Grid at a crossroads: The AI demand shock and the future of power
together—or fragment apart.
The rapid expansion of AI, cloud infrastructure, and digital services is triggering an unprecedented surge in electricity demand—one that is colliding head‑on with the physical, regulatory, and economic realities of the power grid. For utilities and hyperscalers alike, the question is no longer whether demand will grow, but how quickly power can be delivered, at what cost, and under what model.
Based on a KPMG survey of 100 senior leaders across the power, technology, and data center ecosystem, this research examines how AI‑driven demand is reshaping grid economics, accelerating the rise of behind‑the‑meter and hybrid power models, and forcing a fundamental rethink of utility–hyperscaler relationships.
Executive insights
1
2
3
4
5
6
Why this matters now
For decades, utilities planned around predictable, incremental load growth. Today’s reality is fundamentally different. AI‑driven demand is massive, geographically concentrated, and timeline‑sensitive, compressing planning horizons and increasing reliability risk.
For hyperscalers, extended interconnection queues and uncertain delivery timelines threaten capital efficiency and speed to market. For utilities, the risk is equally existential: if the largest loads migrate off‑grid, core rate‑base economics—and the ability to fund grid modernization—come under pressure.
This has created a fundamental gap measured not just in gigawatts, but in years.
Todd Fowler
US Sector Leader, Energy, Natural Resources & Chemicals, KPMG US
Key findings from the research
1 | Grid delivery delays are becoming a breaking point
Large‑load customers—particularly data center developers—are encountering multi‑year waits for power, even after facilities are constructed. This mismatch is reshaping site selection decisions and accelerating demand for alternative power models.
2 | Bottlenecks are structural, not cyclical
Utilities overwhelmingly point to:
- Supply chain constraints, particularly multi‑year lead times for transformers and switchgear
- Permitting and regulatory approvals that routinely extend project timelines
These constraints are largely disconnected from demand signals, limiting the sector’s ability to respond quickly—even when capital is available.
3 | Behind‑the‑meter and hybrid power models are becoming permanent
As confidence in grid delivery erodes, hyperscalers are moving forward anyway. What began as short‑term workarounds are now becoming structural features of the power system.
Utilities are not being displaced—but their role is evolving from sole supplier to integrator, orchestrator, and partner in hybrid power ecosystems.
4 | PPAs now serve a different purpose
PPAs are no longer used solely to support sustainability strategies. They are increasingly deployed to:
- Lock in delivery certainty
- Fund dedicated or on‑site generation
- Reduce exposure to curtailment and grid timing risk
5 | Cost allocation is the defining policy question
As utilities deploy capital at historic scale, the question of who pays has become central. Respondents show conditional support for faster infrastructure development—but only when costs are transparently allocated to the customers driving demand.
Dive into our thinking:
Grid at a crossroads: The AI demand shock and the future of power
New KPMG research explores why tech and power must accelerate together, or fragment apart.
Download PDFFuture of Power infographic
Rising AI demand is exposing grid limitations – triggering a rapid pivot to alternative power solutions
Download PDF- Utilities face a narrowing window to modernize delivery models, rethink rate structures, and reposition themselves as partners in hybrid power ecosystems.
- Hyperscalers must balance speed and certainty against regulatory, political, and stranded‑asset risk as private generation scales.
- Both remain interdependent: fragmentation may solve near‑term constraints, but it increases long‑term risks to reliability, affordability, and resilience.
A proactive checklist for utility leaders
To remain relevant—and competitive—in an AI‑driven energy landscape, utility leaders should focus on five priorities:
1
2
3
4
5
How KPMG can help
KPMG works with utilities and hyperscalers to navigate AI‑driven demand shock by supporting:
- Strategic acceleration of planning and interconnection
- Workforce strategy and supply‑chain resilience
- Quantitative load‑scenario and capital modeling
- Utility–hyperscaler partnership frameworks
- Regulatory strategy and hybrid grid rate design
Explore more
Meet the team
