Time-to-Power Is the New Critical Path

A data center that is built but not energized earns nothing. Every month between mechanical completion and power-on is a month of stranded capital on some of the most expensive real estate ever assembled — which is why the industry has started measuring itself with a phrase borrowed from energy development: time-to-power.

For two decades, power was a procurement line in a data-center business plan. You picked a site near fiber and substations, signed with the utility, and the megawatts arrived before the servers did. That assumption has broken.

The demand shock, briefly

The International Energy Agency projects that global data-center electricity consumption will roughly double from around 415 TWh in 2024 to around 945 TWh by 2030 — growing more than four times faster than electricity demand from everything else. Meanwhile, on the supply side, grid connections have become the queue everyone is standing in: in some major European cities, obtaining one can now take seven to ten years.

Compute can be manufactured in quarters. Grid capacity is built in decades. When demand doubles into a queue like that, the bottleneck moves from compute to power — and the market responds the way markets do: it routes around the queue.

Announcing megawatts is easy. Delivering them is development.

The clearest evidence is behind-the-meter generation — power plants built at or next to the data center, supplying it directly. Research by Cleanview identified roughly 90 GW of behind-the-meter generation planned for US data centers across 59 projects — more than a quarter of all planned US data-center capacity — with 92% of it announced since the beginning of 2025.

And here is the number that matters for this article: of those ~90 GW, only around 2 GW are actually operating.

A contracted megawatt is a promise. A delivered megawatt is a development project that finished on time.

The gap between 90 announced and 2 delivered is not a technology gap. It is land that must be secured, permits that must be obtained, gas and grid infrastructure that must be built, interconnections and substations with their own schedules, and — as several high-profile US projects have already discovered — permitting processes that can stop a private power plant just as firmly as they can stop a merchant one.

The energy developer moves to the center

This is the structural shift underneath the headlines: the parties who know how to take energy infrastructure from screening to commercial operation — developers and IPPs — are moving from the periphery of data-center development to its critical path. Hyperscalers are signing gigawatt-scale agreements with them, partnering with them on co-located campuses, and in some cases bringing development capability in-house.

For the energy developer, the customer changed shape. The offtaker is no longer only a utility or a PPA counterparty; it is a campus with a required energization date, and every development milestone in the power portfolio now maps to someone else's go-live. For the data-center operator, the dependency inverted: the question is no longer "what will power cost?" but "will the projects behind my contracted MW actually deliver on time?"

Same operating model, different delivery context

Developing power for a data-center campus is not identical to developing a merchant solar portfolio — the interdependencies are tighter, the schedule tolerance is smaller, and the load is a counterparty with its own construction program. But the governance questions are the ones energy development has always had to answer: Who owns what? What depends on what? What has changed? What is committed? What is at risk? What needs attention next?

A planned energization date is the end of a chain of assumptions — land secured, permits granted, generation built, substation energized, connection completed. When the chain lives across spreadsheets, the date is an act of faith. When every link is a governed object with an owner, a baseline and a measured deviation, the date becomes explainable: you can say which link moved, when, by how much, and what it does to the campus.

Europe's version of the story

The largest behind-the-meter numbers are American, but the underlying pressure is at least as strong in Europe — and it is redrawing the map. Congested hubs push load toward markets where power and land can still be secured, and Spain has become the prime example: its 2025–2030 grid plan sizes 3.8 GW of new data-center demand, and Spanish energy companies are positioning to serve around 12 GW of data-center-linked projects representing tens of billions of euros of investment, according to reporting by Strategic Energy Europe.

European projects also carry the permitting reality we have written about before: milestones here are administrative acts with legal deadlines, not planning estimates. A power portfolio serving a campus in Aragón or Lombardy has to be governed with the same discipline as any renewable portfolio — per-jurisdiction processes, versioned baselines, quantified risk — with a data center's go-live at the end of the chain.

What this asks of your tooling

Navorda's position in this shift is deliberately precise. Navorda does not manage data centers — no compute, no cooling, no racks, no construction scheduling. It governs the energy project development layer underneath them: the power projects, land, permits, milestones, baselines, risks and ownership structures that a campus's energization date depends on. Same operating model that runs 300+ energy projects across Europe; a new source of demand at the end of the chain.

Because when the critical path of the world's most valuable infrastructure runs through energy development, the developers who can explain their time-to-power — not just promise it — are the ones who will keep winning the contracts.

Navorda for data-center power development — govern the power projects behind the data center

Energy Critical Path: why plants, substations and power lines should not be managed as a single project