This website uses cookies

Read our Privacy policy and Terms of use for more information.

In partnership with

A grid operator in Tennessee receives a 500 MW capacity request. The application says "light industrial." The coordinates place it 40 km from a nuclear plant that has been generating baseload power for three decades. What the application does not say: it is a 1 GW AI factory. The operator approves the connection. Nobody tells the nuclear plant.

The construction schedules do not overlap. The handover timelines do not reference each other. The permitting systems share no common data model. Two facilities that will depend on each other for decades. No shared document. No shared standard. No shared view of the infrastructure between them.

I have spent five editions inside the facility. The documents that disagree. The standards that conflict. The shift supervisor trusting a drawing that was already wrong. Every edition has lived in the gap between construction complete and permit to operate.

That gap is real. But it is not the only one.

The gap between facilities

Edition 1 put a cost on the intra-facility gap. £79m ($100m, €95m) per day for a single AI factory sitting dark. Editions 2 through 5 diagnosed why. Document chaos. Outdated drawings. Standards that conflict. A harmonisation layer that did not exist until someone built one.

But step outside the facility. Look at what surrounds it.

In Tennessee, a nuclear plant and an AI factory campus share the same transmission corridor. In France, EDF is bringing Flamanville 3 online. Three hyperscaler data centres are under construction within 200 km of the same grid node. In Singapore, the Energy Market Authority caps data centre power at 5% of national capacity. The island's energy infrastructure was not designed for this demand. In Brazil, Petrobras LNG terminals feed a grid that is now also powering the country's first wave of hyperscale compute.

Different sectors. Different regulators. Same grid. No coordination.

500 MW is not "light industrial"

The language matters. An AI factory applies for grid connection. It uses the same form as a manufacturing plant or a logistics warehouse. There is no category for "facility that consumes as much power as a city and must never go offline."

The grid operator sees a load request. The nuclear plant sees nothing. The data centre developer sees a site with baseload access. Nobody sees the system.

In the US, the Federal Energy Regulatory Commission regulates interstate transmission. State utility commissions regulate local distribution. Data centre permitting sits with local planning authorities. Nuclear generation sits with the Nuclear Regulatory Commission. Four regulatory bodies. Four databases. Zero shared view of what is being built within transmission range of what already exists.

In the UK, the picture is the same. National Grid ESO operates the system. Ofgem regulates markets. The ONR oversees nuclear safety. Local councils approve data centre builds. The connection queue exceeds 700 GW. Most applications will never be built. Nobody is mapping which sit next to which generating assets.

Others have started mapping. None of them see across sectors.

Organisations have begun building interactive maps of individual sectors. Reactor locations on a globe. Data centre builds on a dashboard. LNG terminals on a shipping chart.

Each map shows one layer. None shows the relationship between layers.

The nuclear map does not show the data centre under construction 40 km away. The data centre map does not show the nuclear plant providing its baseload. The LNG map does not show either.

This is the inter-sector gap. Not a document problem inside one facility. A visibility problem across an entire infrastructure landscape.

What the pattern looks like from a different altitude

Across the facilities Vistergy tracks on Atlas, spanning nuclear, data centres, and LNG, the pattern is structural. Facilities that will share grid connections and regulatory dependencies are being built in parallel. None has a shared view of the others.

I started with the sector I know best. Every reactor worldwide on a 3D globe. Open. Free. Permanent. Because the argument that nuclear is the backbone of future energy infrastructure should not sit behind a paywall.

This week, I'm widening the rule.

Every data centre under construction. Every LNG terminal that feeds the grid those data centres will draw on. All of it. Open. Free. Permanent. One map. Three sectors. One view of the system nobody else has connected.

The argument is the same. If energy is the bottom layer of the AI factory, and if the AI factory cannot work without it, then the view that connects them belongs in public. Not behind a gate. Not locked to the people who can afford a terminal subscription.

Then the questions started arriving.

Which data centres are under construction within baseload range of an operating reactor? Which LNG terminals feed the same grid node as a planned AI factory? Which transmission corridors are about to carry loads they were never designed for?

Those questions cannot be answered from a single-sector map. They require something that connects nuclear generation, compute demand, and energy supply in a single view. That is what I have been building.

The gap between sectors is where the next trillion pounds of risk lives

Every edition of this newsletter has focused on the gap inside the facility. The cost of that gap is measured in months and millions.

The gap between facilities is measured in years and billions. The Tennessee operator's 500 MW approval clears. The nuclear plant 40 km away learns about the load increase when the grid frequency shifts. Voltage sag. Transmission congestion. Emergency load shedding.

The IEA projects global data centre electricity demand will more than double by 2030. That demand is concentrated around existing baseload generation. The convergence is happening. The coordination is not.

This is the next chapter of Still Dark. The facility is still dark for 12 to 18 months after construction. But the grid connecting it to everything else is dark in a different way. Nobody can see the full picture. Not because the data does not exist. Because nobody has connected it across sectors.

What comes next

Edition 7: The grid was not built for what is coming.

Edition 8: The gap is not closing. It is getting wider.

If you build, operate, invest in, or regulate infrastructure anywhere in the world, this is written for you. Subscribe to Still Dark.

This newsletter lives in the gap between digital delivery complete and permit to operate. That gap is where value dies, and where it can be recovered.

I also co-author The Vistergy Brief at vistergy.com/archive. Satellite and geospatial monitoring, facility lifecycle intelligence, and standards architecture across LNG, nuclear, data centres, utilities and construction. Subscribe to both for the full picture.

The podcast is here. Permit to Operate is available at vistergy.com/podcast, Apple Podcasts, and Spotify.

Permit to Operate.

Fast browsing. Faster thinking.

Your browser gets you to a page. Norton Neo gets you to the answer. The first safe AI-native browser built by Norton moves with you from idea to action without slowing you down. Magic Box understands your intent before you finish typing. AI that works inside your flow, not beside it. No prompting. No copy-pasting. No switching apps.

Built-in AI, instantly and for free. Privacy handled by Norton. Built-in VPN and ad blocking protect you by default. No configuration. No extra apps. Nothing to think about.

Fast. Safe. Intelligent. That's Neo.

Reply

Avatar

or to participate

Keep Reading