Reliability and the Grid's Future

Grid reliability gets attention for about three days after a blackout and then disappears until the next one. The decisions that cause or prevent the next one are made years earlier, in resource plans, transmission studies and rate cases that almost nobody outside the industry follows. That gap is where an engaged resident or advocate can do the most good.

Why the lights go out

Blackouts split into local and wide-area events. Local outages, the common kind, come from the distribution system: a tree on a line, a car into a pole, a failed transformer, a storm. Wide-area events come from the bulk system, either a cascade of failures or a shortage of supply that forces operators to shed load on purpose.

Look at the winter storm rows together, because they cut against the claims of every side in the energy argument. Gas plants failed in the largest numbers, and gas fuel supply froze up. Wind turbines froze too. Coal piles froze. Nuclear units were less than 1 percent of affected units in both storms, though one Texas reactor did trip during Uri. The common failure was cold-weather preparation across almost every fuel type, and the main fixes were winterization standards, gas and electric coordination, and better planning for extreme cold. NERC has since strengthened its cold weather reliability standards.

Extreme weather is the main stressor

Heat waves push summer peaks past forecasts. Cold snaps hit gas supply and raise heating load at the same time. Drought cuts hydro, which matters a lot in the Northwest: lower forecast hydro was the main driver of Idaho Power's 2026 power cost increase. Wildfire is its own category in the West, where utilities now plan public safety power shutoffs in high-risk conditions.

What NERC says about the next ten years

NERC's 2025 Long-Term Reliability Assessment, released January 29, 2026, was its starkest yet. Trade press coverage of the LTRA gives these top-line numbers:

• Summer peak demand projected to grow about 224 GW over 10 years, 69 percent more growth than the prior year's forecast. Winter peak growth about 245 GW. NERC called the growth rates the highest since its tracking began in 1995.
• 13 of 23 assessment areas carry an above-normal or high risk rating for resource shortfalls over the next five years.
• Data centers are the main driver of new load in many areas, with large industrial loads and electrification adding more.
• WECC-Northwest: peak load up about 6.6 GW (19 percent), much of it from data centers; risk normal for 2026 to 2028 and high beginning in 2029, with winter peaks and reliance on variable resources named as concerns. NERC also named WECC-Basin, along with MISO, PJM and ERCOT, among areas that could reach high risk by 2030.

A forecast is not a fact. NERC's numbers aggregate what utilities report, and utilities' large-load forecasts include many projects that may never be built or may be counted by more than one utility as developers shop for sites. Whether the forecasts are inflated is a real dispute. Some analysts and consumer advocates argue utilities have an incentive to overbuild against speculative load. Utilities and grid operators answer that the signed contracts and requests for service are real and that underbuilding risks blackouts. Both are reasonable concerns. The way to judge your own utility is to ask how much of its forecast is under signed agreement and how much is "pipeline."

The scale of data center demand

The Department of Energy's report estimated data centers used 176 TWh in 2023, about 4.4 percent of U.S. electricity, and projected 6.7 to 12 percent by 2028. The width of that range is the honest answer. Module 18, on data centers, covers how these facilities use power, and module 19 covers their local impacts.

The bottlenecks: wires and queues

Building generation is often faster than connecting it. Lawrence Berkeley National Laboratory's Queued Up report (2026 edition, data through 2025) counted over 2,060 GW of generation and storage seeking grid connection, more than the whole existing U.S. fleet of about 1,281 GW. Of capacity that requested connection from 2000 to 2020, only 13 percent was operating by the end of 2025; about 75 percent withdrew. Projects built in 2025 had a median wait of more than five years from request to operation.

Transmission is slower still. A major line can take a decade or more from proposal to energizing, crossing federal land, private land and multiple states. FERC's Order No. 1920 (2024) now requires transmission providers to plan at least 20 years ahead and update every five years. In Idaho, the Boardman to Hemingway 500 kV line between Oregon and southwest Idaho has been in development for close to two decades; Idaho Power's 2025 plan targets service by the end of 2027.

Reading an integrated resource plan

Most vertically integrated utilities must file an IRP every two years or so: a 20-year forecast of demand and a "preferred portfolio" of resources to meet it. In Idaho the PUC "acknowledges" a plan rather than approving it. Acknowledgment means the commission found the plan a reasonable working document. It does not pre-approve the costs; each investment still has to pass a prudence review in a later rate case.

Where to look in any IRP

1. Load forecast. Find base, high and low cases. How much comes from large loads, and how many of those have signed agreements? What assumptions about electric vehicles and heat pumps are built in?
2. Existing resources and retirements. Which plants retire or convert, and when? What happens to coal contracts?
3. Resource accreditation. The table assigning capacity credit to each resource type (lesson 4). Small changes here swing the whole portfolio.
4. Cost assumptions. Gas price forecast, cost of solar, wind and batteries, carbon cost assumptions if any. Compare to the last plan.
5. Market and transmission assumptions. How much does the plan count on importing power, or on a line that is not built yet?
6. Preferred portfolio vs. alternatives. IRPs usually test several portfolios. Find the cost difference between the preferred one and the runners-up, and the reasons given.
7. Action plan. The near-term list, often the next two to four years, is what the utility will actually build or buy first. It turns into requests for proposals and later into rate cases.
8. Comments. Read staff and intervenor comments in the IRP docket. They point at the weakest assumptions.