Powering the AI Boom: Managing Risks in the Next Wave of Data Centre Growth

Introduction

Generative AI has triggered an unprecedented surge in data‑centre construction. Hyperscale cloud providers and sectors like entertainment, banking and healthcare are building facilities globally. Aon predicts more than $1.3 trillion will be invested in data centres over the next five years. However, this AI-driven expansion brings intertwined risks: soaring energy demand, supply‑chain bottlenecks, regulatory complexity and geopolitical uncertainties. This post unpacks those risks and outlines strategies for sustainable growth.

Key Risks in the AI-Driven Expansion

• Energy supply constraints. Global data‑centre power consumption is projected to double by 2026. Meeting AI workloads requires unprecedented energy, and sourcing power is a key priority. Land and power supply constraints, coupled with construction delays, create bottlenecks.

• Extreme heat and climate stress. Facilities must handle higher ambient temperatures and increasing heatwaves. Cooling costs and energy use rise, and new technologies (e.g., liquid cooling, immersion systems) are becoming necessary.

• Regulatory and geopolitical complexities. Countries are wary of over-reliance on infrastructure in risk-prone regions. Natural disasters, privacy laws and export controls (e.g., US restrictions on AI chips) affect supply chains and site selection.

• Supply‑chain disruptions. Shortages of semiconductors, storage devices and cooling systems delay projects and increase costs. Trade policies and export controls exacerbate constraints pwc.com.

• Talent shortages. Scaling AI infrastructure requires specialised engineers and IT professionals. The Uptime Institute reports that staffing challenges have neither improved nor worsened in 2024, and demand for skilled labour is surging.

Sustainable Growth and Risk Mitigation Strategies

1. Power‑planning and energy innovation

   • Scenario modelling: Conduct holistic site feasibility studies and plan for power costs using scenario modelspwc.com.

   • Renewable integration and microgrids: Invest in on-site solar, wind, hydrogen or nuclear partnerships. Meta’s 2024 request for 1-4 GW of nuclear capacity for AI operations illustrates the trend aon.co

   • Adaptive cooling and hardware optimisation: Use new accelerators and 3D chips, dynamic liquid cooling and scheduling workloads when renewable power is abundant.

2. Diversified site selection and geopolitical risk management

   • Prioritise regions with stable energy supply, lower climate risk and supportive regulatory frameworks. As Tier‑1 markets become saturated, consider secondary markets where land and grid capacity are available.

   • Diversify across countries to minimise exposure to local disasters or geopolitical tensions. Understand data sovereignty laws and privacy regulations in each jurisdiction.

3. Supply‑chain resilience

   • Develop strong supplier relationships, diversify sourcing regions and maintain approved vendor lists.

   • Use e-procurement and regular performance reviews to enhance transparency and anticipate shortages.

4. Workforce development and collaboration

   • Invest in training and apprenticeships to build the specialised talent pipeline. Partner with universities and technical schools to develop curricula for AI data‑centre operations.

   • Use managed services or strategic partnerships to supplement capacity during rapid scaling.

5. Holistic governance and risk frameworks

   • Establish cross-portfolio governance models to ensure issues at one facility are communicated across the network.

   • Integrate cybersecurity and physical security risk management, as automation and scale increase the attack surface.

Conclusion

The AI boom is creating extraordinary opportunities, but success depends on proactively managing interconnected risks. By planning for energy, diversifying sites, strengthening supply chains, investing in people and embedding governance frameworks, organisations can build data‑centre capacity that keeps pace with AI demand while safeguarding reliability and sustainability.