Explained: Pantheon AI's US$58bn Data Centre Campus

2026-05-06 · Source: AI Magazine · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure · Depth: Intermediate, short

Summary

Pantheon AI, an investment group, is developing a hyperscale AI data center and innovation campus in Topusko, Croatia, with an initial EU€12bn (US$14bn) construction phase starting in early 2027 and full operations by Q1 2029. The total investment, including tenant equipment, is projected to reach EU€50bn (US$58bn). This 310-acre facility, designed to NVIDIA GW-Scale AI factory standards, aims to address Europe's data center capacity gap, particularly for AI workloads. It will feature up to 1 GW total capacity with 800 MW IT load, supported by 5.2 GW of renewable energy, including a 500 MW on-site solar plant and 8,000 MWh battery storage. The campus will also integrate robust connectivity via four independent fiber routes and the GreenMed subsea cable.

Key takeaway

For CTOs and VPs of Engineering evaluating European AI infrastructure, this US$58bn Pantheon AI campus in Croatia signals a significant new option for hyperscale compute. You should consider this facility's NVIDIA GW-Scale standards, 800 MW IT load, and integrated renewable energy as a potential solution for deploying demanding AI training and inference workloads, especially given existing capacity constraints in established European hubs.

Key insights

A US$58bn hyperscale AI campus in Croatia aims to close Europe's AI compute capacity gap.

Principles

• Hyperscale AI requires gigawatt-scale power and advanced cooling.
• Redundant power and fiber are critical for Tier IV AI operations.

Method

The project integrates large-scale compute capacity above Tier IV availability, supported by 5.2 GW renewable energy, a 500 MW on-site solar plant, 8,000 MWh battery storage, and four independent 400 kV lines.

In practice

• Deploy high-density AI workloads requiring sustained power.
• Utilize on-site renewable energy and battery storage.
• Connect via multiple independent fiber routes for low-latency.

Topics

• Pantheon AI Campus
• Hyperscale AI Data Centers
• Europe's AI Capacity Gap
• Renewable Energy Infrastructure
• NVIDIA GW-Scale Standards

Best for: CTO, VP of Engineering/Data, AI Architect, Director of AI/ML, Investor

Related on AIssential

• Can France’s Stargate-style buildout succeed? — France's ambitious €51bn AI compute buildout faces skepticism regarding its alignment with Europe's diverse AI infrastructure needs.
• Solstice and Tensorx to Buy $1 Billion in AI Infrastructure to Support EU Sovereign AI Demand — Solstice and TensorX are financing EU sovereign AI infrastructure via a \$1 billion fund and a new yield asset, aiUSX.
• SpaceX Is Spending $2.8 Billion to Buy Gas Turbines for Its AI Data Centers — SpaceX is investing heavily in dedicated energy infrastructure for its AI data centers.
• World’s Largest AI Datacenter — $100B Disaster — Building leading AI models now demands extreme infrastructure decisions, prioritizing compute, power, and rapid deployment.
• Nebius plans $10 billion AI data center in Finland near Russian border — Strategic data center location optimizes energy costs and leverages renewable power for AI infrastructure.
• Verda raises $117M to scale AI cloud infrastructure built on clean Nordic power — Verda secured $117M to scale its AI cloud infrastructure, emphasizing renewable energy and vertical integration.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Magazine.