The short answer
Data centre capital is plentiful; deliverable power is not. Interconnection timelines of roughly three to seven years sit against build cycles of one to two, so the projects that get financed are the ones that can evidence a credible path to energised capacity. Power procurement has become the first diligence item, not the last.
For most of the last decade, a data centre developer's hardest problem was capital. That is no longer true. The five largest hyperscalers are expected to spend somewhere in the region of $745–775 billion on capital expenditure in 2026 alone, and the debt market has responded with structures that arrive earlier in the development cycle than they ever used to. The hard problem now is electricity.
This matters to anyone financing digital infrastructure, and it also matters to real estate sponsors who have never considered a data centre, because the same competition for power is reshaping utility economics, interconnection queues, and land values in the markets they operate in.
The timing mismatch that defines the sector
The structural problem is simple to state. A data centre can be designed, permitted, and built in roughly twelve to twenty-four months. Getting a large new load interconnected to the grid takes, in many territories, thirty-six to eighty-four months. Those two numbers are incompatible, and no amount of capital fixes the gap.
The scale of demand explains why the queues are what they are. In ERCOT alone, data centre interconnection requests reached roughly 356 gigawatts as of March 2026 — a figure that vastly exceeds what will actually be built, but which nonetheless has to be processed, studied, and queued. Nationally, data centres are on track to add on the order of 300 terawatt-hours of incremental annual electricity demand in the United States by 2030, which is the largest sustained demand increase the US grid has faced in decades.
How developers are responding
The industry's answer has been to stop waiting. Developers have announced on the order of 101 gigawatts of behind-the-meter natural gas generation in the United States, with a substantial share having placed equipment orders and a smaller tranche already under construction. More than 130 gigawatts of energy resources of various kinds are now proposed specifically to serve planned data centre projects.
In effect, a parallel generation system is being built to serve a single demand class. That has consequences for how these projects are financed, because a data centre with its own generation is not one asset. It is two, with different useful lives, different revenue mechanics, different regulatory exposure, and different natural lender pools.
What the capital markets have adapted to
Several financing patterns that were unusual a few years ago are now established practice.
- Debt arriving earlier in development. Facilities against land and pre-development cost, previously an equity-only stage, are now debt-financeable for sponsors with credible power and offtake positions.
- Equipment and hardware financing. Long-lead electrical gear — turbines, transformers, switchgear — has become financeable in its own right, partly because the lead times are now so long that securing equipment is itself a competitive position worth capitalising.
- Contracted-cashflow structures. Where a facility has creditworthy contracted revenue, financing is structured against that contract in ways that look more like project finance than real estate lending.
- Forward sale and monetisation structures that let developers realise value earlier rather than holding through stabilisation.
The common thread is that lenders are willing to take more development risk than they used to, in exchange for evidence on the one variable they cannot control. That evidence is power.
What lenders actually diligence now
If you are financing a data centre project, expect the power position to be examined before the real estate. In practice that means:
- Interconnection status in specifics — queue position, which studies are complete, what deposits have been posted, what the utility has committed to in writing, and what the realistic energisation date is rather than the optimistic one.
- The behind-the-meter or bridge generation plan, if there is one, including fuel supply, air permits, equipment procurement status and delivery dates, and how the facility transitions to grid supply if and when it becomes available.
- Offtake. Who is contracted to take the capacity, on what credit, for how long, and what happens on delay.
- Water and cooling, which in some markets is now nearly as constrained as power and considerably more politically sensitive.
- Local and political durability. Large loads attract scrutiny, and a project that depends on a contested approval carries schedule risk that a lender will price.
The spillover into conventional real estate
Sponsors with no interest in digital infrastructure are still affected, and it is worth understanding how.
Utility cost and capacity are becoming a differentiator between otherwise comparable markets. Industrial and flex sites with existing heavy electrical service have acquired an option value they did not previously have. Interconnection congestion slows every project in a queue, not just data centres, which matters for any development with meaningful new load. And in territories where large-load growth is driving generation and transmission investment, that investment shows up in tariffs — which is precisely why on-site generation and storage economics have improved for ordinary commercial property.
The practical implication for a real estate sponsor is that electrical capacity has become a diligence item on acquisitions where it used to be an afterthought. Knowing what service a site has, what it would cost and take to upgrade, and where the local utility sits on large-load requests is now part of underwriting.
Where this is heading
The binding constraint will stay on the electrical side for the foreseeable future, because generation and transmission take longer to build than either data centres or capital markets take to adapt. The projects that get financed on good terms will be the ones that treat power procurement as a first-order development activity — with the same seniority as land assembly and entitlement — rather than as a utility coordination task to be handled once the site is under control.
Frequently asked questions
Why is power, rather than capital, the constraint on data centre development?
Because the timelines do not match. A data centre can be built in roughly twelve to twenty-four months, while interconnecting a large new load to the grid takes roughly thirty-six to eighty-four months in many territories. Capital is available and has adapted with earlier-stage structures; deliverable electricity has not, and cannot be accelerated by financing.
What is behind-the-meter generation and why are data centres building it?
It is generation sited on the customer's side of the utility meter, serving the facility directly rather than delivering into the grid. Data centre developers are pursuing it at scale — roughly 101 gigawatts announced in the United States — because it can be delivered on a timeline compatible with the build cycle, bypassing the interconnection queue for the load they need to serve immediately.
Should I read headline interconnection queue figures as a demand forecast?
No. Queues include speculative and duplicative filings, with the same project queued in multiple territories or multiple sites filed to preserve optionality. The figures are reliable evidence of congestion and study burden, not of capacity that will be built. In diligence, what matters is the specific queue position, study status, and deposit history of the project you are financing.
How does this affect a sponsor with no data centre exposure?
Three ways. Interconnection congestion slows any project adding meaningful electrical load. Utility investment driven by large-load growth eventually shows up in tariffs, which improves the economics of on-site generation and storage. And sites with existing heavy electrical service have gained option value. Electrical capacity is now a genuine acquisition diligence item rather than an afterthought.
Sources
This article is general information about capital structures and financing mechanics. It is not investment, tax, accounting, or legal advice, and it is not an offer to sell or a solicitation to buy any security. Programme rules, statutory deadlines, and market terms change. Confirm anything you intend to rely on with advisers engaged on your specific facts.