New England and New York

Load growth in New England and New York is outpacing planned supply additions, and sustained delays in offshore wind have prompted near-term diversification that increasingly includes natural gas. Both states carry aggressive clean energy mandates that sit in direct conflict with near-term reliability needs, and that tension is shaping procurement decisions with long-term consequences for the resource mix. ISO-NE’s forward capacity market is under strain, with ongoing debate about whether the current structure can attract the investment the region requires. With winter peaks and tight capacity margins, buyers are ordering long-lead equipment earlier, particularly transformers, and in some cases paying premiums for firm supply or onsite generation to secure schedules.

The concentration of data center development in Northern Virginia has created localized transmission constraints and land-use friction among the most acute in the country. PJM's interconnection queue carries hundreds of gigawatts of proposed capacity, a significant share of which will not advance to completion, complicating supply adequacy planning for both developers and utilities. FERC Order 1920, issued in May 2024, requires PJM to conduct long-term transmission planning over a 20-year horizon using scenario-based analysis across seven defined categories. PJM made its formal compliance filing in December 2025, though the planning process is not expected to be operational until mid-2027 at the earliest, and the rule faces ongoing legal challenges that add further uncertainty to the timeline. Separately, FERC issued a co-location order in December 2025 directing PJM to develop new transmission service options for large loads, such as data centers, co-located at generating facilities, a mechanism that could allow some hyperscale customers to bypass traditional interconnection queues and accelerate their path to power.

TVA and neighboring utilities are absorbing significant industrial load growth alongside a rapidly expanding data center sector. Data center demand climbed to 18% of TVA's industrial load in 2025, with the utility projecting that figure to double within its region by 2030. Georgia Power received approval from state regulators in December 2025 for nearly 10,000 megawatts of new energy infrastructure primarily to serve projected data center demand, a scale of procurement that has drawn scrutiny over whether the underlying load forecasts are realistic. Demand-response programs are expanding, but 24/7 data center operations and advanced manufacturing generally lack the flexibility to participate in curtailment-based programs, pushing customers to pair grid supply with onsite natural gas reciprocating engines or smaller gas turbines, typically filed as separate interconnections to preserve schedule flexibility. On the longer-term supply side, TVA signed the first US utility power purchase agreement with an advanced nuclear plant when it committed to purchase power from Kairos Power's Hermes 2 SMR reactor, expected online at Oak Ridge, Tennessee by 2030, and separately announced a 6 GW small modular reactor deployment program framed explicitly around serving regional data center load growth.

ERCOT’s market structure and siting flexibility continue to attract large load development, though supply chain constraints and evolving interconnection timelines are moderating some of that advantage. PV module sourcing and tariff exposure have created schedule pressure on large projects, contributing to diversification toward storage and onsite thermal generation. During periods of high wholesale pricing, economic incentives alone have prompted gigawatt-scale voluntary load reduction, demonstrating genuine latent flexibility in the ERCOT load base. That response has been driven by economics during stress events rather than formal VPP aggregation frameworks, and ERCOT’s ancillary services market remains relatively immature for structured VPP participation.

MISO and SPP are evaluating extra high voltage AC (765 kV) as a supply-chain-aware alternative to HVDC for backbone expansion, with the maturity of domestic 765 kV equipment supply chains factoring into that assessment. MISO’s Long Range Transmission Planning (LRTP) process is one of the more substantive grid expansion efforts currently underway in the US, with multiple project tranches aimed at improving deliverability for large new loads and integrating renewable generation. Increased FACTS and STATCOM deployments reflect the need to stabilize weaker grid segments and manage power quality challenges from a growing concentration of power electronics loads.

West and Mountain West

New data center and large load development is expanding into Wyoming, the Dakotas, and other Mountain West states. CAISO remains one of the most operationally complex grid environments in the country, managing steep ramp requirements from the duck curve as solar output peaks through midday and demand rises sharply in the early evening. Integrating the volume of renewables that California’s policy targets require while maintaining reliability is an ongoing challenge shaping how large-load customers approach onsite generation and storage decisions. Wildfire risk adds a further layer of transmission vulnerability that is underweighted in most national grid planning discussions. Lines crossing high fire-risk terrain face both physical exposure and the risk of planned de-energization events, creating real availability uncertainty.

Resilient societies and liveability

Jonathan Howie

April 15, 2026

Securing tomorrow’s grid: How to navigate evolving power challenges across the US

In this article, Jonathan Howie, Senior Managing Consultant, Energy & Renewables, discusses the overall challenges facing the industry, breaks down how different regions are addressing the issues, and layouts what the industry should be doing now to prepare for a secure power supply in the future.

Share on facebook Share on linkedin Share on twitter

The US power grid was built for slow, predictable load growth. However, that assumption no longer holds. Hyperscale data centers are connecting to grids not designed for gigawatt-scale, always-on demand. Manufacturers are electrifying industrial processes that previously ran on gas. Electric vehicles and heat pumps are reshaping residential load curves faster than most utility planning models anticipated. The result is sustained structural pressure across the system, driven by a fundamental shift in how electricity is consumed. In this article, we outline regional power challenges across the country. But each challenge reflects a variation of the same underlying tension, with approaches varying by resource mix, grid topology, and policy environment.

Key observations

The combination of demand-side flexibility tools, including virtual power plants (VPPs) and behind-the-meter assets, with physical grid investment in transmission, substations, and power quality infrastructure such as Flexible AC Transmission Systems (FACTS) devices including STATCOMs will be critical.
Nuclear is re-entering long-term procurement discussions as a credible option for customers requiring firm, around-the-clock carbon-free power.
The most effective project strategies run mixed supply tracks in parallel, secure long-lead equipment ahead of utility process timelines, and account for region-specific risks such as water availability in the Southeast and wildfire exposure in the West.

What challenges does the US power sector face in 2026?

Speed to power and interconnection reform: Developers and large-load customers are racing to secure interconnection positions, often advancing preliminary engineering to strengthen queue applications amid evolving reform processes at several RTOs and ISOs. Multi-year study timelines remain the norm. New-to-transmission customers, particularly in the technology and real estate sectors, frequently need guidance navigating requirements for mixed onsite and grid supply strategies.

Supply chain constraints: Transformers remain on long lead times, with quality variability an increasing concern. PV modules and inverters are sensitive to tariff and domestic-content changes, while large gas turbine deliveries can stretch past five years. Many projects now order long-lead equipment in parallel with utility processes to protect schedules, with some developers reserving hot spares or pursuing refurbished equipment options.

Grid technology upgrades: The proliferation of large power electronics loads, including data centers, EV chargers, and variable-speed drives, is injecting harmonics and reactive power swings into transmission systems not designed to absorb them at this scale. FACTS devices and STATCOMs are increasingly standard infrastructure at major load interconnections. A STATCOM can inject or absorb reactive power within milliseconds, stabilizing voltage profiles that would otherwise deteriorate under the non-linear demand patterns of a hyperscale campus. Regions are simultaneously evaluating HVDC versus extra high voltage AC (765 kV) for backbone expansion, with supply chain availability influencing that choice.

Demand flexibility and market signals: VPPs are scaling from pilot programs toward gigawatt-level aggregation, and price-responsive load is becoming more operationally visible. During ERCOT’s 2021 winter stress event, market price signals alone prompted gigawatt-scale voluntary load reduction, demonstrating the latent flexibility available even without formal ancillary services frameworks.

Reliability, harmonics, and onsite solutions: Around-the-clock data center operations struggle to curtail for dynamic response programs, pushing interest in onsite generation (recips and smaller gas turbines), power electronics, and power-quality solutions. Mixed supply (grid plus onsite) is often best pursued as parallel interconnection tracks to de-risk timing.

Nuclear power: Interest in nuclear has increased, driven by hyperscalers seeking firm, around-the-clock carbon-free power that intermittent renewables cannot reliably provide. The recommissioning of shuttered plants and growing investment in small modular reactors (SMRs) reflect a shift in how large load customers are approaching long-term procurement. For data center operators requiring continuous high-availability power, nuclear is attracting serious consideration that was largely absent five years ago.

How are states addressing power limitations?

What does the power industry need to do to stay ahead of demand?

Demand is growing faster than most planning frameworks anticipated, and closing the gap between projected need and deliverable capacity requires coordinated action across developers, utilities, regulators, and large-load customers.

1. Utilities need to modernize both how they plan and how they operate.

Utility planning and operations are lagging behind the pace of demand change on two fronts. First, integrated resource plans built on 1-2% annual load growth assumptions are being rendered obsolete in data center corridors where growth is arriving an order of magnitude faster. Forecasting models need to be rebuilt around actual interconnection queue data and signed customer commitments rather than historical trends. Second, and equally important, utilities need to extract more from the infrastructure they already have. Dynamic thermal ratings, real-time grid modeling, and topology optimization can meaningfully increase the capacity of existing lines without new steel in the ground. VPP and demand response programs, properly structured, give utilities a dispatchable resource that sits inside the load itself.

2. Large-load customers need to own their power strategy earlier while continuing to assess load flexibility.

Hyperscalers and industrial customers are increasingly showing up to interconnection queues without adequate preparation, treating power as a late-stage infrastructure problem rather than a site selection variable. The most effective operators are now embedding power procurement and grid strategy into decisions made 3–5 years before a campus goes live. Load flexibility is an underutilized lever in this planning process, particularly for AI data centers. While the perception is that AI workloads demand uninterrupted, always-on power, the reality is more nuanced. A meaningful share of AI compute workloads, including model training, batch inference, and data preprocessing, are not inherently time-sensitive and can be shifted to off-peak periods without material impact on operations. A data center that is designed from the outset to tolerate even modest curtailment, in the range of less than 1% of annual operating hours, can materially reduce its interconnection cost, accelerate its path to energization, and provide utilities with a dispatchable resource that improves overall grid stability. The operators who will have the greatest advantage in securing power at scale are those who engage with utilities early, understand their own workload characteristics in sufficient depth to offer genuine flexibility commitments, and structure their power agreements accordingly.

3. The skilled labor shortage is an underappreciated execution constraint.

Schedule slippage on well-funded projects with equipment on order and permits in hand is increasingly being attributed to one cause: not enough qualified people. The shortage spans the full project lifecycle. At the engineering level, experienced power systems engineers, protection and controls specialists, and substation designers are in short supply at a time when the volume of complex interconnection projects is at an all-time high. Many engineering firms are turning away work or extending delivery timelines not because of capacity constraints in their offices but because the talent pipeline has not kept pace with demand. In the field, qualified electrical workers, high-voltage commissioning crews, and substation erection teams face similar pressure. This bottleneck rarely appears in integrated resource plans or regulatory filings, but it is consistently raised by developers, EPCs, and utilities as one of the most binding constraints on actual project delivery.

Conclusion

Meeting demand that is outpacing legacy planning will require pairing near-term "grid pragmatism", including VPPs, mixed onsite-plus-grid supply, FACTS and STATCOMs, and dynamic grid operations, with long-lead investment in transmission, substations, and firm low-carbon options such as nuclear. If utilities, regulators, developers, and large-load customers plan earlier, run parallel interconnection and procurement tracks, and address the skilled labor bottleneck, the US can deliver a resilient, reliable, and cleaner grid that keeps pace with demand growth.

Want to know more?

Jonathan Howie
Sr Managing Consultant, Energy & Renewables

RELATED INSIGHTS

Jan 28, 2026
Ramboll inventors revolutionize power grid stability with “universal shock absorber”
Ramboll inventors create “universal shock absorber” for the grid, making it easier and less risky to add renewable energy sources to the grid, preventing power outages, cutting costs of developing power systems around the world, and increasing resilience.
Decarbonise for net zero | Feb 19, 2026
Key considerations for land developers entering the data center market in 2026
The global appetite for cloud services, AI training, and edge computing continues to surge, making data centers one of the most sought-after real-estate asset classes. 2026 promises to be a year of growth and development for data centers in the US – with several large projects already in the works or planned. In this article, we discuss the key factors that land developers should consider if they are planning to enter the data center market.
Article
Decarbonise for net zero | Sep 25, 2025
Gridlocked: What is holding up tomorrow’s resilient electricity grid?
There is no green energy transition without a large-scale grid build-out. Dr Mohamed Rashwan, managing director of TransGrid Solutions (TGS) that was recently acquired by Ramboll, explores the issues that can impact grid expansion – from too few suppliers to too many permits - and how to overcome them.