© photology1971 - stock.adobe.com
AI and EV Boom Could Outrun the Copper Behind Power Grid Equipment
In A Nutshell
- Under a high-growth scenario, the US could fall about 28% short of needed power grid equipment by 2030.
- Copper becomes the first metal to run out, starting in 2027, with steel and nickel tightening later.
- Cutting metal supplies from less predictable trading partners by 70% widens the gap to 36%.
- A transformer-focused fix trims the gap to about 22%, but the results are scenarios, not forecasts.
Every new data center, electric vehicle charger, and solar farm needs a hidden middleman to plug into the power grid, and the US may not be able to build enough of them. A new study finds that if electricity demand surges, the country could end up about 28% short of that equipment by 2030, with copper running out first.
These middlemen include transformers, which adjust voltage so power can travel safely, and inverters, which turn solar and battery power into the kind of electricity a home can use. Data centers also rely on backup power systems. Johns Hopkins University researchers call all of it grid-supporting equipment, and their analysis in the journal Cell Reports Sustainability asks whether the metals needed to build it can keep pace with demand. Grid planning usually counts power plants and transmission lines, then assumes the hardware in between will simply show up.
Early signs say otherwise. Wait times for a new transformer have stretched to as long as two years, four times longer than before 2022, according to figures the authors cite. Much of the equipment already in place is also old and needs replacing, which piles on top of new growth.
Old Equipment and Data Centers Drive Grid Equipment Demand
This was a computer modeling study, so no people were surveyed. Authors Boyu Yao and Yury Dvorkin used federal records of US power generators to rebuild how much grid equipment has been installed over the years, then estimated how much is wearing out. They tested two versions: one where equipment lasts about as long as expected, and one where heavy use wears it out sooner. Transformers are the oldest equipment on average, so a growing share of their future demand comes from replacement.
Next, the team mapped out two possible futures. A baseline scenario assumed electricity use rising 1.8% a year and 33 million electric vehicles on the road by 2030. A high-growth scenario assumed 2.8% yearly growth, a bigger data center boom, and 42 million EVs. Data centers alone could consume 9% to 17% of US electricity generation by 2030, according to figures the study cites.
Every machine has a metal recipe, so the researchers compiled average ingredient lists for each type. A standard transformer is mostly steel, with some copper, and converters for one type of wind turbine generator were the most copper-hungry of the equipment studied. A global trade database then showed where the metals come from. Steel and aluminum are mostly sourced in the US, at about 79% and 75% by 2030. Copper is slipping, with the domestic share falling from around 70% in the mid-2010s to roughly 61% by 2030 as more arrives from Canada, Mexico, and elsewhere. A final model checked each year whether the available metal could cover the equipment needed.
Copper Shortage Begins in 2027, and High Growth Pushes Gaps to 28%
In the baseline scenario, the model showed no shortfall in 2025 or 2026. By 2027, about 7% of needed equipment could not be built, and by 2030 that gap grew to 16%. Under high growth, the shortfall reached about 28% in 2030, whether equipment lasted a long time or wore out early.
In both scenarios, the model used up all available copper from 2027 onward. By 2030 under high growth, steel and nickel were close behind, with 97% and 93% of their supply spoken for. A test giving grid equipment more copper eased the gap, but the squeeze simply shifted to steel.
Transformers were the last to run short, partly because the model protects them first, so the result says little about how sturdy transformer supply is in real life. In the baseline, early gaps showed up in wind turbine converters, data center equipment, and EV charging hardware. Delays for data center and charger projects would be a reasonable knock-on, though the study did not test that. By 2030 under high growth, even transformers came up 14.5% short.
Trade Disruptions Deepen the Copper Shortage, but Transformer Technology Offers Only Partial Relief
Two what-if tests followed. In the first, metal supplies from less predictable trading partners, meaning most countries outside NATO, other close US allies, and North American trade deals, were cut by 70%. That pushed the high-growth shortfall from 28.5% to 36%, and transformers took the biggest hit, with their gap jumping from 14.5% to 25%.
A technology called dynamic transformer rating, which squeezes more use out of transformers already in place, was assumed to cut new transformer needs by 10% in the model. That trimmed the overall shortfall to about 22%. It barely helped inverters, converters, and other electronics, and copper stayed maxed out.
Authors describe these numbers as scenarios, not forecasts, and the model uses metal supply as a stand-in for factory capacity, so real life could land better or worse. Even so, a grid plan that counts power plants but skips the copper behind the hardware overlooks the first resource the model shows running dry.
Disclaimer: This article summarizes a published scientific study for general informational purposes only. Its figures come from modeled scenarios, not confirmed outcomes, and it is not professional engineering, investment, or policy advice. Readers should consult the original paper for full details.
Paper Notes
Limitations
Metal supply serves as a stand-in for real manufacturing capacity, since factory output, skilled labor, and specialized parts are often private or unavailable. Actual shortages could be worse if factories cannot keep pace even when metals are on hand. The ranking in which transformers are protected longest comes partly from a priority rule built into the model. Material recipes are class-level averages rather than exact product specifications, and the share of the broader electrical machinery sector assigned to grid equipment is an estimate; the historical low and high bounds moved the 2030 high-growth shortfall between 36.7% and 21.4%. Recycling, refurbishment, and part reuse were not modeled. The authors call the results conditional, scenario-based indicators of potential supply risk rather than direct forecasts.
Funding and Disclosures
The work was supported in part by the Johns Hopkins University Discovery Award and in part by the US National Science Foundation under grant OISE-2330450. The authors declare no competing interests. They also disclosed using ChatGPT for grammar checking and LaTeX formatting, and state that they reviewed the content and take full responsibility for it.
Publication Details
Title: “Grid-Supporting Equipment Supply Chains Constrain the Feasible Pace of Power System Expansion” Authors: Boyu Yao and Yury Dvorkin, Johns Hopkins University (Department of Civil and Systems Engineering; Department of Electrical and Computer Engineering; Ralph O’Connor Sustainable Energy Institute) Journal: Cell Reports Sustainability, volume 4, article 100923; available online November 17, 2026, in an issue dated January 22, 2027 DOI: https://doi.org/10.1016/j.crsus.2026.100923







