According to Liu Xiangyang, enterprise core competitiveness lies in capability and efficiency. Capability primarily depends on research and development, while efficiency mainly relies on digital intelligence transformation. Efficiency is what keeps enterprises alive, as competitive moats will eventually be filled; therefore, digital intelligence transformation is what preserves enterprise survival.

Liu believes enterprise digital intelligence transformation consists of three stages: The first stage involves using digital systems to express business operations, essentially moving business processes from offline to online, or digitalizing business operations. The second stage uses digital systems to optimize business processes, employing digital methods such as big data analysis and optimization algorithms to enhance business processes and achieve optimized business results. The third stage involves digital systems taking over business operations, where digital systems replace human workers in completing business tasks. The three stages can be summarized with three keywords: expression, optimization, and takeover.

Although everyone discusses AI nowadays, for most enterprises, AI has not brought revolutionary changes because their digitalization foundations are not yet solid. The fundamental skill for AI is digitalization, and the fundamental skill for digitalization is the digital foundation. The importance of a digital foundation to digital business operations is equivalent to the importance of a foundation to a building.

What constitutes a digital foundation? It encompasses all capabilities unrelated to business logic but required by every business system, such as computing platforms, database platforms, and big data platforms. While digital foundations are crucial, building them presents CIOs with extremely challenging difficulties.

Digital foundations can be built either within enterprise data centers, using public clouds, or through a hybrid approach combining both private data centers and public clouds. The problem with building digital foundations within enterprise data centers lies in the difficulty of proper construction, mainly due to five reasons:

(1) Outdated Technology: Most enterprise data centers still rely on computing virtualization technology from the 1990s rather than public cloud technology systems.

(2) Fragmented Products: Various commercial software and open-source software have been procured, creating isolated systems that are difficult to integrate and coordinate systematically.

(3) Poor Stability: Digital foundations with low capabilities and poor architecture inevitably suffer from stability issues. Major IT failures are almost always architectural and systemic problems.

(4) Low Efficiency: Poor foundational capabilities force upper-layer business software teams to handle their own requirements, such as building and maintaining middleware and databases, naturally reducing overall digital department efficiency.

(5) Weak Security: Information security construction is a systemic issue. Poor digital foundation capabilities make it difficult to establish robust information security. For example, without cloud computing's virtual network technology and relying solely on physical networks, isolation between application systems becomes unfeasible.

Does using public clouds solve the digital foundation construction challenge? Unfortunately, it still causes CIO frustration because managing multiple clouds is difficult. Why do many enterprises use multiple clouds? There are various reasons, including horizontal departmental factors (different departments have the authority to choose different clouds), vertical historical reasons (different CIOs in different periods chose different clouds), business overseas expansion needs, financial cost reduction requirements (single cloud dependency makes price negotiation difficult), and business stability requirements.

Multiple clouds create three major challenges:

(1) Public cloud incompatibility issues: Since each public cloud provider wants to lock customers into their single cloud environment, migrating business software from one cloud to another typically requires at least six months of business software modification and adaptation, delaying business launches by half a year.

(2) Cloud silos problem: Business operations and data deployed on different clouds form isolated islands that are difficult to coordinate, making it challenging to achieve multi-cloud active-active configurations.

(3) Multi-cloud operations issues: Different clouds have different configurations and parameters, with increasingly complex parameters making performance optimization increasingly difficult.

**Breaking Boundaries**

At Midea Group Co.,Ltd., Liu is responsible for the entire group's cloud computing capability construction and information security capability building. In cloud computing capability construction, he has built a complete global multi-cloud unified digital foundation at Midea, called the "neutral cloud."

Technical innovations include the following aspects:

(1) Achieved unification between on-premises and cloud environments, unifying digital foundations in enterprise data centers and public clouds;

(2) Achieved multi-cloud unification, implementing unified digital foundations across multiple public clouds;

(3) Achieved software-hardware decoupling in cloud computing;

(4) Realized unified scheduling of cloud and on-premises resources, enabling enterprises to achieve "one global network, one global cloud";

(5) As an operating system-level platform, it shields differences between different vendor public clouds and differences between enterprise data center digital foundations and public clouds;

(6) When application systems migrate from enterprise data centers to any public cloud, or from any public cloud to another, no business modifications are required, whereas the industry typically requires over six months of modification due to public cloud incompatibility.

In information security capability construction, Liu has built a complete information security and compliance system at Midea, proposing the "Eight Security Axes": prevent intrusion, enable detection, prevent leakage, ensure compliance, emphasize operations, maintain regular governance, conduct frequent offensive-defensive exercises, and cultivate talent. In terms of innovation, he collaborated with enterprises to develop the industry's first security large language model.

**Anti-Involution**

The neutral cloud built by Liu at Midea does not compete homogeneously with existing public clouds (neutral cloud provides only software, while existing public clouds provide both software and hardware services), but rather offers complementary advantages and mutual cooperation.

From a complementary advantages perspective, existing public clouds cannot solve multi-cloud problems due to conflicts of interest, while neutral cloud can unify all public cloud resources and all data center resources of an enterprise into one resource pool within one network space, achieving true "one global network, one global cloud."

From a mutual cooperation perspective, neutral cloud can operate on existing public clouds and collaborate with existing public cloud vendors to serve customers.

**About Liu Xiangyang:** Liu Xiangyang currently serves as Chief Information Security Officer (CISO) and Dean of Software Engineering Institute at Midea Group Co.,Ltd. He obtained his Ph.D. in Computer Science from The University of Texas at Austin in 2006. He previously served as Assistant Professor, Tenured Associate Professor, and Tenured Full Professor at Michigan State University, and as Chief Scientist at Ant Group.