Grid operator PJM plans to require large data centers to either bring their own power generation or curtail electricity use to prevent a large-scale outage
Grid operator PJM Interconnection on Friday unveiled a plan to manage surging power demand from data centers that Big Tech needs …
Context & Ripple Effects
PJM’s proposal follows a federal push for data centers to reduce load or use backup power during grid stress, moving that reliability approach into the country’s largest grid market.
It also comes as PJM has reduced its projected summer peak after excluding projects without firm service or construction commitments, underscoring that speculative and committed data-center demand cannot be treated alike in planning.
First-order effects
- Large data centers seeking service in PJM would face a choice between securing on-site generation and accepting curtailment when reliability is at risk.
- PJM gains a demand-management tool aimed at protecting existing grid users from outage risks as high-load facilities connect.
Second-order effects
- Developers and their cloud customers will need to weigh the cost and operating constraints of firm power arrangements against the value of uninterrupted compute availability.
- The plan puts a premium on projects with demonstrable power commitments, reinforcing PJM’s distinction between prospective demand and projects ready to take service.
Third-order effects
- If adopted and replicated, grid access for data centers could shift from a standard utility connection toward a conditional service product with explicit flexibility or self-supply obligations.
- The policy tests a broader allocation question: whether new large loads should absorb more of the reliability measures their connection requires rather than spreading those costs and risks across other customers.
The trend: AI-driven data-center expansion is turning electricity availability and load flexibility into core constraints on where and how compute infrastructure can be built.