Why Power Terms Belong in the Shortlist
Power claims still get a provider onto a long list, but power terms increasingly decide whether it stays there. If a facility cannot explain who pays for upgrades, what happens during grid stress, and which milestones make capacity firm, a buyer should treat the offer as preliminary.
That is a change in the shortlisting process. For years, many enterprise infrastructure teams could leave utility details to legal and engineering after choosing a preferred market and building. With larger AI and high-density requirements, the commercial structure behind the power has become a first-round diligence item.
What has changed in the buyer's first review?
Buyers are seeing more capacity claims paired with more conditions. Some conditions are sensible: a utility has to study a load, construct facilities, and protect the broader system. The problem comes when sales language smooths over those conditions and a customer discovers them after it has spent months on design work, executive approval, and contract redlines.
Recent policy and tariff developments make the point. Duke Energy's proposed North Carolina settlement would apply to new qualifying large-load customers signing service agreements after June 1, 2026, requiring nonrefundable upfront payments for dedicated facilities, security for broader grid upgrades, and service under a high-load-factor rate. The stated threshold is 50 MW or more at an 80% load factor, subject to approval (Duke Energy).
A customer does not need to operate in North Carolina for that development to matter. It is a useful signal of the questions utilities and regulators are increasingly asking: Is the load committed? Who carries the cost if supporting infrastructure must be built? What protects other customers if the project changes course?
That means a buyer's initial facility comparison needs to go beyond available megawatts, cabinet density, and a target go-live date. Those are still important. They are no longer enough.
Which power terms should buyers ask for early?
Ask for a short written power basis during the shortlist stage. Not a full utility agreement; most operators cannot hand one over before a deal is real. But the operator should be able to provide a clear commercial explanation of where the capacity stands and what could change it.
The useful questions are practical:
- Is the stated capacity energized, under construction, contracted with the utility, or merely included in a development plan?
- What utility studies, permits, equipment deliveries, or customer commitments remain before the customer can draw the stated load?
- Are network or transmission upgrades expected, and who bears their cost if estimates rise?
- Does the service arrangement include curtailment, load shedding, load shifting, or minimum-load obligations?
- What deposit, security, take-or-pay, or termination exposure would flow through to the customer?
- Can the operator show a phased commissioning path rather than one broad in-service date?
The distinction between operational capacity and future potential is especially important. Applied Digital recently announced that its Polaris Forge 1 campus had reached 250 MW of operational critical IT load, while separately describing an agreement for up to 1 GW of potential capacity in Finland with initial availability anticipated in 2028 and development contingent on customer, regulatory, and commercial milestones (Applied Digital). Both statements may be relevant to different buyers. They should not be scored as equivalent capacity in an RFP.
Honestly, this is where a lot of procurement templates fall short. They ask for “available power” in one cell. That invites a single impressive answer when the buyer really needs several answers, each tied to a date and a condition.
How should a shortlist score power certainty?
Do not force every site into a binary “has power” or “does not have power” category. Use a certainty view alongside the normal price, location, network, and resilience criteria.
One simple approach is to classify the customer's required capacity by the evidence behind it. Capacity that is energized and serving comparable load belongs in a different band from capacity supported by a signed utility agreement, which belongs in a different band from capacity awaiting studies or a future substation. Then record the dependency that moves it to the next band.
For each finalist, the buying team should leave the diligence meeting able to state:
We can place this initial load on this date because these facilities are live. Our expansion depends on this named upgrade, currently at this stage, with this party carrying this specific risk.
If no one can say that clearly, the proposal is not ready for a final shortlist.
This does not mean every buyer should avoid development capacity. A fast-growing platform may rationally reserve future blocks before they are energized. But it should pay and contract for that uncertainty knowingly. The commercial structure for a live deployment should not be copied and pasted onto a capacity option that still depends on utility work, financing, permitting, or a customer's own commitment.
What does grid-stress language tell a buyer?
Load flexibility is becoming part of the commercial conversation. Proposed U.S. Senate amendment text printed in the Congressional Record would call for dedicated data-center interconnection queues, payment for necessary generation and transmission upgrades, and binding load-reduction or load-shifting agreements during grid stress. It is proposed text, not enacted law (Congressional Record).
The immediate takeaway is not that every customer will soon face the same rule. It is that a buyer should stop treating curtailment language as boilerplate. Ask what event triggers it, how much load is affected, how often it could be called, whether backup generation is permitted, and what service-credit or economic protections apply.
For an enterprise running a noncritical analytics workload, a managed load-shift arrangement may be acceptable and even financially attractive. For a production AI training cluster with fixed delivery commitments, it may be a material operational risk. Same clause; very different answer.
Buyers should also ask whether the provider has designed a credible operating model around flexibility. Finland's grid operator Fingrid has backed a proposal to prioritize constrained connections in a way that encourages large projects to locate where capacity exists and support the system with flexibility and new generation (Fingrid). That is a local policy proposal, but the direction of travel is familiar: flexibility can affect access to scarce capacity.
How can operators answer without exposing sensitive contracts?
You do not need to publish utility contracts to give buyers a usable answer. In fact, withholding everything until legal review usually creates more doubt than protection.
Build a buyer-facing power diligence sheet for each campus and update it when the underlying facts change. It should separate live capacity from contracted expansion and early-stage development. It should name the next gating milestone, summarize the anticipated customer obligation, and identify any load-management terms that matter to operations.
Avoid vague phrases such as “utility confirmed” or “power secured” without context. A stronger statement is: “The first deployment block is supported by existing energized infrastructure; the second block requires a utility upgrade now in design; customer commitment is required before construction release.” That is not flashy. It is credible, and credibility keeps a deal moving.
Sales should be trained to bring power, legal, and development colleagues into the opportunity before the customer asks a second time. By the time an enterprise issues final redlines, the right answer needs to be documented, not reconstructed from a chain of emails.
Common questions
Should we remove a site from consideration if expansion power is not live?
Not automatically. Keep it in play if the provider can distinguish the initial deployment from the expansion path, document dependencies, and offer terms that match the uncertainty. Remove it when the claimed date relies on unnamed assumptions or the provider cannot explain cost exposure.
Is a utility agreement enough proof of available capacity?
It is strong evidence, but it is not the whole answer. Buyers should still ask about construction milestones, upgrade scope, customer security requirements, and the operating conditions attached to service.
When should procurement bring power experts into the process?
Before the finalist list, not after it. Once a team has invested in architecture and commercial negotiations, it becomes harder to walk away from an attractive site with weak power terms.
Do curtailment provisions always make a facility a bad choice?
No. They can be workable when the affected workload, notice period, compensation, and operational response are understood. The issue is whether the provision matches the customer's actual uptime and delivery commitments.
What to do with this
Make power evidence a scored part of the shortlist, and insist that each claimed block has a status, a dependency, and a risk owner. The providers that answer plainly will often be safer partners than those offering the biggest unqualified number. GridReach helps data center and energy companies turn expertise like this into qualified pipeline.