Why is the US power grid running out of spare capacity?

The US power grid has just 3% spare generating capacity left. That's the headline finding from a new ICF report, and it leaves almost no buffer against demand spikes, extreme weather, or unplanned outages. In the country's two fastest-growing electricity markets, there's effectively no buffer left at all. When we put the question to our network, 52% said modernising the grid was the most urgent priority, ahead of new generation at 30%, with talent and permitting reform tied at 9% each.

That result is worth examining closely, because the more revealing part of this story isn't the capacity number. It's that the industry has nearly $1.3 trillion committed to fixing it, and may still not have enough people to spend it on schedule.

How thin is the margin, really?

ICF's report, Electricity Demand Growth: How Will the Grid Keep Pace?, finds the US power system currently has only about 26GW of excess generating capacity above minimum resource adequacy requirements, roughly 3% of total capacity. ERCOT and PJM, the nation's fastest-growing electricity markets, have no excess generating capacity available to reliably support additional demand growth at all.

The pressure isn't easing on its own. In PJM and ERCOT, the cushion has largely disappeared already, as load growth has outpaced resource additions and any pre-existing excess capacity has been quickly absorbed. Total US electricity demand is expected to rise 21% by 2030 and 39% by 2035, with peak demand climbing 14% and 25% over the same periods respectively, growth concentrated in exactly the regions seeing the most data centre, industrial, and electrification activity.

New capacity is coming, but not fast enough to rebuild much of a buffer. Roughly 445GW of new generation is forecast to be added between 2026 and 2030, but only 68GW is expected online in 2026, and margins are projected to stay low through the decade, with excess capacity recovering to only around 20GW by 2030. Not all of that new capacity counts equally toward reliability during peak demand either. Intermittent solar and wind contribute less to the margin than firm generation does.

If it's not a capital problem, what is it?

This is the part of the story that gets less attention than it should. Utilities are not short of money to spend on this. S&P Global's Regulatory Research Associates forecasts approximately $1.3 trillion of aggregate capital expenditure across US energy utilities between 2026 and 2030, a record, and roughly a 29% increase on the prior year's forecast, driven largely by data centre and large-load demand.

Capital at that scale doesn't build itself. It has to be designed, engineered, permitted, constructed, and commissioned by people. That's where the real shortage sits, and it's a finding our poll results reflect almost exactly: the industry has clearly settled on modernisation over new generation as the priority, but talent and permitting, the two mechanisms that actually deliver modernisation, ranked lowest of all four options at 9% each.

The workforce gap behind the capacity gap

A joint study by consulting firm Kearney and the IEEE Power & Energy Society finds the global power sector will need between 450,000 and 1.5 million more engineers by 2030 to build, implement, and operate the infrastructure this investment is paying for, more than doubling today's workforce in some estimates. Already, 40% of power executives report significant difficulty hiring skilled workers, citing talent competition and insufficient skills as the primary barriers.

Retention is working against the industry just as hard as recruitment is. Almost half of all power engineers have either changed jobs within their company, moved to another employer, or left the industry altogether in the past three years. The reasons cited include burnout and limited scope for interesting, challenging work, not just pay. In the nuclear sector specifically, where reliability demands are the strictest anywhere in the industry, that churn rate reaches 58%.

The pipeline behind the current workforce isn't refilling fast enough either. University enrolment in power engineering programmes has stagnated, as engineering students increasingly gravitate toward data science, software, and AI, fields seen as more current and better paid. The result is an industry trying to build out $1.3 trillion of infrastructure with a workforce that's both shrinking and ageing out faster than it's being replaced.

Two crises, one root cause

The 3% capacity margin and the engineer shortfall aren't separate problems running in parallel. They're the same problem viewed from two different angles. A thin capacity margin means every delay in bringing new generation and transmission online carries real reliability risk. And every delay in hiring, training, or retaining the engineers needed to deliver that generation and transmission is exactly the kind of delay a 3% margin can't absorb.

Our poll result puts a number on how the industry currently weighs this. Modernisation is seen as the clear priority. Talent, the thing that actually delivers modernisation, isn't yet being treated with the same urgency.

What needs to happen

• Treat workforce planning as seriously as capital planning. A $1.3 trillion build-out needs a matching hiring and retention strategy, not an afterthought.
• Address burnout and career progression directly, since compensation alone isn't stopping a churn rate approaching 50%.
• Invest in the pipeline now, given university enrolment in power engineering isn't recovering on its own.
• Prioritise nuclear retention specifically. A 58% churn rate in the highest-reliability part of the sector is a compounding risk, not just a statistic.

Spencer Ogden partners with utilities, developers, and grid operators to build the engineering teams this next phase of capacity expansion depends on. Get in touch to talk through your grid and power engineering workforce strategy.