Digital Infrastructure Fund — Data Centres & Fibre

Digital Infrastructure Fund: The Physical Backbone of the AI Economy

Data centres and fibre optic networks have quietly become among the most critical — and most defensible — infrastructure assets in the global economy. Every streaming video, cloud application, financial transaction, and AI model training run requires physical infrastructure: rack space in a cooled, secure building connected to high-capacity fibre networks. As artificial intelligence workloads drive an unprecedented surge in compute demand, the data centre and connectivity infrastructure needed to support that demand has become one of the most compelling long-term investment themes of the decade.

This AIFMD-regulated fund invests in European data centre campuses and regional fibre optic networks, targeting a net IRR of 9–12% over an eight-year fund life.

The AI-Driven Demand Surge

Data centre demand was already growing before the generative AI wave of 2022–2026. Cloud computing migration, video streaming, and fintech growth were driving sustained capacity additions across European markets. The emergence of large language models and AI workloads has materially accelerated this demand.

Training a frontier AI model requires months of continuous compute time across tens of thousands of specialised processors. Serving AI applications at scale requires dedicated inference infrastructure that must be available 24/7/365 with near-zero downtime tolerance. The result is that major AI companies and cloud hyperscalers are now signing multi-decade lease commitments for large-scale data centre capacity — long before the facilities are built.

This contracted demand profile transforms data centres from speculative assets into infrastructure assets with utility-like revenue certainty.

Asset Focus

Hyperscale data centres: Large-scale campuses (typically 50MW–200MW of IT load capacity) purpose-built to house compute, storage, and networking infrastructure for cloud and AI tenants. These facilities are leased on long-term triple-net contracts — tenants pay base rent plus all operating costs, typically with annual CPI escalation clauses.

Wholesale colocation: Mid-scale data centres where corporate enterprises and growing technology companies lease dedicated space, power, and cooling rather than building their own facilities. These assets serve a broader tenant base than hyperscale, reducing single-tenant concentration.

Long-haul and regional fibre: Backbone fibre routes connecting major European cities and data centre clusters. Fibre assets generate long-term wholesale revenue from telecoms operators and data centre interconnects. Once laid, fibre has very low ongoing maintenance cost and long asset life.

Edge data centres: Smaller distributed facilities close to population centres, enabling the low-latency processing required by autonomous vehicles, industrial IoT, and real-time AI applications. A structural growth segment as edge computing requirements expand.

Revenue Structure

The fund's target assets share a common feature: very long-term contracted revenues. Data centre leases with hyperscalers typically run 15–20 years with break clauses available only to the tenant (making them similar to institutional property leases). Fibre indefeasible rights of use (IRUs) are typically sold for 20–30 year terms.

This long contracted revenue provides:

• Predictable cash flow for investor distributions from years 2–3
• Natural inflation protection through CPI escalation clauses
• Low operational risk relative to development-stage or uncontracted assets

Risk Considerations

Technology obsolescence risk: Data centres require regular capital investment to remain at the technological frontier. Air-cooled facilities may require significant upgrade or replacement as liquid-cooling and direct-chip-cooling become the standard for AI workloads. The fund focuses on assets capable of accommodating evolving cooling technologies.

Tenant concentration: Hyperscale data centres may have a single major tenant. The fund's diversification across multiple assets and tenant types mitigates but does not eliminate the risk of tenant non-renewal at lease expiry.

Power and grid risk: European data centre development is constrained by grid capacity limitations in key markets. Securing grid connections is a critical development risk, and energy cost inflation affects operating margins in power-purchase agreements not fully hedged.

Regulatory risk: European data protection, energy efficiency regulation, and planning policy for large-scale data centre development are evolving rapidly. New requirements could increase operating costs or constrain development pipeline.

Liquidity: Eight-year closed-end fund with no early redemption mechanism. Investors commit for the full fund life.

Suitability

This fund is appropriate for sophisticated investors with long investment horizons seeking infrastructure returns from the structural growth in digital economy infrastructure. It suits investors who can commit capital for eight years and who understand the technology and operational risks specific to data infrastructure assets.

How to Invest

Contact our investment team to receive the fund's KIID, prospectus, and current portfolio development pipeline documentation. Full investor categorisation under AIFMD required before subscription. Minimum investment €150,000.

Important: Capital is at risk. Past performance is not a guarantee of future returns. This is for information purposes only and does not constitute a personal recommendation. Seek independent financial advice before investing. This fund illustrates the type of digital infrastructure opportunity Global Investments advises on — it is not a live investment offer.