Key findings
- AI infrastructure growth increasingly depends on power availability, grid connections and delivery speed rather than compute hardware alone.
- The IEA base case projects global data-centre electricity demand to reach about 945 TWh by 2030, more than double current levels.
- National Grid’s £40bn UK investment programme and data-centre power exposure create a measurable infrastructure thesis, but execution, regulation and valuation remain material risks.
- This is an opportunity watch, not a recommendation to buy or sell any security.
The signal
The constraint in the AI build-out is broadening from chips to electricity. The International Energy Agency projects global data-centre electricity consumption at around 945 TWh by 2030 in its base case, with accelerated servers accounting for a large share of incremental demand. ↗ That makes connection queues, transmission capacity, substations, generation and power-delivery timelines part of the AI investment stack.
Why the UK infrastructure angle matters
National Grid says its wider £40 billion UK investment programme is intended to modernise strategic electricity infrastructure supporting homes, factories and data centres. In September 2026 it announced £624 million of contracts across 78 companies, including 69 SMEs. ↗ This does not make every grid supplier an attractive investment; it does make the spending pipeline observable enough to monitor.
A public-market example, not a recommendation
National Grid also agreed a $1.75 billion investment for a 35% stake in Joulent, a business developing contracted power solutions for large US energy users including data centres. ↗ The investment thesis to watch is therefore not simply “AI grows”; it is whether regulated and contracted infrastructure can convert rising power demand into durable cash flows without capital costs, regulation or valuation eroding returns.
2026–2030 scenarios
Base case: grid investment grows steadily as AI and electrification increase connection demand, but delivery is paced by regulation, planning and construction. Upside: data-centre demand and accelerated connections pull forward network capex and supplier order books. Downside: efficiency gains, project cancellations, planning delays, financing costs or regulatory outcomes reduce expected returns even if electricity demand rises. The strongest confirmation signal would be sustained contracted investment plus delivered connections; the clearest invalidation signal would be a widening gap between announced projects and completed, revenue-producing infrastructure.
Decision use
For founders and investors, the practical question is to map where a bottleneck has pricing power, long-duration contracts or regulated returns — and where it merely has high capital requirements. Sanctuary’s management consultancy work can use the same scenario discipline when testing capital-intensive growth opportunities.
Research foundation
References
- International Energy Agency (2025). Energy demand from AI — Energy and AI. International Energy Agency.Source ↗
- National Grid (2026). National Grid investment programme supports UK supply chains and regional growth. National Grid.Source ↗
- National Grid (2026). National Grid Ventures to invest $1.75bn to accelerate power solutions for U.S. data centers and AI. National Grid.Source ↗
- Sanctuary Insights & Research (2026). Sanctuary editorial concept illustration. Sanctuary Consulting & Development Group.Image source ↗
Related Sanctuary capabilities
From analysis to implementation.
Discussion
No approved comments yet.