Flexibility Is Not Free: Who Bears the Risk in a High-Load Grid?

The U.S. grid has always relied on demand response. What’s different now is the scale and economics of the demand.

At Operating Under Load Growth, one idea quietly cut through the technical complexity:

Large loads can be flexible — but flexibility is a political and economic question, not just a technical one.

---

The Assumption: Big Loads Will Behave

Sean Kelly (CEO, Amperon) emphasized that many large loads already participate in demand response programs, and that flexibility is wider than many assume.

“These large loads are surprisingly flexible,” he said.

He described how forecasting can anticipate stress periods and help customers respond—whether through coincident peak programs or price signals. He also pointed to Winter Storm Fern as a real-world demonstration that large loads curtailed meaningfully when called upon.

The technical story is hopeful:

• Flexibility exists

• Forecasting is improving

• Operational coordination can work

But the panel quickly exposed a second layer: not every large load has the same incentive to respond.

---

The Economic Reality: Not All Megawatts Are Equal

Brenden Millstein (Cofounder and President, Centinel) reframed the issue in blunt economic terms.

He laid out a rule-of-thumb hierarchy: utility cost per square foot is small relative to rent, and rent is small relative to the economic activity inside the building. For data centers, he argued, the economic density is orders of magnitude higher.

His conclusion was stark:

“The idea that a data center will curtail because the grid asks nicely, I think is not going to happen.”

Instead, he suggested hyperscalers will choose between two strategies:

• install batteries, or

• use influence to reshape the rules

“Call me cynical,” he added, “but that’s my hot take.”

The implication is uncomfortable but necessary: flexibility is not just an engineering lever. It is a distributional and market power issue.

---

The Engineering Layer: Speed, Not Size, Is the New Risk

Antonio J. Conejo (Professor, The Ohio State University) introduced a separate risk category: dynamic instability.

He argued that large loads can change “in a dramatic way quickly in a particular location,” and that this dynamic behavior—especially combined with high renewable penetration—can “eventually” lead to blackouts if not managed through safeguards.

His point shifts the conversation from megawatts to acceleration: how fast load moves, and how sharply it changes across locations.

In Conejo’s framing, “good behavior” won’t emerge naturally from economics alone. It must be enforced through operational standards and regulation.

---

The Legal Dimension: Curtailment Becomes a Property-Rights Question

Joshua Macey (Professor of Law, Yale Law School) pushed the conversation into institutional design.

If load growth outpaces grid capability at certain moments, curtailment priorities must be assigned. That creates a system-level question: who is curtailed first, and under what rules?

Macey suggested that secondary markets could emerge—where high-value loads buy out curtailment obligations from lower-value loads—because curtailment has real economic value.

Whether allocation is administrative or market-based, the core issue remains: flexibility is not neutral. It shifts costs and risk across participants.

---

The Core Tension

The panel’s underlying message is that flexibility will matter more than ever—but it cannot be assumed, and it will not be distributed evenly.

• Some large loads can curtail cheaply.

• Others will treat curtailment as unacceptable.

• Some will self-supply with batteries or generation.

• Others will carry the reliability burden.

Operating under load growth means grappling with this reality directly.

Flexibility is possible.

But it is not free.

And the central question for the next phase of grid planning is simple:

Who bears the risk when the system is under stress?