Delta unveils advanced cooling and power solutions for AI data centers at NVIDIA GTC 2025

Delta is expanding its power infrastructure and cooling technologies for AI data centers, showcasing its latest innovations at NVIDIA GTC 2025. The company’s comprehensive solutions address the growing power and thermal management demands of high-performance computing environments, while integrating with NVIDIA’s platforms to enable more efficient and sustainable data center operations. These developments highlight the increasing importance of specialized infrastructure to support AI workloads as computational demands continue to escalate.

The big picture: Delta is unveiling comprehensive power and cooling solutions specifically designed for AI and high-performance computing data centers at NVIDIA’s GTC 2025 conference.

• The company is showcasing an AI-based smart manufacturing demo line that integrates with NVIDIA’s Omniverse and Isaac Sim platforms for virtual reality synchronization.
• These solutions address the growing power and thermal management challenges that accompany the increasing computational demands of modern AI systems.

Key innovations: Delta’s new offerings span several critical data center infrastructure categories, focusing on maximizing efficiency and performance.

• Power solutions include a new Power Capacitance Shelf with lithium-ion capacitors and a 19-inch 1RU 66kW AC-DC Server Power Shelf that achieves 97.5% efficiency.
• Advanced cooling technologies feature a 1.5MW liquid-to-liquid Coolant Distribution Unit, Cold Plate Loops for NVIDIA GPU/CPU, and a 3D Vapor Chamber with 1000W cooling capacity.
• The company is also introducing specialized passive components, including High Density Power Chokes and Core Shell Liquid Cooling Busbars.

Why this matters: As AI systems become more computationally intensive, the underlying infrastructure must evolve to support their growing power and cooling requirements.

• These specialized power and cooling solutions are essential for organizations looking to deploy advanced AI applications while managing energy consumption and operational costs.
• The integration with NVIDIA’s platforms demonstrates the increasing convergence of hardware and software solutions in the AI data center ecosystem.