Huawei Unveils Innovative Interactive Network Strategy for AI Data Centers

On May 15, 2026, Huawei launched its strategy for Interactive AI Data Centers (AIDC) during the Global AIDC Industry Summit held in Dongguan, China. The event gathered nearly 1,000 leaders, technical experts, and key ecosystem partners from sectors such as energy, smart processing, and telecommunications. Participants shared insights on the evolution of next-generation AI data center architectures and the latest technological innovations, witnessing the unveiling of Huawei's new strategy.

Hou Jinlong, a member of Huawei's board and President of Huawei Digital Power, delivered a keynote speech highlighting the burgeoning AI industry and the substantial energy demands generated by widely adopted large language models and numerous AI agents. According to him, electricity is essential for data processing and serves as the foundation for the long-term growth of AI. The integration of processing and energy is deepening, gradually constructing frameworks that connect new energy systems with AI infrastructure. Reliable power supply is crucial for the sustainable development of AIDC, ensuring quality operations and network-friendly functions for long-term, stable data center performance.

In the face of significant industry transformation opportunities, Huawei Digital Power aims to lead innovations in AIDC and become a long-term trusted strategic partner. By integrating technologies related to bits, watts, heat, and batteries (the 4T), the company has developed comprehensive competencies in renewable energy generation, network formation, high-density computing power supply, liquid cooling, and synergies between processing and energy.

Bob He, Vice President of Huawei Digital Power, also addressed attendees, discussing the need for an AIDC strategy based on innovations featuring a "3+1" model. He noted that the global AI boom is causing a surge in token demand, indicating a shift towards a token-centric AIDC era. However, the diverse and high-density processing of data poses challenges related to power density, scaling, and load fluctuations. As renewable energy's share in the energy mix increases, the associated load fluctuations further compound AIDC reliability challenges.

To address these challenges, Huawei's strategy aims to create reliable, energy-efficient, rapidly deployable, and grid-friendly AIDC solutions to maximize token production per watt.

The innovation in energy management involves a reimagining of power systems transitioning from grid structures to integrated chip architectures. The diversity in load requirements will necessitate the coexistence of AC and DC power supplies in AIDC operations for a significant period. Huawei plans to develop a multi-input and multi-output (MIMO) power architecture, comprising grid-friendly uninterruptible power supplies (UPS) and energy storage systems (ESS) that evolve towards integrated energy routers based on semiconductor transformer technology. This architecture leverages Huawei's expertise in power electronics, network shaping, and management of high and low voltage power supply to meet AIDC demands dynamically.

Furthermore, thermal management innovations enhance operational reliability throughout the entire lifecycle, covering everything from chips to external environments. The advancement in liquid cooling systems is not merely about cooling hardware; it encompasses a holistic thermal management strategy that extends the lifespan of devices by maintaining effective and reliable heat dissipation. Huawei's AI-powered MW-class liquid cooling systems push the boundaries of large-scale cooling adoption, focusing on long-term reliability rather than mere availability.

The digital transformation of data centers through AI technologies allows for greater visibility and operational efficiency throughout the lifecycle of the systems, ensuring reliability at every stage and facilitating efficient maintenance practices. Additionally, construction innovation will transform the investment realization model through engineering, prefabrication, and modularization, permitting design, production, testing, and verification processes to occur in factories, with only assembly needed on-site. This approach significantly reduces time to market, enhances execution quality, and enables rapid product replication.

Bob He emphasized that the ultimate capability of AI computing is limited by the available energy supply, aligning with the principle of "energy as a token." In this evolving AIDC landscape, traditional energy efficiency metrics based on Power Usage Effectiveness (PUE) fail to capture the full value of data centers. The industry must transition from efficiency metrics to value metrics. To facilitate this, Huawei proposes the TokEnergy Index—a ratio measuring energy efficiency against token generation in AIDC contexts.

As we move further into the AI era, data centers are not just server rooms; they are super factories for token production. Positioned at a pivotal moment in the sector's development, Huawei Digital Power is committed to continued investments in core technologies, fostering leading-edge interactive solutions in AIDC to drive high-quality growth, maximize tokens per watt, and propel the AI era forward.