Exascale Labs and Compal Signed GPU Platform Deal

The companies have agreed to develop a U.S.-based GPU system platform featuring advanced power architecture.

Updated on Oct. 2, 2026 in Data Centers

Isometric editorial illustration of a modular server rack and power unit, representing data center AI infrastructure development. — Exascale Labs and Compal Electronics have announced a partnership to develop a US-based GPU system platform, focusing on 800-volt power architecture for AI data centers. AI Illustration. Upload story photo >

Live Poll

Do you believe new AI data center technologies will make high-density AI more accessible to you?

Exascale Labs and Compal Electronics have signed a Letter of Intent to build a new platform for GPU systems within the United States. This collaboration aims to validate native 800-volt direct current power across a complete data center system.

Why it matters

This partnership targets the development of high-density AI infrastructure by integrating specialized power and cooling technologies. The collaboration seeks to create a modular, scalable framework suitable for the growing demands of modern artificial intelligence workloads.

The system utilizes an 800-volt direct current architecture combined with solid-state transformer technology to manage power. Compal provides the GPU servers and liquid cooling systems, while Exascale supplies the modular data center design and software.

The players

Exascale Labs

Exascale Labs is a technology firm specializing in modular data center systems and advanced power architecture design.

Compal Electronics

Compal Electronics is a Taiwan-based global manufacturer of notebook computers, tablets, and server hardware.

The details

The platform functions as a comprehensive testing environment where electrical, thermal, and mechanical interfaces can be validated before full deployment. This integration of modular data center design and power architecture is intended to support high-density AI scaling.

Exascale Labs and Compal Electronics announced their partnership agreement on October 2, 2026.

The Tech Race

As data centers face increasing pressure to host massive AI models, developers are pivoting away from legacy power distribution toward high-voltage direct current systems. This shift enables the cooling and power density required to operate the next generation of GPU-accelerated computing infrastructure.

For users and developers, this project aims to create more efficient and scalable hardware environments for AI. Over time, advancements in power and cooling efficiency may lower operational hurdles for deploying high-density computing resources.

The takeaway

The move toward 800-volt direct current power signals a critical shift in how data centers manage high-density workloads. Efficiency in power delivery and liquid cooling remains the primary bottleneck for the scaling of AI infrastructure.