China Mobile Unveils AI Trusted Computing: Addressing the Core Security Challenges of Cloud AI

A significant announcement was made at the main forum of the 2026 China Computing Conference on September 12. China Mobile officially launched its new security architecture, dubbed "AI Trusted Computing" (AITC). This release directly targets the fundamental trust deficit that enterprises, particularly in data-sensitive sectors, face when adopting cloud-based artificial intelligence.

Beyond Compute: Crafting a Unified AI Security Lifecycle

AI Trusted Computing is not a single tool but a systematic security framework. Its central aim is to merge the convenience of public cloud with the security and control of a private cloud, enabling clients to confidently deploy their most critical data and AI models in the cloud.

The solution is built upon China Mobile Cloud's "fully localized, heterogeneous computing infrastructure." This implies autonomy and control from the hardware layer upward, establishing a solid foundation for security. On this base, AITC deeply integrates three core technologies:

  • Confidential Computing: Ensures data remains encrypted even during processing in memory, preventing cloud providers or potential attackers from accessing sensitive information.
  • Domestic Cryptographic Technology: Employs nationally compliant cryptographic algorithms to build all-link encryption protection covering transmission, storage, and computation.
  • Information Flow Privacy Protection: Strictly monitors and safeguards data flow throughout the entire AI pipeline, preventing unauthorized diffusion of information during model training or inference.

"Universal Confidential Computing" and "Confidential Tokens"

To achieve these goals, AITC offers two key service modules. "Universal Confidential Computing" provides unified, confidential computing-protected resources for diverse AI workloads (like large-scale training or high-concurrency inference), allowing users to maintain performance and flexibility without compromising security.

The "Confidential Token" acts as a secure pass, enabling precise control and auditing of every access attempt to confidential data and model assets. It ensures only authorized and trusted operations can proceed, effectively safeguarding an enterprise's most valuable digital property.

Who Stands to Benefit?

The application scenarios for this technology are clear-cut. Sectors like finance, government, industry, healthcare, and academic research have long been plagued by critical questions: Will core customer data, medical records, or research findings be leaked during cloud processing? Could expensive, proprietary AI models be copied or stolen? Is the cloud computing environment truly trustworthy and free of backdoors?

China Mobile's AI Trusted Computing seeks to answer these questions systematically. It aims to deliver a one-stop security barrier covering the "entire lifecycle of cloud-based AI training, inference, and data usage," thereby lowering the barriers and alleviating concerns for these vital industries as they move toward intelligence and cloud adoption.

As data becomes a core factor of production, securely leveraging it to create value is a global imperative. China Mobile's launch represents not just a product upgrade, but an exploration of a path toward AI democratization that meets high-standard security requirements.