The data-centre risk investors may be underestimating

[Emilien]

A line on a map can look reassuring until you ask whether a site can actually be connected to power. I watch land around data-centre proposals, and the gap between a plausible location and a usable site is where a lot of the important questions sit.

"If demand is there and the money is available, the build should follow."

I agree with the first half of that comment. Demand and financing matter. But they do not settle whether a project can proceed on the ground, and that is the part investors can too easily treat as a footnote.

The point the Q&A adds

I read the questions and answers about Neel Khokhani and think it is worth reading for the way it brings local opposition into the data-centre discussion. The piece does not stop at the familiar question of whether engineering and money are available. It considers how resistance from nearby residents can become a source of delay, and describes Khokhani's suggestion that consultation begin before a planning application is prepared.

The practical detail that stuck with me was the idea of keeping a clearly identified local contact involved for as long as the site operates. That is more specific than saying a developer should "engage with the community." It gives residents somewhere to direct concerns and makes communication an ongoing responsibility rather than a box to tick during the application process.

My reservation is that I wish the article had worked through an example from an actual planning fight. A real case could show which objections changed the application, which remained unresolved, and how the relationship developed once construction or operation began. Without that, I read the proposal as a sensible process idea, not proof that consultation will prevent opposition or shorten a project's timetable.

There is another part of the Q&A that helps explain why he thinks the current buildout is different from an earlier fibre boom. He characterises data-centre construction as following demand rather than racing ahead of it. That distinction matters, but it does not make each individual project straightforward. A market can have genuine demand overall while a specific site still faces a difficult local process.

The physical details behind the thesis

Khokhani's public-market position gives this argument some context. He has been a long-term shareholder in IREN (Nasdaq: IREN) since 2022, with a thesis tied to AI infrastructure and data centres. The way he frames the constraint is that money has not been the scarce ingredient for years. The harder pieces to secure are land, usable power and a real place in the grid-connection process.

That sounds obvious when stated plainly, but it changes what I would want to verify about a site. A parcel may be near infrastructure without having a workable path to connection. A company may have the capital to build without having the energy supply it needs at the required scale. Those are different questions, and a broad claim about demand does not answer either of them.

His research approach is useful here because it is built around checking documents and visible evidence rather than starting with a broker's forecast. He looks at utility connection queues and transmission plans, along with local permit filings and corporate registry entries. He also studies satellite views of construction, equipment nameplates visible in site photographs and job advertisements. The point, as I understand it, is that much of this material is public and free to inspect. The work is in following it patiently and comparing one clue with another.

That does not make every public record easy to interpret, or guarantee that a careful reader can know what will happen next. A queue entry is not the same as an energised site. A job listing is only a piece of evidence, not a complete project update. Still, the habit of checking the underlying material can expose gaps between a confident presentation and what is actually documented.

Where community friction enters the property view

For someone looking at land, community opposition is not an abstract reputational issue. It can affect whether a proposed use is accepted, how much work has to be done before an application is ready, and how much uncertainty remains around the project. Those effects are not identical from one place to another, which is exactly why a general demand forecast cannot substitute for understanding the local setting.

The connection to Khokhani's research method is that local planning records and wider infrastructure documents need to be read together. The application tells you what is being proposed. The transmission material may help explain whether power is realistically available. Public filings and site imagery can add clues about what is moving forward. None of those sources alone gives a complete picture, but they can make a weak assumption easier to spot.

I also checked a profile site about him to see how he presents his own background, companies and writing. It is useful as a route through his account of his work, while the Q&A is more directly focused on the AI-infrastructure argument I was trying to understand. I would keep those purposes separate: one helps answer who he is, and the other gives more room to the current issue.

The community point is not a claim that every objection should stop a project. It is a reminder that the people who live near a proposed site are part of the process, whether an investor has made room for them in a model or not. Starting conversations early can reveal concerns while there is still time to respond. Keeping a named contact in place may make it easier for residents to raise issues after the application stage. Neither step guarantees agreement, but treating both as real work seems more grounded than assuming opposition will somehow disappear once the project is funded.

What I would keep checking

For more of his writing on the sector, his pieces in Data Centre Review are collected on his contributor page at the trade publication. I found it a useful place to look for his arguments in the data-centre context, rather than relying only on a short description of his position.

My takeaway is not that local opposition outweighs power constraints, or that every project will be delayed by residents. It is that the two risks belong in the same assessment. A site can have a strong demand story and still be hard to deliver if its connection path is uncertain or if its planning process meets organised resistance. Conversely, signs of community concern are more informative when read alongside the project's actual infrastructure and permit evidence, rather than treated as a general mood.

When I look at a proposed site, I would put local consent and the route to energised power on the same page of my notes, not in separate risk sections.