BTM Is No Longer Optional: The Grid Won’t Wait

Grid interconnection queues in major data center markets have stretched to the point where utility-scale projects are being delayed or canceled outright

Decision Lens

The structural shift is now quantifiable: time-to-power runs roughly 1.5 to 2 years longer than previously expected, and approximately 30% of all planned U.S. data center capacity intends to rely on behind-the-meter resources. These two facts create a compound problem — not just operational delay, but a permanent reconfiguration of what it means to own and operate a data center. Operators did not sign up to run power plants, yet the grid cannot deliver power fast enough to support AI infrastructure timelines. Behind-the-meter is no longer a niche hedge; it is becoming the default entry point for new capacity.

90-Second Brief

In recent days, grid interconnection delays have made behind-the-meter generation the fastest path to operational capacity for new U.S. Data center builds. A March 2026 Bloom Energy survey found that time-to-power now runs roughly 1.5 to 2 years beyond prior expectations. Around 30% of planned U.S.

What’s Actually Happening

The shift is structural, not opportunistic. Grid interconnection queues in major data center markets have stretched to the point where utility-scale projects are being delayed or canceled outright. CBRE’s February 2025 North American data center report flagged declining under-construction capacity in Northern Virginia and Dallas-Fort Worth, with power procurement named as a contributing obstacle.

In response, operators are stacking BTM assets — fuel cells providing 9 to 12 months of speed-to-power advantage over turbine orders, hybrid solar-plus-storage configurations, and microgrid orchestration layers that allow safe islanding and, increasingly, grid export. Natural gas turbines remain a near-term anchor despite supply pressure: Wood Mackenzie estimated lead times for certain turbine classes at 243 weeks as of Q2 2025.

The White House’s nonbinding Ratepayer Protection Pledge, signed by leading hyperscalers in early March 2026, formalized what was already an industry direction: operators absorb generation costs rather than burden ratepayers. That political signal reinforces the operational logic driving BTM adoption — community and regulatory resistance to utility grid reliance is adding a permitting dimension to the power availability problem.

The emerging decision framework, per analyst commentary from Bank of America Securities, follows a sequenced hierarchy: secure fast capacity with gas turbines or engines, firm and smooth it with battery storage, then layer in solar as the lowest-cost marginal source. Within this framework, storage has shifted from a project-level optimization tool to a grid and load-serving necessity.

Why It Matters for Global Heads of Data Center Energy?

For energy heads managing multi-GW portfolios, BTM is now a capacity strategy, not a technology experiment. The 30% BTM share of planned U.S. data center capacity signals that procurement teams need to build new competencies: generator procurement, fuel supply contracting, permitting for on-site combustion sources, and interconnection agreements that allow export without unsafe backfeeding.

The operational model changes materially. Behind-the-meter assets require redundancy planning that differs from utility-grid assumptions — EPRI has flagged the cost and complexity of building redundancy into on-site equipment. Permitting for BTM systems, particularly those involving gas or combustion sources, adds regulatory layers that traditional grid interconnection projects do not face. Communities and utilities are increasingly requiring grid export commitments as a condition of project approval.

The risk profile also shifts. Basis risk on a Texas PPA is a known quantity; fuel supply volatility on an on-site gas turbine fleet is a different exposure class. Long-term PPA commitments remain relevant for RECs and Scope 2 accounting, but the physical power strategy now requires a separate lane — one that operates more like generation asset management than energy procurement.

The most consequential near-term decision is whether BTM assets are treated as bridge infrastructure or as permanent baseload. That framing determines capital allocation, technology selection, and how operators engage with utilities under emerging flexible power frameworks.

The Forward View

The regulatory context is evolving alongside the technology. EPRI and more than 65 organizations are developing the Flex MOSAIC classification framework, which would standardize how large loads describe their ability to adjust and support the grid — the foundation for a new class of utility-data center contracts. Operators who engage early with this framework are likely to hold a structural advantage in permitting and community relations.

Supply chain dynamics will shape the technology mix for the next two to three years. With turbine lead times at multi-year levels and fuel cell supply scaling incrementally, no single BTM technology can absorb all new demand. Hybrid portfolios — gas fast-start backed by battery energy storage, with solar as the marginal layer — are likely to become the operational standard before nuclear or long-duration storage options reach commercial scale at relevant timescales.

Policymakers are increasingly treating on-site generation as an expected feature of new data center proposals. Projects that cannot demonstrate a defined power solution from day one are facing longer permitting timelines and greater community resistance.

What We’re Uncertain About?

Whether BTM remains bridge or becomes baseload: Much BTM activity is framed as a stopgap pending grid interconnection — but at what point does a five-year bridge become permanent infrastructure? The answer depends on whether interconnection queue reforms materially accelerate timelines, which remains unresolved at the regulatory level.
Cost and redundancy economics at scale: EPRI has flagged redundancy cost as a genuine challenge for on-site systems, but no confirmed figures are available for large-scale BTM portfolios. Operator cost data across technology types would be required to assess true total cost of ownership versus grid interconnection.
Flexible power frameworks and their commercial terms: The Flex MOSAIC framework is under development, and the commercial structure of grid export obligations — pricing, curtailment rules, liability — has not been defined. How utilities translate this into tariff structures will materially affect BTM economics.
Fuel cell long-term performance degradation: Critics highlight performance fade over time as a fiscal and operational concern. Confirmed long-term operational data at AI-density scales is not yet available, which creates procurement risk for operators committing to fuel cell baseload.

One Question to Bring to Your Team

Given that supply chain constraints and grid delays make a single BTM technology bet operationally risky, does your current capital allocation and vendor strategy support a hybrid BTM portfolio — and have you mapped the permitting and utility engagement requirements for each technology layer at your five highest-priority sites?