Who's Financing and Building this AI Wave?

AI data centers could attract $3T - $8T of capital by 2030.

McKinsey & Company outlines 3 scenarios: constrained momentum, continued momentum, or accelerated demand.

A key assumption: power capacity costs $2.2–$3.2B per gigawatt (GW) —depending on power mix (gas, gas + storage, or solar) and regional differences.

👉 Who's financing and building this wave? McKinsey points to five archetypes: • Builders • Energizers • Technology developers & designers • Operators • AI architects

Here's the translation I find most intriguing: ⚡ 1 GW of energy = one billion watts of generation capacity on the grid 💻 1 GW of data center demand = one billion watts consumed by servers, cooling, and support infrastructure—delivering exaflops-scale AI performance (10¹⁸ calculations per second)

Big picture, why the distinction matters: • Building 1 GW of grid capacity → costs ~$2 - $3B, based on scenarios ranging from fully gas-powered to solar + storage combinations per McKinsey's analysis • Building 1 GW of data center → costs ~ $40B - $50B, driven by high-end equipment such as GPUs and specialized cooling

Same unit—1 GW—but radically different implications for risk and timing.

For example, grid assets operate for decades; data center IT equipment (GPUs) cycles every 4 years.

Link (McKinsey & Company, the cost of compute: A $7 trillion race to scale data centers, 2025): https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-cost-of-compute-a-7-trillion-dollar-race-to-scale-data-centers#/

________

Clean Power Whisperer™ Perspectives |

Educational Insights on Risk, Insurance & Capital for AI, Energy, and Climate.

← All Perspectives · Next: Roads, Not Carriages →