Connected risk management for digital infrastructure and data centers

The explosive growth of artificial intelligence is straining global energy grids and water supplies, making connected risk management for data centers a critical imperative to prevent widespread disruptions.

Data centers, the backbone of digital infrastructure, are expanding at an unprecedented rate due to AI demands. Global capacity is projected to double by 2030, with nearly 100 gigawatts of new facilities added between 2026 and 2030. This surge follows the AI boom that accelerated in 2023 with tools like ChatGPT, but intensified in 2025 as enterprises shifted from pilots to production-scale deployments. By 2027, inference workloads—real-time AI applications—are expected to overtake training as the dominant driver, yet both require immense computational power.

Energy consumption has become the flashpoint. Data centers could claim up to 12% of U.S. electricity by 2028, up from 4.4% in 2023, adding 460 terawatt-hours of demand by 2030. In Virginia alone, they already consume 25% of generated power. Grids in regions like the Southeast face potential shortfalls, with utilities forecasting loads that exceed probable outcomes by factors of 500-to-1 in some models. This has led to delays: 57% of U.S. projects slipped by three months or more in 2025, costing developers up to $14.2 million monthly per 60-megawatt facility in lost revenue.

Water scarcity compounds the issue. Cooling needs could rise 170% by 2030, stressing supplies in drought-prone areas like Arizona and the Rocky Mountains. In Mesa, Arizona, a hyperscale center's approval for over one million gallons daily sparked community backlash and rate hikes. Backup power from gas generators adds air pollution and greenhouse gases, clashing with decarbonization goals—fossil fuels still supply 56% of data center electricity.

Economic ripples affect multiple stakeholders. Utilities risk overbuilding infrastructure, passing costs to households via higher rates. Businesses face outages that disrupt operations; a single downtime event can erase millions in productivity. Investors confront overbuild risks if AI adoption slows, alongside technological obsolescence from rapid chip advances like Nvidia's Vera Rubin generation, pushing rack densities to 250 kilowatts. Construction costs have climbed 7% annually, from $7.7 million to $10.7 million per megawatt between 2020 and 2025.

Regulatory pressures are mounting. The EU's AI Act, effective from 2025, mandates risk assessments and transparency for AI hosting. In the U.S., states like Maryland and Colorado have introduced similar rules, while global FDI in data centers rose 14% in 2025 but concentrated in fewer regions, heightening vulnerability to geopolitical tensions.

Talent shortages exacerbate execution risks, with 58% of organizations reporting severe IT skills gaps. Cyber threats loom larger, with 61% of risk professionals citing data breaches as top concerns amid AI-enabled attacks. Natural catastrophes, amplified by climate change, threaten facilities in exposed areas.

Connected risk management—integrating these interconnected threats—matters because isolated approaches fail against compounding factors like grid constraints and sustainability demands. Without it, the digital economy risks stalling, affecting everything from e-commerce to national security.