A power project has approval, investors and a construction schedule. It can still spend years waiting to connect to the grid. Electricity has to travel through a network with real limits on where and when additional supply or demand can be accommodated. New power stations and data centers therefore create questions about substations, transmission, queues and who pays for upgrades. The capacity to generate electricity is not identical to the capacity to deliver it.

A new solar farm can have financing, land and equipment ready while the project remains unable to sell much electricity. A factory may need more power than the nearby network can reliably deliver. A data centre can secure its building and servers only to discover that grid access will take much longer than construction.

These projects appear unrelated. They share a physical constraint: electricity networks have limited capacity at particular places and times. Installing more generation equipment does not automatically solve the challenge of moving power across those networks.

The problem has become large enough to influence energy policy and investment decisions. The International Energy Agency's Electricity 2026 report estimates that more than 2,500 gigawatts of renewable, storage and large-load projects, including data centres, are stalled in grid connection queues worldwide. This figure describes an assortment of proposed supply, storage and demand projects. It is not 2,500 gigawatts of completed power plants waiting for someone to flip a switch.

That distinction matters because connection queues measure proposed demand for grid access. Some projects may be speculative, change size, secure another location or never be built. Even so, a congested queue can signal a genuine problem for projects that are ready to proceed.

The grid is a network with local constraints

Electricity has to be balanced continuously across a network. Operators must keep supply and demand aligned while staying within the physical limits of lines, transformers and other equipment. A power plant producing inexpensive electricity far away may be of limited help if the lines between that plant and a city are already heavily loaded.

The network also has a geography. New wind or solar facilities are often planned where natural resources and land are favorable, not necessarily near existing industrial demand. Meanwhile, a growing city or data-centre cluster can need large amounts of electricity in an area where substations and transmission corridors were designed for much smaller loads.

A grid connection therefore involves more than approving a cable. Operators must study whether a new generator, battery or major consumer will affect flows, voltages, system stability or the ability to maintain service during equipment failures. If the network needs reinforcement, developers and utilities must determine how the costs, timing and risks will be shared.

All of that can continue after the generating equipment itself is technically straightforward to build.

Why new electricity is not automatically cheap electricity

People often judge an energy project by its generation cost: how much it costs to produce a unit of electricity. That is useful, but consumers pay for an entire system. The final bill also reflects networks, reliability requirements, taxes and other charges, with different market rules across countries.

The least expensive plant on paper may require a costly transmission upgrade to serve demand. Electricity delivered at the wrong hour may require storage or flexible demand to be useful. And a new facility can fail to lower local prices if congestion prevents it from reaching the area where electricity is scarce.

Grid upgrades may be expensive and slow, but avoiding them has costs as well. Congestion can force operators to limit output from otherwise available power plants, a practice known as curtailment. It can also require costlier generation closer to demand. Neither outcome is captured by simply adding up the capacity of new projects.

The IEA's 2026 analysis reports that grid investment has lagged behind investment in generation. It estimates current annual global grid investment at approximately $400 billion and projects that meeting electricity needs through 2030 would require investment to grow by roughly 50% from that level by 2030. That is a planning estimate, not a guaranteed spending plan or a fixed tariff increase.

There is an economic tension here. Transmission and distribution networks have to be built before all the future customers and generators exist, yet the cost of construction must ultimately be financed. Waiting for demand to be certain can produce shortages; building indiscriminately can burden consumers with underused infrastructure. The timing and location of grid investment are therefore central decisions.

A connection queue is not a construction schedule

A developer normally needs studies, agreements and sometimes network upgrades before a proposed project can operate. In a crowded queue, multiple proposals may be assessed against an electrical system that is also changing as other projects enter or leave. The results can affect which upgrades are needed and who is expected to pay.

Many regions have tried to improve the sequence. In the United States, the Federal Energy Regulatory Commission's Order No. 2023 changed federal generator-interconnection procedures to address long backlogs, including studying groups of projects together and placing greater emphasis on project readiness. FERC issued a subsequent clarifying order, Order 2023-A, in March 2024. These are United States regulatory examples, not rules that apply to every electricity grid worldwide.

Studying proposals together can make sense when several facilities affect the same transmission area. Requiring developers to demonstrate readiness can also discourage speculative applications from filling the queue. But a faster study cannot manufacture missing transformers, a new high-voltage corridor or an unavailable workforce. Process reform and physical investment solve related, different problems.

That is why queue figures should be read carefully. A smaller queue can mean projects were connected, or merely that applications were withdrawn. A faster decision can help development even if a connection still requires expensive upgrades. Good reporting needs more than a headline total.

Using the network more intelligently

Building new lines remains essential, especially where demand is rising for years. But the electricity grid is not equally busy at every hour. Some lines have spare capacity when demand is low or when weather conditions allow equipment to operate more efficiently.

The IEA describes several ways of making better use of existing assets. One is a non-firm connection, under which a generator or large consumer can connect sooner but agrees to limits during constrained periods. A battery might reduce charging when the network is congested. A flexible industrial load might shift some operations to times when capacity is available. The terms depend on the grid and the economics of the customer.

Another family of options is grid-enhancing technology. Dynamic line rating uses conditions such as weather to estimate how much power a line can safely carry at a given time, rather than relying only on a conservative fixed limit. Advanced power-flow controls can help redistribute electricity through available paths. Reconductoring, which replaces existing line conductors, can increase capacity in some corridors without creating an entirely new route.

These tools have trade-offs. A non-firm connection can constrain revenue when a generator is told to reduce output. A data centre may have less flexibility to interrupt computing work than a storage operator has to shift charging. Hardware improvements also require engineering studies and regulatory permission; they are not an excuse to ignore safety limits.

The IEA estimates that regulatory changes and grid-enhancing measures could make enough room to connect roughly 1,200 to 1,600 gigawatts of advanced-stage queued projects worldwide. The range is a modelled potential, dependent on timely reforms and deployment. It should not be presented as capacity already unlocked.

The AI boom adds a concentrated source of demand

The grid question has moved beyond renewable generation. New electricity demand can arrive in large blocks, particularly when data centres cluster in locations with attractive land, fiber connections or commercial incentives.

According to the IEA's Energy and AI analysis, data centres accounted for about 1.5% of global electricity use in 2024, while their effects were much larger in specific local markets. The agency projected strong growth toward 2030, with the size of the increase depending on the development of AI and other digital services.

A small global share can still create a serious local engineering problem. A new load of substantial size must be connected at a particular place. Its demand may arrive faster than transmission lines can be approved and built. Local households, factories and future projects may then compete for constrained capacity or face higher network-upgrade costs.

This does not make every data centre a poor investment or every new power plant a solution. It strengthens the case for asking developers early how their projects will interact with the grid: where they will connect, how flexible they can be and whether they contribute to network costs.

Who pays for a faster connection?

Every proposed solution involves distributional choices. A utility can recover an upgrade through regulated charges, a project developer can pay some of the cost, or a government may provide financial support. The right answer depends on who benefits and how a country's electricity system is governed.

If a single factory needs a dedicated connection, charging that factory for the relevant work may be straightforward. A new regional transmission line may benefit many future users, including households who were not involved in requesting it. Spreading its costs more broadly can be justified, but requires transparency about expected benefits and the risk that forecasts are wrong.

Grid planning also affects fairness between places. Regions with existing infrastructure can attract new investment more easily, while remote areas with excellent renewable resources struggle to connect. Communities hosting large projects may bear construction disruption without seeing equal local benefits. A serious grid plan must acknowledge the people living along a proposed line, not only the megawatts at its endpoints.

What to watch in the next few years

A credible progress report would ask how long ready projects take to reach connection, how much new capacity can actually operate, how often congestion forces generators to curtail output and whether households or businesses face rising network costs.

It would also track practical reform: completed transmission corridors, new transformers, better connection studies and contracts that reward flexible use of electricity. A press release about installed generation capacity answers only one part of the question.

The energy transition is changing how electricity is produced and consumed. The grid determines how much of that progress can travel from a promising project to a useful socket. Building more generation is valuable; connecting it safely and affordably is the work that makes it count.

---

SOURCES & FURTHER READING

References and background

Our explainers link to source documents so readers can explore the evidence and interpret it in context. Websites and policies may change after publication.

  1. International Energy Agency, *Electricity 2026*, executive summarywww.iea.org
  2. International Energy Agency, *Electricity 2026*, grids chapterwww.iea.org
  3. Federal Energy Regulatory Commission, *Explainer on the Interconnection Final Rule*www.ferc.gov
  4. International Energy Agency, *Energy and AI*, executive summarywww.iea.org
Spot a factual issue? Request a correction ↗Read our editorial standards ↗