RM新时代有限公司

RecentlyThe national standard "Network Security Technology - Security Framework for Artificial Intelligence Computing Platforms" was officially released. This standard is managed by the National Cybersecurity Standardization Technical Committee, led by Huawei Technologies Co., Ltd., with CloudWalk Technology as one of the core drafting units, deeply participating in the standard formulation.GB/T 45958-2025"Network Security Technology - Artificial Intelligence Computing Platform Security Framework" was officially released. This standard is managed by the National Cybersecurity Standardization Technical Committee, led by Huawei Technologies Co., Ltd., and CloudWalk Technology, as one of the core drafting units, deeply participated in the standard formulation.

This national standard, which will be officially implemented on a specific date, establishes a systematic security framework for AI computing platforms, specifying security functions, security management, and role security responsibilities, effectively addressing new security threats such as hardware vulnerabilities, training data leakage, and adversarial sample attacks.2026Year2[Month]1The national standard officially implemented on that day establishes a systematic security framework for AI computing platforms, specifying security functions, security management, and role security responsibilities, effectively addressing new security threats such as hardware vulnerabilities, training data leakage, and adversarial sample attacks.

From Wuhan Practice to National Standard: CloudWalk's "Controllable Training Ground" Experience Recognized"Controllable Training Ground" Experience Recognized

As a leading enterprise in China's AI field, CloudWalk Technology integrated its practical experience at the National Network Security Base into the standard-setting process.2025Year6In a certain month, CloudWalk Technology signed a project with Wuhan Linkonggang Development Zone for the National Network Security Base Large Model Training and Inference Integration Project. This project, in deep cooperation with Huawei, aims to build a leading security foundation in the Central China region.AIsecurity foundation.

Zhang Li, Vice President of CloudWalk Technology, introduced that the "Controllable Training Ground" built by the project is a large-scale, safe and controllable model training factory based on trusted data space, and also a security testing center integrating quality inspection, detection, and filing, which exactly aligns with the national deployment of "strengthening the supply of computing power, algorithms, and data.""Controllable Training Ground" is a large-scale, secure and controllable model training factory based on trusted data space, and also aAISecurity Testing Center integrating quality inspection, testing, and filing, exactly aligning with the national deployment of "strengthening the supply of computing power, algorithms, and data".

At the Network Security Base Training and Inference Integration Center, CloudWalk Technology has created a safe and controllable model training environment, effectively solving the high costs, high barriers, and potential security risks faced by current large model training. This practice has become a secure foundation for the large-scale implementation of cutting-edge technologies such as spatial computing.

CloudWalk Technology has long been involved in AI security standardization. During the standard-setting process, CloudWalk contributed the practical experience from the National Network Security Base Training and Inference Integration Project.

The "Controllable Training Ground" built by the project relies on trusted data space technology (in line with the privacy computing international standard in which CloudWalk participates) to achieve full-process security control of large model training. It also integrates CloudWalk's attack and defense detection capabilities accumulated in biometric anti-spoofing national standards, forming a "training-detection-filing" integrated security closed loop, and continuously follows up on technical standard research in areas such as large model capability requirements and evaluation methods. "Controllable Training Ground" relies on trusted data space technology (in line with theIEEEprivacy computing international standard that CloudWalk participated in), achieving full-process security control of large model training, while integrating CloudWalk's attack and defense detection capabilities accumulated in biometricfeaturerecognition anti-counterfeiting national standard (GB/T 38671, GB/T 41772), forming a "training-testing-filing" integrated security closed loop, and continuously following up on technical standard research projects in areas such as large model capability requirements and evaluation methods.

CloudWalk Technology has always regarded security as the cornerstone of AI development. The company's technical team has accumulated rich experience in preventing new security threats such as model theft, data leakage, and adversarial sample attacks. This experience has been integrated into the technical specifications of the national standard.

Under the guidance of the national standard, CloudWalk Technology is building a more complete security industry ecosystem in Wuhan. By constructing an industrial innovation complex of "government guidance, enterprise leadership," gathering the three core forces of "scenario parties (demand), capability parties (technology), and service parties (implementation)," it promotes safe and controllable collaborative innovation and achievement transformation.AIsecurity industry ecosystem. By building an industrial innovation complex of "government guidance, enterprise leadership", gathering "scenario parties (demand)"+capability parties (technology)+service parties (deployment)" three core forces, promoting secure and controllable collaborative innovation and achievement transformation.

Responding to the National Strategy: Safeguarding the "Artificial Intelligence+" Action"Artificial Intelligence+Action Escort

7[Month]31Recently, the State Council executive meeting reviewed and approved the "Opinions on Deepening the Implementation of the 'Artificial Intelligence+' Action," clearly requiring the strengthening of computing power, algorithm, and data supply, building an open-source and open ecosystem to provide strong support for industrial development and growth.+Action Opinions", clearly requiring strengthening the supply of computing power, algorithms, and data, building an open-source ecosystem to provide strong support for industrial development and growth, and proposing 2025The AI industry scale is expected to reach one trillion yuan.5The goal of trillions of yuan.

CloudWalk Technology's practice shows that trusted data spaces and controlled training grounds are precisely the core hubs supporting"Artificial Intelligence+the achievement of the action goals. Through the innovative model of the network security base training and inference integration center, the company providesAIa replicable "foundation paradigm" for the safe, deep integration and large-scale application of technology in various industries.

CloudWalk Technology is anchoring with security to help ChinaAISeize the commanding heights of the global track.

The release of this national standard marks a key step in the construction of China's AI security system, laying a solid foundation for the healthy and sustainable development of AI technology.

In the future, CloudWalk Technology will continue to leverage itsAItechnical advantages and practical experience in the security field, promote the implementation of security standards, and build a trustworthy, reliable, and controllable security barrier for the intelligent transformation of thousands of industries.