AI power diligence — grid queues, microgrids and capacity realism

AI infrastructure is now gated less by capital headlines than by time-to-power, interconnection evidence, PPA bankability, and the credibility of behind-the-meter alternatives.

Reader guide

How to read this Signal

Evidence → Uncertainty → Trigger → Decision implication

Each Signal separates the sourced change from what remains uncertain, identifies the trigger to watch and states why it matters to a decision.

See the full Approach →

Key points

What the evidence indicates for the decision.

  • IEA's Energy and AI work projects global data-centre electricity consumption to more than double to around 945 TWh by 2030, with AI as a major demand driver.
  • IEA's Electricity 2026 demand chapter expects 2026-2030 to add materially more electricity demand per year than the past decade, with data centres among the major end-use contributors.
  • Reuters reporting on US storage firms points to widening interest from AI data centres in front-of-meter and behind-the-meter storage, but flags interconnection timelines that can exceed data-centre build periods.
  • Khazna's announced plan to add more than 1 GW of hyperscale capacity across multiple countries by 2030 is a practical indicator of GCC/EU AI-infrastructure expansion pressure.
  • The diligence implication is simple: treat power, interconnection milestones, PPA counterparty quality, OEM lead-times and microgrid optionality as a single delivery-risk stack.

Actions

Practical next steps if this Signal touches your mandate.

  1. 01Request a power evidence pack before IC approval: interconnection status, PPA heads, grid studies, on-site generation/storage design, OEM lead-times, and key assumptions.
  2. 02Separate 'available capacity' from 'bankable capacity': document who bears delay, curtailment, fuel, grid-upgrade and balancing risk.
  3. 03Use a 30/60/90-day trigger board for grid queue movement, permitting milestones, PPA pricing windows and OEM delivery slips.

Sources and evidence

Publication and access dates are shown with the source record.

Discuss how this affects your decision

We can pressure-test assumptions and map the next verification steps.