Clean Power Has to Arrive on Time

The photograph is attractive. A field of panels, a bright sky, a company name. Beneath it sits a promise about clean energy.

We should want that promise to become real. More useful computing supported by cleaner supply is a better ambition than more useful computing with no concern for how it is powered. The question is what the claim means and what has been built to support it.

A contract, a certificate, a generating plant and electricity available at a particular connection are different things. They can be related in valuable ways. Treating them as identical makes it harder to judge the value.

Consider a simple imagined account. A facility uses ten units of electricity during daylight and ten after dark. A renewable project associated with it produces twenty units during daylight and none in the second period. Across the day, production matches consumption. During the second period, the facility still needs another arrangement for supply.

Nothing in that arithmetic says the renewable project is useless. It may make a valuable contribution. The arithmetic tells us that matching totals and matching time answer different questions.

A clean-energy claim should tell us which question it answers.

Start with the service the system must provide

A data center does not buy an abstract annual sum and disappear until the accounting period ends. It operates through hours with different conditions. Its promised service may require continuous availability even when its particular workload changes.

The electricity system must meet the relevant demand while respecting its technical limits. If a project has arranged annual renewable matching, ask what serves it during the hours that do not match. If it has storage, ask what duration and operating conditions the storage can cover. If it has firm supply from another source, ask where and how that supply reaches the connection.

These are ordinary design questions. They should not be treated as an attack on clean energy or an excuse to avoid building it.

Imagine a buyer comparing two offers. The first provides a large annual quantity at a favorable price. The second provides a smaller quantity but is available during periods when the buyer has a harder time securing supply. Comparing only total annual output would miss part of the value. Comparing only one difficult hour would miss another part.

A sensible portfolio may combine different resources, network access, storage and agreed flexibility. Which combination works depends on the setting. A national argument conducted as a contest between favorite technologies can obscure the practical task of building an adequate local and regional system.

We should ask the system to perform before asking it to confirm our identity.

Location is part of clean supply

Return to the imagined renewable project. Suppose it generates the right amount at the right time but is connected where the network cannot carry the needed power to the load. The account may look balanced while the physical constraint remains.

This is why place belongs beside time. A commercial relationship can support investment without making every physical route unconstrained. The public explanation should distinguish the financial contribution from the service actually available at the relevant location.

That distinction also helps us recognize real benefits. A new project might improve the broader supply even if it does not directly serve one building in every hour. Its value should be described honestly rather than stretched into a claim that invites distrust.

For our imagined town, a useful clean-power plan would name the supply arrangements, the required network works and the conditions under which the facility must adapt. It would show which parts exist, which are under construction and which remain proposals. It would explain what happens if a planned resource is late.

An opening date should not depend on an unbuilt solution that is mentioned only when someone asks a difficult question. If an interim arrangement is needed, it should be assessed as an interim arrangement, with its own effects, costs and end conditions.

Temporary deserves a date and a test. Otherwise, it is only a comforting description of something that may last much longer.

Additional to what?

A company may say it supports additional clean supply. The word additional points toward an important question: did the arrangement help bring about something that would not otherwise have happened in the same way or on the same schedule?

That can be difficult to establish. Investment decisions have several causes. A project may have other buyers, public support or a long development history. A responsible claim should describe the company’s role without taking credit for every outcome nearby.

For a proposed public bargain, I would ask for evidence appropriate to the claim. What commitment was made? What did it enable? When will the resource operate? What alternatives were available to the project? Who verifies the milestone?

This is a way to encourage useful participation, not to demand philosophical certainty about an alternative universe. We can distinguish a meaningful long-term commitment from a loose association. We can check whether a promised project was built. We can avoid counting the same benefit several times in several announcements.

The public should be able to see whether the clean-supply promise is a plan with responsibilities or merely a purchase of flattering language.

A growing industry has a chance to become a serious customer for better infrastructure. That chance is worth taking seriously enough to measure.

A policy that joins several conditions

Ireland’s regulator adopted a new data-center connection policy in December 2025. It combines location-specific review with generation or storage requirements and at least 80 percent annual demand matched by additional renewable projects in Ireland, allowing a six-year glide path. This describes the policy decision, not a verified account of every later implementation step. Annual matching still does not establish clean supply in every operating hour.¹

The example is useful because it treats connection as a package of obligations rather than a single yes. It does not provide a ready-made rule for every country. Different networks, laws, resource choices and institutions require different designs.

The general question is what a new large user should contribute to the conditions that make its service possible. A policy can connect permission to build with commitments about supply, timing and location. It can also become too complex to administer or too rigid to respond to better options. The rule needs a clear purpose, evidence that its requirements serve that purpose and a way to review performance.

We should borrow questions before borrowing percentages.

A familiar number can look authoritative in a new setting while doing very little work there.

Storage is a bridge with a length

Storage can move usable energy from one time to another. That makes it valuable in a discussion where timing matters. But a storage label alone does not tell us which gap can be crossed.

In an imagined system, a battery might cover a brief disturbance, shift use across several hours or support a planned reduction in grid demand. Those are different duties. The available power, stored energy, starting charge and operating constraints determine what it can do.

A public claim should name the duty. If the battery is reserved for the facility’s own emergency needs, the same capacity should not casually be promised to the grid at the same moment. If it is expected to recharge during a constrained period, that requirement belongs in the assessment. If its service depends on a control system, the control system should be tested.

There is nothing anti-technology about asking these questions. They are how a technology becomes dependable enough to earn trust.

The same reasoning applies to any resource. A plant’s label cannot tell us whether it will be available under the conditions assumed. A contract cannot create delivery through an unavailable connection. A proposed future technology cannot serve today’s load merely because it would be impressive if completed.

Plans can include emerging options. The present commitment should identify how much depends on them and what happens if they do not arrive on time.

Hope can guide research. It should not be entered as completed capacity.

Backup has a neighborhood

An on-site generator can be important to a facility’s resilience. It can also affect the area around it when operated. The assessment needs to distinguish emergency service, routine testing and any proposed use for wider grid support.

These uses may involve different operating patterns and permissions. A device intended for occasional backup should not acquire a new role through a vague statement that the facility can help the system. If the role changes, the relevant air, noise, fuel and technical requirements need examination.

Imagine a town agrees to one limited pattern of operation, then learns that a financial incentive makes more frequent running attractive. The equipment has not moved. The practical bargain has changed. Residents deserve a process that recognizes the change before it becomes normal.

A good contract can prevent this confusion by naming the approved services and the review needed for additional ones. It can require operating records at a useful level of detail. It can connect grid benefits with compliance rather than treating one as a reason to overlook the other.

Reliability is important. So is the way it is obtained.

A community should not have to trade a clear obligation in one system for an unclear burden in another.

Count the work that improves the whole system

Some of the most useful clean-power work will not look glamorous in a campaign photograph. Better forecasting, sensible siting, careful scheduling and well-maintained equipment can matter alongside new generation.

For an imagined campus, moving a task to a less constrained period might reduce the need for an expensive response. Locating a later stage where suitable infrastructure exists might avoid unnecessary works. Coordinating maintenance could help preserve service. Each proposal needs evidence and measurement; none should be awarded a benefit just because it sounds efficient.

There is also a public opportunity in building shared capabilities. A stronger network can serve several forms of useful activity. Training can support workers across industries. Public data about constraints, released with appropriate security protection, can help developers choose better locations before spending heavily on the wrong ones.

That is a more ambitious vision than bargaining over one facility at a time with no wider plan. It treats the arrival of new demand as a reason to improve the system deliberately.

The danger is using a broad vision to excuse a weak project. A local agreement still needs its own facts. A strategy becomes credible when particular decisions are consistent with it, including the decisions to redirect or refuse proposals that do not fit.

A map earns its value through the routes it actually supports.

The climate claim and the project decision

People sometimes want a single answer to whether AI will help or harm the climate. This book cannot settle every future use of a changing technology, and neither can one energy purchase.

A useful application may help solve a problem. A wasteful application may consume resources without much benefit. Infrastructure can become cleaner or remain dependent on a poor arrangement. Adoption can expand. Efficiency can improve. The total result depends on these developments and on choices outside the technology itself.

That uncertainty should not paralyze present responsibility. We can assess the facility being proposed, the service being procured and the commitments being offered. We can require honest reporting. We can support demonstrated improvements and stop counting imagined future savings as though they have already canceled present costs.

Nor should present impacts erase the value of research that could lead to better outcomes. The reasonable response is to evaluate and improve the work, not assume that every possible benefit is either guaranteed or fraudulent.

A mature public position can hold both ideas: useful discovery deserves support, and current obligations must be met with current evidence.

There is enough ambition in that position to build a great deal.

Ask for a plan that can be read in bad weather

Our imagined town now receives a revised clean-power proposal. It includes expected supply in ordinary conditions and a response to difficult ones. It names the resources already operating, the projects still being developed and the commitments that support them. It explains which claims concern annual accounting and which concern delivery during particular periods.

The plan also has a failure path. If a resource is delayed, the operator must meet specified conditions before expanding. If flexibility is promised, it is tested. If backup operation changes, the relevant permissions and local effects are reviewed. If the public claim overstates performance, the record can be corrected.

This is less elegant than the photograph. It is more valuable.

A good plan should remain readable when the sky is dark, a component is unavailable and a deadline has slipped. That is when the division of responsibility matters most.

It should also make success visible. When a new resource arrives on time, when a facility performs the service it promised, when cleaner supply expands without pushing avoidable costs onto households, the public should be able to see that too. Accountability should recognize delivery as well as failure.

The clean-energy transition needs serious customers, skilled builders and institutions that can connect private demand to public needs. Computing can be part of that effort. It will earn that place through what its agreements cause to be built and how its facilities behave when the system needs help.

The photograph can remain on the brochure.

Behind it, let there be a plan that works at night.