The US power grid has a delivery problem. In 2025, grid congestion cost US consumers more than $17 billion, up from $12 billion just a year earlier. When transmission constraints prevent lower-cost generation from reaching demand, it follows that more expensive generation must run instead[1].
The common response is to build more transmission. A necessary response, but also only part of the solution. A less discussed reality is that much of today’s congestion is not a constant problem. According to the US Department of Energy, most congestion is concentrated in just 5% of hours[2].
It’s not dissimilar from a typical highway that works perfectly well most of the day but becomes congested during the busiest periods. You could build more lanes to remove those bottlenecks, but that would mean investing heavily in extra capacity that is only really needed for a small part of the day. Another approach is to use the existing infrastructure more efficiently and manage traffic better when pressure peaks.
When people think about congestion, they often think about physical bottlenecks like in the analogy above, however, that same logic increasingly applies to the power system. Congestion in the power grid is typically driven by high net load, cold weather, high shares of intermittent generation, and larger differences between day-ahead and real time prices [2]. The result is that the most valuable hours in modern power markets are increasingly determined by timing rather than volume.
When congestion is concentrated in a small number of hours, the challenge is therefore not only about building more infrastructure. Alleviating congestion is increasingly about making better use of the infrastructure that already exists through optimising, leveraging flexibility and dispatching energy resources more efficiently instead of only building more lanes.
The grid is increasingly constrained at exactly the wrong moments
Texas offers a useful example, endowed with some of the world’s most competitively priced renewable resources. Large volumes of low-cost wind generation are plentiful across West Texas and the Panhandle, while demand growth is increasingly concentrated elsewhere, driven by data centres, industrial activity, and electrification.
The result is a familiar pattern, hardly isolated to the Lone Star State. Across PJM that serves approximately 67 million customers in 13 U.S. states and the District of Columbia, congestion costs reached about $3.2 billion in 2025[3]. During summer heat waves, similar transmission constraints frequently limit deliveries of lower-cost power produced in the west to East Coast demand centres.
Adding new capacity is not particularly easy either. At the end of 2025, more than 2,000 GW of generation and storage projects were waiting in interconnection queues, while projects reaching commercial operation had spent a median of more than five years progressing through the process[4].
Demand is growing today, while much of the infrastructure needed to accommodate it will not arrive for years. It’s a reality that only underlines that growing need for optimisation skills.
Trading moves energy across space and time
When we look at congestion and the price differences it creates, we see them not just as symptoms of constraints in the system, but as signals that help the market respond. Price differences reveal where electricity, flexibility and storage are most valuable at a given point in time.
That makes the ability to anticipate where renewable output will be highest, where demand will emerge, how transmission constraints may develop and where prices could separate across locations increasingly important. Forecasting, analytics and optimisation therefore play a growing role in how market participants navigate the physical realities of the grid.
Market participants respond to these signals by adjusting schedules, optimising flexible assets, charging and discharging storage, and positioning generation according to changing system conditions. Collectively, these actions can help make better use of the infrastructure already in place, however, doing this effectively increasingly depends on sophisticated forecasting and automation. Modern energy trading uses predictive analytics, machine-learning-based forecasting and algorithmic execution to respond continuously as weather, demand, generation and market conditions change.
The objective is straightforward: reduce surprises, improve dispatch decisions and use available transmission capacity more efficiently. Trading cannot remove a physical constraint in the grid. But by revealing its economic impact and enabling market participants to respond, markets can help ensure that the capacity we already have is used as efficiently as possible.
A familiar challenge on both sides of the Atlantic
While Europe and the US have very different market structures, they face remarkably similar challenges. Electricity demand is growing, renewables are scaling quickly, infrastructure development is struggling to keep pace, and congestion, curtailment, and connection queues are becoming familiar features of both systems.
Europe has spent the past two decades learning how to operate increasingly interconnected and renewable-heavy power markets. The lesson has been clear: building infrastructure matters, but so does maximising the value of infrastructure that already exists.
The European Energy Regulator ACER estimated that cross-border electricity trade delivered €34 billion in economic benefits in 2021[5]. The same principle applies to any grid under pressure: when power can move to where it is valued most, scarce capacity goes further.
So, where does that leave US?
There is no doubt that new transmission infrastructure is essential, and significant investment is already underway. The U.S. Department of Energy points to major transmission portfolios recently approved by MISO, SPP, PJM and ERCOT, including MISO’s largest-ever portfolio in 2024. But new transmission takes years to plan, permit and build, while demand from data centres, manufacturing and electrification is growing now.
That makes better use of the existing system increasingly important. Better forecasting can provide a clearer picture of where supply and demand are heading and where constraints are likely to emerge. Flexible assets can respond as conditions change, while trading responds to price differences across locations and time, helping direct electricity and flexibility towards where they are most valuable.
The US therefore does not face a choice between building more grid infrastructure and making better use of what is already there. It needs both. New transmission provides the additional physical capacity the system will need over the longer term. In parallel, forecasting, trading, optimisation and flexibility can help the existing system respond more efficiently to changing conditions.
This is where experience from other increasingly complex power markets becomes relevant. Over more than two decades, Centrica Energy has built capabilities in forecasting, trading, optimisation and flexibility across European markets shaped by growing renewable generation and greater interconnection. The point is not to transplant a European market model to the US. It is to apply that experience to a challenge both markets increasingly share: how to get more from an increasingly complex power system while the infrastructure around it continues to evolve.
Sources
[1] Grid Strategies (2026), Transmission Congestion for 2025
[2] U.S. Department of Energy (2026), National Transmission Needs Study, Draft for Consultation and Public Comment
[3] US Grid Congestion Hit $17.1 Billion in 2025, Below the 2022 Peak, With PJM at $3.2 Billion, mgrid.org
[4] Lawrence Berkeley National Laboratory (2026), Queued Up: 2026 Edition
[5] ACER (2022), Final Assessment of the EU Wholesale Electricity Market Design
