Compal–Verda Partnership Aims to Scale AI Infrastructure in Europe and APAC

The global race to build AI infrastructure is accelerating as cloud providers, hardware manufacturers, and data‑center operators invest heavily in high‑performance GPU clusters. A new strategic partnership between Taiwan’s Compal Electronics and Finland‑based AI cloud provider Verda illustrates this trend, aiming to expand advanced AI computing capacity across Europe and the Asia‑Pacific region.

At the same time, large GPU purchases—such as Datasection Inc.’s order for hundreds of Nvidia B300‑based servers—signal how rapidly demand for AI training and inference infrastructure is growing.

A Partnership Focused on AI Infrastructure Expansion

On May 7, 2026, Compal Electronics and Verda announced a strategic partnership to accelerate the deployment of next‑generation AI infrastructure in Europe and APAC. Under the agreement, Compal will supply GPU server systems that support Verda’s cloud platform, which is designed for frontier‑scale AI workloads such as advanced model training and agent‑based inference services .

Verda, headquartered in Helsinki, operates AI‑focused cloud infrastructure and provides high‑density GPU computing resources for organizations developing large AI models and applications. The partnership is intended to expand the company’s computing footprint in regions where demand for localized AI capacity is rapidly rising .

By pairing Compal’s hardware expertise with Verda’s cloud infrastructure operations, the companies aim to accelerate the rollout of AI data‑center capacity tailored for large‑scale machine‑learning workloads.

The Role of High‑Density Liquid‑Cooled GPU Servers

A central component of the collaboration is Compal’s high‑density, liquid‑cooled GPU server platform. These systems are engineered to support demanding AI workloads, including large‑scale model training, high‑throughput inference, and agent‑based applications that require high concurrency and long‑context processing .

Liquid‑cooling technology is becoming increasingly important for modern AI data centers. GPU clusters used for training large models consume large amounts of power and generate substantial heat. Liquid‑cooled architectures help:

• Improve thermal efficiency and cooling performance
• Enable denser GPU deployments in limited data‑center space
• Maintain stable performance during extended training runs
• Reduce energy usage and operational complexity

These capabilities allow data‑center operators to run large AI workloads continuously while managing the power density and thermal constraints that come with modern GPU hardware.

Datasection’s Nvidia B300 Server Purchase Signals Rising Demand

A separate but related development highlights the broader market demand for this type of infrastructure. Japan‑based Datasection Inc. announced an agreement to acquire 635 GPU servers equipped with Nvidia’s B300 processors from Compal Electronics, representing a total of 5,080 GPUs and an investment of roughly $325 million .

According to the company, the systems will serve as the core infrastructure for its next‑generation AI data‑center project. The goal is to provide large‑scale computing capacity capable of meeting the needs of major cloud providers and AI developers building generative‑AI systems .

Such large cluster purchases demonstrate how AI operators are locking in GPU supply to ensure sufficient computing power for training increasingly complex models and delivering inference services at scale.

Why Dense GPU Infrastructure Matters for Generative AI

Both the Compal–Verda partnership and Datasection’s GPU purchase point to the same industry reality: AI computing capacity has become a critical bottleneck.

Modern generative‑AI systems require:

• Massive GPU clusters for training large language and multimodal models
• High‑throughput inference infrastructure for real‑time applications
• Scalable architectures that can support agent‑based or autonomous AI systems

High‑density GPU servers combined with efficient cooling allow operators to deploy these clusters more quickly and efficiently, enabling AI services that demand large computational resources.

The Bigger Picture: Building the Next Generation of AI Data Centers

Together, these developments illustrate how the AI infrastructure ecosystem is evolving. Hardware manufacturers like Compal are increasingly supplying specialized AI server platforms, while cloud providers such as Verda deploy those systems to build large GPU clusters for enterprise and AI‑model developers.

Meanwhile, organizations like Datasection are investing hundreds of millions of dollars in GPU hardware to construct new AI data centers capable of supporting generative‑AI workloads and next‑generation applications .

Public details about the exact deployment timelines and data‑center locations for the Compal–Verda infrastructure expansion remain limited. However, the partnership reflects a broader industry push to scale AI computing capacity globally as demand for training, inference, and agent‑driven AI systems continues to surge.