Jonathan Siegel
The build–out of infrastructure to support artificial intelligence has already started showing up on electric bills.
At PJM, the regional transmission organization that serves about 67 million people from Illinois to the Mid-Atlantic, the capacity clearing price for most of the region jumped roughly ninefold in the July 2024 auction, and the following two auctions each cleared at their Federal Energy Regulatory Commission-approved cap. PJM’s independent market monitor estimated that data-center demand accounted for roughly 40% of that auction’s capacity-market revenues.
Separately, PJM estimated that the July 2025 auction would add between 1.5% and 5% to some customers’ bills for the delivery year that began in June 2026. Before those higher bills become an investment risk, they are a risk to household finances. Families with no say in the matter are paying more to keep the lights on so that the build-out can proceed. Rising bills—and the resentment they breed—also tend to harden into the kind of organized opposition that stalls projects, which is the part allocators overlook.
Weighing the Risks
I invest in this kind of infrastructure on the private side, so I have a stake in how allocators weigh these projects. I am neither an engineer nor a grid operator; my job is to underwrite these assets. And I will be plain about a conviction that shapes how I read all of this: An industry preparing to spend on this scale should not be loading its costs onto the nearest community, and the projects worth backing tend to be the ones that protect local ratepayers and give the people next door to the centers a real share of the upside. Weigh that however you like, coming from an investor. The useful part of the argument does not require you to share my opinion, because doing right by these communities has started to look a lot like just good underwriting.
Start with the mismatch at the center of the build–out. Demand for power can move quickly. The system that supplies it cannot. Transmission and grid upgrades take years to permit and build. Supply chains are stretched too: The International Energy Agency found that new large power transformers can take up to four years to secure, while the acquisition time for some specialized grid components can exceed five years. Meanwhile, Berkeley Lab estimates that approximately 8,200 projects were actively seeking grid interconnection in the U.S. at the end of 2025, representing 1,312 gigawatts of proposed generation and roughly 749 GW of storage. The typical project that came online in 2025 had spent more than five years in the queue. None of that comes online quickly. Power infrastructure does not respond to a price signal the way a software company responds to demand.
For a CIO with capital in private infrastructure investments, this falls under execution risk, as do cost overruns and counterparty default. It is a material execution risk, and a concentrated one in the asset class now drawing the most capital. A permit that never gets regulatory approval can leave a stranded asset, and a stranded asset rarely earns its way back. Last year, by Data Center Watch’s count, U.S. data-center projects representing more than $150 billion in publicly disclosed proposed investment were blocked or delayed amid local opposition, moratoria or litigation. The demand was real; the delivery was not.
The response is to underwrite the factors that decide whether a project actually gets built—and to price them, rather than note them and move on.
Consideration for Communities
There is a hierarchy of risk for CIOs to understand and evaluate. Power comes first: A developer that tells investors that power for their project is “secured” is usually describing a hope, when what matters for the viability of the deal is queue position, firm capacity and a real energization date. The significance of water is underrated, and siting a thirsty facility in a region already stressed can invite the permit fight and the litigation that derail a timeline. The cost of insurance for a project should be underwritten the same way: If a facility cannot document how it runs, an allocator should assume coverage may become more expensive or harder to maintain. The community is the variable most often evaluated least—until the day community opposition sets the schedule back.
It is tempting to file local opposition under NIMBYism and permit around it. I think that reading is both wrong and expensive. People object for good reasons. They watch their electricity bills climb and their water tables fall so a data center two counties away can run, while most of the jobs and the profit from the sited facility land somewhere else. These are the neighbors who will live with the project for its whole life, and they deserve a real seat at the table.
If one changes who bears the costs and who shares the benefits, the opposition often softens. Some consumer and environmental advocates have begun proposing versions of that model in which new data centers pair their own clean-power supply with ratepayer protection and direct benefits for nearby households. I would hold those additions to the deal economics to the same independent verification I would apply to any projection, including my own. The projects that get built tend to be the ones that can give a community a credible reason to say yes.
Here is the uncomfortable part for people who do what I do. In my experience, few managers present all of this in one standardized, independently verifiable form. So investment managers raise capital on capacity, pipeline and demand, none of which tells an allocator whether the thing in which they are investing can be delivered or whether it will be a decent neighbor once it is built. The investor ends up carrying power, water, permitting and community risk on the manager’s word, and that is where some of the write-downs may come from if the market underwrites this too loosely.
The Right Questions
If you allocate to these kinds of projects, a short list of questions closes much of the gap, and none require an engineering degree:
- Power. What is the project’s interconnection queue position? What firm capacity has it secured—with dates?
- Permitting. What is the local approval path, where does it stand today, and is there any organized opposition on record?
- Water. What is the source, the volume and the basin? Which permits hang on them?
- Insurance. Can the asset document an operating profile a carrier will price? What coverage runs over the hold?
- Community. What does the host community receive, in writing—lower bills, a share of the benefit, enforceable protections—and who is accountable if those terms slip?
- Verification. Which of these answers has an independent party checked? Which are self-reported?
A manager that has done the work tends to have the answers ready. A manager that has underwritten demand and hoped for success tends to have a story instead.
The build-out is already underway; capital has been committed, and demand remains real. The live question for asset owners is how carefully to underwrite their share of it.
For years, the binding constraint on this build-out was capital. That is changing. As deliverability becomes the bigger constraint, the discipline that protects investors and the discipline that protects the people living next to these projects increasingly look like the same discipline.
Allocators that treat power, water, permits, insurance and community terms as real numbers to be priced in and verified are better placed to tell, early enough to matter, which projects are real. And the communities those projects touch are more likely to be heard in the process.
Jonathan Siegel is a partner in Altes Capital and managing partner of Altes Ascent, the firm’s impact investing division. He writes as an investor in the sector with a commercial interest in how these assets are evaluated.
For discussion purposes only. Not an offer to sell or a solicitation to purchase any security or advisory service. This material does not constitute investment, legal, or tax advice and is not a recommendation to buy or sell any security or to pursue any investment strategy. Statements regarding future events, trends, or outcomes reflect current beliefs and assumptions as of the date of publication, are subject to significant uncertainty, and should not be construed as a guarantee of future results. References to Altes Capital’s underwriting approach, investment criteria, or current firm practices reflect general views and practices as of the date of publication, are subject to change without notice, and do not guarantee the outcome of any particular investment decision. Third-party data and sources cited herein are believed to be reliable but have not been independently verified by Altes Capital LLC, which makes no representation as to their accuracy or completeness. Altes Capital LLC and its affiliates, including the author, may have financial interests in the types of investments discussed.
This feature is to provide general information only, does not constitute legal or tax advice, and cannot be used or substituted for legal or tax advice. Any opinions of the author do not necessarily reflect the stance of ISS STOXX or its affiliates.