分类： technology

China has officially commissioned a groundbreaking national network infrastructure that demonstrates unprecedented data transmission capabilities, marking a significant advancement in global information technology. The China Environment for Network Innovation (CENI), the country’s first national science and technology infrastructure in the ICT sector, recently passed national acceptance in Nanjing and is now operational.

This revolutionary network testbed achieved what previously seemed impossible: transferring 72 terabytes of data generated by China’s Five-hundred-meter Aperture Spherical radio Telescope (FAST) from Guizhou province to Huazhong University of Science and Technology in Hubei province in merely 1.6 hours. Traditional internet methods would have required approximately 699 days to accomplish the same data transfer.

According to Liu Yunjie, academician of the Chinese Academy of Engineering, CENI represents China’s entry into the world’s advanced echelon in network technology innovation. “CENI’s passage of national acceptance enables China to deliver higher-quality services for all industries to embrace AI,” Liu stated.

The infrastructure boasts large-scale, programmable, virtualized, and open-sharing features, providing a simple, efficient, and low-cost test environment for researching innovative architectures of future networks. The development team has pioneered a service-customized network architecture internationally and created the world’s first distributed large-network operating system.

After over a decade of construction, CENI now spans 40 cities nationwide with a total optical transmission length exceeding 55,000 kilometers. The network can support 128 heterogeneous networks and 4,096 parallel heterogeneous service tests while seamlessly interconnecting with existing internet infrastructure and global test facilities.

The project has generated substantial intellectual property, including 206 international and domestic standards, 221 authorized invention patents, and 139 software copyrights. Multiple achievements have received national-level scientific and technological awards.

Wu Hequan, another academician of the Chinese Academy of Engineering, emphasized that CENI will support the development of 5G-Advanced and 6G technologies while allowing research institutions and enterprises to test various new technologies.

Notably, CENI offers dramatic improvements for AI development. Liu explained that “training a large model with 100-billion-level parameters across domains requires over 500,000 rounds of iterative development. With CENI, each training round takes only about 16 seconds,” resulting in significant time and economic savings.

Looking forward, CENI will provide open testbed support to key industries including manufacturing, energy and power, education, healthcare, and the low-altitude economy, thereby accelerating China’s digital transformation initiatives.

In a landmark advancement for neurotechnology, Chinese researchers have demonstrated unprecedented brain-machine interface (BMI) capabilities enabling paralyzed patients to interact with their physical environment through thought alone. Two individuals with high-level paralysis have successfully operated power wheelchairs, directed robotic dogs to retrieve objects, and manipulated robotic arms to perform tasks like drinking water—all without physical movement.

The achievement, announced by the Center for Excellence in Brain Science and Intelligence Technology of the Chinese Academy of Sciences, represents a significant leap beyond previous BMI applications that were limited to virtual interactions. The patients, both in their 30s, received minimally invasive neural implants developed through collaboration between the CAS center, Shanghai Huashan Hospital, and industry partners.

Technical innovations were crucial to this breakthrough. Researchers overcame substantial challenges including neural data compression, environmental noise interference, and maintaining ‘cross-day stability’ through neural manifold alignment techniques. Most impressively, the team achieved an end-to-end delay of under 100 milliseconds from neural signal to device execution—faster than the body’s natural neural transmission speed.

The system’s practicality is enhanced by its cost-effective design and minimal invasiveness. The implant requires only a 5-millimeter cranial puncture, the smallest reported globally, and is approximately half the size of Neuralink’s device. According to lead scientist Zhao Zhengtuo, the technology is progressing toward clinical applications with the goal of making implantation as simple as ear piercing.

Future development timelines anticipate motor and language function restoration within three years, sensory restoration and treatment for neurological disorders within five years, and highly minimally invasive systems for both medical and consumer applications within a decade. All procedures underwent rigorous ethical and technical reviews, with patients reported in good health throughout the trials.

In a landmark advancement for neurotechnology, Chinese researchers have demonstrated unprecedented brain-machine interface (BMI) capabilities enabling two paralyzed patients to perform complex physical tasks through thought alone. The breakthrough achievements—announced December 17, 2025, by the Center for Excellence in Brain Science and Intelligence Technology of the Chinese Academy of Sciences—include the world’s first mental control of a power wheelchair and robotic dog for delivery retrieval, plus precise manipulation of a robotic arm for drinking activities.

The two patients, both males in their thirties suffering from high-level paralysis, received minimally invasive BMI implants developed through collaboration between the CAS center, Shanghai Huashan Hospital, and corporate partners. The implant procedures—conducted in June and October 2025 through a mere 5-millimeter cranial puncture—represent the smallest such implementation globally, with the implant device itself being approximately half the size of Neuralink’s comparable technology.

This research signifies a quantum leap beyond the team’s previous March 2025 achievement where a patient mentally controlled a computer cursor. The new cases enable three-dimensional interaction with the physical environment, dramatically expanding possibilities for self-care, employment, and social participation. One patient has already secured employment as an intern product sorter, utilizing brain control to perform online data annotation for AI verification systems in vending machines.

Technical innovations were crucial to these advancements. Researchers developed high-compression, high-fidelity neural data compression techniques and created hybrid decoding models capable of extracting usable signals in noisy environments. By addressing ‘cross-day stability’ challenges through neural manifold alignment, the team maintained reliable decoder performance despite environmental interference and physiological variations.

Perhaps most impressively, the system achieves end-to-end delay of under 100 milliseconds from neural signal acquisition to device execution—faster than the body’s natural neural transmission speed. Patients describe the experience as intuitive and seamless, comparable to controlling video game characters without conscious effort.

According to lead scientist Zhao Zhengtuo, the research represents significant progress toward practical clinical applications. The team anticipates scalable applications for restoring motor and language functions within three years, breakthroughs in sensory restoration and neuropsychiatric disorder treatment within five years, and highly minimally invasive systems enabling both medical and consumer uses within a decade. Future goals include achieving sufficiently fine control for activities such as playing the piano with mind-controlled robotic fingers.

In an unprecedented strategic maneuver, global AI giants OpenAI, Google, and Perplexity have launched an aggressive campaign to capture India’s massive user base through complimentary premium subscriptions. This calculated initiative represents more than mere user acquisition—it constitutes a crucial effort to gather diverse multilingual training data from the world’s most populous nation.

Market intelligence from Sensor Tower reveals staggering adoption metrics: ChatGPT now commands 73 million daily users in India, more than double its U.S. user base, following a remarkable 607% year-on-year surge. Similarly, Google’s Gemini has experienced a 15% growth spike since November, reaching 17 million Indian daily users compared to just 3 million in the United States.

The strategic free offerings are substantial: Google provides its $400 Gemini AI Pro subscription at no cost for 18 months to 500 million Reliance Jio subscribers. OpenAI has made its ChatGPT Go plan completely free for one year across India, previously priced at $54 locally. Perplexity has similarly waived its $200 annual Pro subscription fee for Airtel telecommunications users.

Behind these generous offers lies a critical data acquisition strategy. Five AI analysts confirm that India’s linguistic diversity presents a unique training opportunity for AI models. The complex communication patterns, multiple languages, and distinctive dialects characteristic of Indian users provide invaluable data that current training sets lack.

Sagar Vishnoi, co-founder of Future Shift Labs, explains: ‘These complimentary initiatives effectively address gaps in AI training datasets that currently lack comprehensive information on regional user behavior patterns.’

The approach leverages India’s exceptional digital landscape: 730 million smartphone users consuming 21 gigabytes of data monthly at just 9.2 cents per gigabyte—among the world’s most affordable mobile data rates. This price sensitivity has historically responded well to free access models, as demonstrated by Reliance’s successful customer acquisition strategies.

While users like Hyderabad PhD student Anees Hassan enthusiastically utilize these tools for research and presentation tasks, concerns regarding data harvesting practices persist. Hassan acknowledges implementing opt-out features to prevent his data from contributing to AI training, highlighting the privacy considerations accompanying these free access programs.

In a landmark development for artificial intelligence advancement, a coalition of China’s premier academic institutions and technology enterprises has established the Artificial Intelligence Open Alliance in Guangzhou. This collaborative initiative brings together 17 leading universities—including Tsinghua University, Peking University, Shanghai Jiao Tong University, Zhejiang University, and Beijing University of Posts and Telecommunications—alongside eight prominent technology corporations and research organizations.

Operating under the guidance of the Ministry of Education’s Department of Science, Technology, and Informatization, this international, open, and nonprofit platform represents a strategic effort to consolidate resources for AI innovation and practical implementation. The alliance will serve as a coordinating body for major national AI research initiatives while developing an independent, controllable AI infrastructure system.

Tsininghua University President Li Luming, who chairs the alliance, emphasized that this collaboration constitutes a significant measure to address national strategic requirements and establish leadership in global AI development. The alliance will concentrate on several core objectives: transforming educational methodologies, constructing innovative talent cultivation systems, organizing cutting-edge technological research, developing autonomous infrastructure, and facilitating international exchanges.

Five specialized committees have been constituted to advance these goals, with leadership distributed among Tsinghua University, Peking University, Shanghai Jiao Tong University, The Chinese University of Hong Kong, and The Hong Kong University of Science and Technology. These committees will address critical domains including AI in education, research applications, fundamental engineering, ethical governance, and international cooperation.

The alliance secretariat will be headquartered at Tsinghua University, providing administrative coordination for this comprehensive effort to build integrated application innovation, coordinated technology infrastructure, and interconnected governance systems.