Huawei’s Energy Chief Issues a Reality Check: The AI Revolution Will Be Won or Lost on the Power Grid

Huawei Digital Energy President Hou Jinlong warns that energy infrastructure is the ultimate limit of AI growth, advocating for a massive shift toward high-voltage architectures and liquid cooling to support the next generation of AI data centers.

A modern humanoid robot with digital face and luminescent screen, symbolizing innovation in technology.

Key Takeaways

  • 1AI data center (AIDC) installation scale is projected to break records globally as large-scale AI models move to production.
  • 2Energy is identified as the foundational constraint for long-term AI development, summarized by the phrase 'the end of computing is electricity.'
  • 3Technical requirements are shifting toward high-voltage, DC power, and power electronics to accommodate ultra-high-density computing.
  • 4Liquid cooling is now categorized as a 'must-have' solution for the heat generated by intensive AI workloads.
  • 5Data center delivery speed is now a primary metric for ROI and business success in the competitive AI market.

Editor's
Desk

Strategic Analysis

Huawei's focus on 'Digital Energy' represents a calculated strategic pivot. While US-led sanctions have targeted Huawei's access to high-end semiconductors, the company is doubling down on the infrastructure that makes those chips useful: the power systems. By positioning itself as the indispensable architect of the AI power grid, Huawei is carving out a niche that is harder to sanction and increasingly vital. The shift toward liquid cooling and high-voltage DC systems reflects a broader global trend where the 'physics of the data center'—heat and power—have become as important as the 'logic of the algorithm.' For global markets, this signals that the next phase of the AI boom will likely benefit infrastructure and power-management specialists as much as chip designers.

China Daily Brief Editorial
Strategic Insight
China Daily Brief

As the global race for artificial intelligence supremacy accelerates, the industry’s most formidable bottleneck is shifting from silicon to the socket. Hou Jinlong, a director at Huawei and President of Huawei Digital Energy, recently underscored this reality at the 2026 Global AIDC Industry Forum. He predicted that the installation scale of Artificial Intelligence Data Centers (AIDCs) will continue to see explosive growth as massive models and intelligent agents move toward large-scale deployment.

Hou’s central thesis—that "the end of computing power is electricity"—serves as a stark reminder that the digital frontier remains tethered to physical infrastructure. For companies like Huawei, the AIDC is no longer merely a storage facility but a "computing power production system." In this new paradigm, the speed of delivery and the reliability of the power supply are directly correlated with investment returns and business viability.

To sustain this trajectory, Hou argues for a fundamental overhaul of how data centers are powered and cooled. He advocated for a transition toward high-voltage, direct-current (DC) architectures and the widespread adoption of power electronics. These upgrades are essential to support the ultra-high-density loads required by modern AI training, which traditional power grids are currently ill-equipped to handle.

Furthermore, liquid cooling has transitioned from a high-end luxury to an absolute necessity for high-density computing. As thermal loads reach the limits of air-cooled systems, Huawei is positioning itself as a leader in high-reliability liquid cooling and intelligent lifecycle maintenance. This strategic pivot highlights Huawei’s ambition to dominate the "digital energy" space, ensuring that the AI revolution does not stall due to an outdated electrical backbone.

Share Article

Related Articles

📰
No related articles found