intel will enter the mainstream data center AI market with a differentiated architecture.

3. Google will build an AI and data center hub in India with a total investment of US$15 billion

1. Located in Visakhapatnam, Andhra Pradesh, with an investment period of 2026-2030, the project is positioned as one of the largest AI infrastructures in Asia.
2. Cooperate with AdaniConneX and Airtel to provide gigawatt-class computing power and submarine cable landing stations to improve east-west international connectivity.
3. The project will link local R&D with the global computing network to serve the regional expansion of Google Cloud and AI applications.

4. Microsoft will lease Nscale's data center AI capacity in Sines, Portugal

1. The project is located in the Start Campus and is scheduled to go online as early as early 2026.
2. Deploy approximately 12,600 Blackwell Ultra GB300 GPUs to supplement the European and cross-regional AI computing power pool.
3. Continuing the "New Cloud Commerce" cooperation model, Microsoft alleviates self-built supply constraints through third-party capacity.

5. Broadcom releases a new generation of data center network chips, accelerating AI cluster interconnection

1. The new product focuses on high bandwidth, low latency and enhanced optical interconnection capabilities, directly targeting the network bottlenecks of large-scale AI training/inference clusters.
2. During the period of tight supply and demand for AI infrastructure, network capabilities become the key variable in the effective utilization of computing power.
3. Ecological adaptation with GPU/accelerator manufacturers will affect the speed of its penetration in AI data centers.

6. Shanghai releases the "Action Plan for High-Quality Development of the Smart Terminal Industry (2026-2027)"

1. The goal is to exceed RMB 300 billion in industry scale by 2027 and create more than three consumer-grade terminal brands with global influence.
2. Propose the "three 10 million" production capacity goals: annual production of more than 10 million units of AI computers, AI mobile phones, and new AI terminals.
3. Deploy end-side AI chips, AI computers, and edge intelligent computing machines to improve the "end-edge-cloud" collaborative computing network.

7. Huawei releases AI-Centric "AI WAN" solution at UBBF 2025

1. Reconstruct operations, security, and computing network collaboration with "AI-Centric" for the Net5.5G R2 phase.
2. Industry cases show that cross-domain computing efficiency can be increased from approximately 40% to over 95%, and AI operations and maintenance can be used to shorten fault diagnosis and recovery time.
3. Launch new capabilities such as "application-level value management" for operators, emphasizing differentiated experience and room for ARPU improvement.

8. AI-enabled consumption: JD.com reported rapid growth in multiple categories during its 11.11 shopping festival

1. Data on October 14th showed that the transaction volume of intelligent robots increased fivefold year-on-year, AI glasses increased more than three times, AI tablets increased by 200%, 3D printers increased by 120%, and AI mobile phones and AI learning machines both increased by 100%.
2. The platform provides more than 50 AI tools and digital human services such as "JoyAI" to improve merchant efficiency and user experience.
3. On the logistics side, the "super brain + wolf tribe" cluster and digital twin decision-making system have been implemented, improving both order volume and efficiency.

9. State Council Special Study: Promoting High-Quality Development through Standard Upgrading, Emphasizing Digital and AI Empowerment

1. On October 15, the State Council’s 16th special study focused on “standard guidance and guarantee” and proposed to accelerate the construction of a standard system adapted to high-quality development.
2. Emphasize optimizing supply, strengthening implementation and internationalization, and clarify "strengthening the empowerment of digital technologies such as artificial intelligence."
3. Propose to guide the adoption of high-level standards in industrial policies, government procurement, and bidding, and improve the governance mechanism.

10. Report: AI has penetrated all aspects of drug R&D, driving CDMOs' transformation from "cost competition" to "value creation."

1. KPMG released the "Bioscience Pioneers 50" report, stating that AI plays a role in the entire chain including target discovery, synthetic pathways, and clinical design.
2. In areas such as precision medicine, deep model optimization is improving the accuracy of gene editing and accelerating clinical translation.
3. With the coordination of policies and capital, the domestic CDMO market space has expanded, and the regions where innovative drugs "go overseas" have become more diversified.

Frequently Asked Questions (Q&A)

Q: When exactly does OpenAI’s content policy take effect and who does it affect?

A: It will take effect from December for adult users who have completed age verification; minors are still subject to platform classification protection, and it is in line with the industry trend of social platforms imposing PG-13 restrictions on minors.

Q: What is the core focus of Google’s AI hub in India?

A: A five-year investment of US$15 billion will be made in Visakhapatnam. We will work with operators and data center partners to build gigawatt-level computing power and submarine cable landing stations as key AI infrastructure and R&D nodes in Asia.

Q: What does Huawei's "AI WAN" mean to operators?

A: We provide reusable solutions for computing-network collaboration, experience grading, and AI operations and maintenance. According to disclosed cases, cross-domain computing efficiency is expected to increase from approximately 40% to more than 95%, and through application-level operations and autonomous driving networks, operation and maintenance costs can be reduced and ARPU can be increased.

Q: Why is the Shanghai Action Plan seen as a signal for the "AI computer" industry?

A: Clearly define the quantitative goal of "annual production of tens of millions of AI computers", supplemented by the supporting equipment of terminal chips, edge intelligent computing all-in-one machines and intelligent computing server clusters, to form a closed-loop cultivation path of terminals-computing power-applications.

Q: What are the key words in the collaboration between Microsoft and Nscale?

A: "Leasing + Rapid Deployment." Approximately 12,600 GB300 GPUs will be imported into Sines, Portugal, with production expected to begin in early 2026. This will complement the European and cross-regional AI computing power pools and alleviate the long lead time for self-built capacity.