PHENEXS Launches Full-Stack Power Resilience Solution for AI Data Centers

As artificial intelligence (AI) continues to reshape industries worldwide, the rapid expansion of AI Data Centers (AIDCs) is creating unprecedented demand for reliable, resilient, and intelligent power infrastructure. Ensuring stable power delivery has become a critical challenge as AI workloads drive higher power density, faster load fluctuations, and increasing energy consumption.

At the AI Computing Era: Energy Storage Innovation and Application Summit held in Shanghai, PHENEXS introduced its Full-Stack Power Resilience Solution for AI Data Centers, presenting an innovative Power-Computing Synergy architecture designed to support the next generation of AI infrastructure.

Power-Computing Synergy for Next-Generation AI Infrastructure

Traditional power architectures, built around utility power, UPS systems, and diesel generators, are primarily designed to respond to power outages. However, they are not optimized to handle the microsecond- and millisecond-level power fluctuations generated by modern AI computing workloads.

PHENEXS addresses this challenge through a multi-layer power resilience architecture that delivers coordinated energy support across different response times, ensuring continuous, stable operation for high-performance AI data centers.

A Three-Layer Full-Stack Power Resilience Solution

• Cabinet Level – Supercapacitor Energy Support

At the cabinet level, high-power supercapacitors work alongside lithium batteries to provide microsecond-level response. While lithium batteries supply backup energy, supercapacitors instantly absorb rapid power fluctuations, stabilize GPU voltage, and protect critical computing processes during intensive AI workloads.

• Power Distribution Level – High-Rate Battery Systems

For millisecond-to-second load variations, PHENEXS deploys high-rate lithium-ion and sodium-ion battery systems capable of delivering rapid discharge performance. These systems effectively mitigate sudden power spikes caused by AI computing tasks while enhancing system stability, operational safety, and power quality.

• Campus Level – Grid-Forming Long-Duration Energy Storage

At the campus level, grid-forming long-duration energy storage systems provide 2–4 hours of energy capacity to support large-scale load variations. Advanced grid-forming control technology enables voltage and frequency regulation while improving renewable energy integration, peak shaving, and overall grid resilience.

Working together, these three layers establish a comprehensive power resilience framework capable of supporting the demanding operational requirements of AI data centers.

Hybrid Energy Storage Powers the AI Era

During the event, Dr. Zou Qifan, Chief Technology Officer of PHENEXS, shared his insights on the future of energy storage for AI infrastructure. He emphasized that AI data centers require a combination of complementary energy storage technologies rather than relying on a single solution.

Lithium-ion batteries continue to provide cost-effective and commercially mature long-duration energy storage, while sodium-ion batteries offer advantages in high-rate discharge capability and operational safety. Supercapacitors further enhance system performance by responding instantly to ultra-fast GPU power fluctuations, protecting critical computing equipment and ensuring uninterrupted AI operations.

By integrating these technologies into one coordinated architecture, PHENEXS enables a smarter and more resilient power supply system tailored to the evolving needs of AI computing.

Accelerating the Future of Intelligent Energy Infrastructure

PHENEXS believes that the future of energy storage is driven by application-oriented innovation. Through its Power-Computing Synergy strategy, the company is redefining how energy storage supports AI infrastructure—improving power resilience, increasing energy efficiency, and enabling more reliable AI computing.

Looking ahead, PHENEXS will continue to focus on high-value application scenarios, delivering advanced energy storage solutions that empower global AI data centers and accelerate the development of intelligent, sustainable energy infrastructure.

Industry Recognition

During the summit, PHENEXS was honored with the 2025–2026 Aurora Award Top 10 Grid-Side Energy Storage Benchmark Projects for its Miaoling 1GW/2GWh Shared Energy Storage Demonstration Project.

This recognition highlights PHENEXS' expertise in utility-scale energy storage and reinforces its commitment to delivering innovative solutions that support the global transition toward a smarter and more resilient energy future.