This covers the 2026 US policy fight over amorphous-core transformer mandates — who’s pushing to keep them, who’s pushing to unwind them, and what it signals for buyers of amorphous and grain-oriented electrical steel outside the US. It doesn’t cover the metallurgy of why amorphous cores behave differently from silicon steel — we’ve written about that separately, linked below.
Core Key Points
- In April 2024, the US Department of Energy finalized a rule letting 75% of new distribution transformers still use grain-oriented electrical steel (GOES) cores, with the rest expected to shift toward amorphous metal — a compromise down from a January 2023 proposal that would have pushed roughly 90% of the market onto amorphous cores.
- In June 2026, DOE opened a formal public comment period reconsidering that same 2024 rule, asking specifically how it affects “domestic manufacturing capacity, supply chain resilience, and the availability and cost of key materials.”
- Cleveland-Cliffs, the only domestic GOES producer in the US, supported the 2024 rule at the time. In 2026, it reversed course, arguing the amorphous requirement will weaken domestic supply chains and increase reliance on imported material.
- Metglas, the sole US amorphous-metal producer, runs one plant in Conway, South Carolina, at roughly 45,000 tonnes of annual capacity — and its parent company, Proterial, is now expanding amorphous production in India instead, not the US.
- On April 20, 2026, a presidential determination under the Defense Production Act named grid infrastructure — including “electrical core steel” specifically — essential to national defense, citing distribution transformer backlogs that had reached 12+ months.
- The 2029 compliance deadline for the original rule hasn’t moved. What’s genuinely uncertain in mid-2026 is whether the rule itself survives intact, gets loosened, or gets replaced before that date arrives.
We’ve been tracking this file since the January 2023 proposal first landed; honestly, the whiplash between then and now still catches us a little off guard.
What the 2024 Rule Actually Did
To follow what’s happening in 2026, it helps to know what the 2024 rule actually did. DOE’s January 2023 proposal would have required amorphous cores in roughly 90% of new distribution transformers.
Manufacturers pushed back hard. GOES capacity, tooling, and stamping lines were already built around silicon steel. A near-total switch to wound amorphous cores in three years wasn’t a design change; it was a rebuild.
The April 2024 final rule softened that considerably. It allows 75% of the market to keep using GOES cores, with only a smaller efficiency tier requiring amorphous metal or equivalently low-loss designs. Compliance was pushed out to 2029, five years out instead of three.
DOE framed it as a middle ground: real efficiency gains, without forcing existing GOES production lines into early retirement. Utility groups and manufacturers both signed off. For a while, this looked settled.
So what changed? It didn’t stay settled. By June 2026, the same rule was back under formal review, and not from environmental groups pushing for more amorphous. The pressure came from a very different direction.
June 2026: DOE Asks If It Got This Wrong
In June 2026, DOE published a request for comments revisiting the 2024 rule. The stated focus: whether the amorphous-core requirement is straining domestic manufacturing capacity and material supply, not whether it cuts enough energy loss.
That framing matters. This isn’t a dispute about whether amorphous cores work; the physics were never really in question. It’s a dispute about whether the US can actually build enough amorphous-core transformers on the timeline the rule assumes.
Reactions split cleanly by role, which is usually how you can tell a policy fight is real rather than performative. The Edison Electric Institute, representing utilities that actually buy these transformers, opposed rolling the rule back outright. Its position: “compliance flexibility preserves the efficiency gains,” paired with a request for even more time given how much supply conditions have worsened since 2024.
The National Electrical Manufacturers Association took the opposite stance: leave the rule alone. Its argument was that transformer manufacturers had already built 2029 compliance plans around the current numbers, and reopening it now just adds uncertainty nobody asked for.
Both of those reactions are what you’d expect from a utility group and a manufacturer group. But which side did the sole domestic GOES producer take? Not the one you’d guess.
Cleveland-Cliffs Changed Its Mind — That’s the Surprising Part
Cleveland-Cliffs is the only company that makes GOES domestically in the United States. In 2024, it backed the amorphous-inclusive rule. That made sense at the time; a rule that reserves 75% of the market for GOES is not a threat to a GOES producer.
By 2026, Cleveland-Cliffs had reversed that position entirely. Its argument now is that the amorphous-core requirement, even at a partial 25% share, pushes US transformer manufacturers toward material the country can’t fully produce domestically, increasing import dependence rather than reducing it.
That’s a striking reversal for the sole US GOES maker to make. Why make it? Not because amorphous cores are inferior — nobody in this fight is making that claim. It’s because the supply chain for amorphous cores isn’t domestically resilient enough yet to build federal policy around, in Cleveland-Cliffs’ own telling.
Whether that argument wins is genuinely unresolved, and we’re not going to pretend we can call it. What it does confirm, though, is that the amorphous-vs-GOES conversation in the US has stopped being a technical debate; it’s become a supply-chain and trade-policy one, and that changes how buyers should be reading transformer core news in 2026.
Capacity Is Expanding — Just Not Necessarily Where the Mandate Assumed
Cleveland-Cliffs’ supply-chain argument gets a lot more concrete once you look at where amorphous capacity is actually being built in 2026.
The US has one amorphous-metal producer: Metglas, a Proterial subsidiary, running a single plant in Conway, South Carolina. Its annual capacity sits at roughly 45,000 tonnes — not nothing, but a single point of failure for a national mandate.
On February 27, 2026, Proterial announced a new amorphous production site. Not in the US; in Sri City, Andhra Pradesh, India, through a joint venture with Shirdi Sai Electricals. Proterial holds 74% of the venture.
Initial planned capacity is about 30,000 tonnes per year, with roughly $77-80 million committed to phase one and a further $155-160 million earmarked for a second expansion phase. Construction started in January 2026.
Commissioning is targeted for around September 2026. If that timeline holds, India will likely have meaningful new amorphous capacity online before the US comment period even closes; worth sitting with for a second.
None of this means US amorphous supply is shrinking, to be clear. Proterial has stated it’s building a three-site structure across the US, Japan, and India to meet rising global demand. But new capacity is landing where the economics work best right now, and that isn’t automatically the US.
A National-Security Label Gets Attached to Core Steel
A month before Proterial’s India announcement, US policy took a different turn entirely. On April 20, 2026, a presidential determination under Section 303 of the Defense Production Act identified grid infrastructure as facing a critical, dangerous supply shortfall.
The determination specifically named electrical core steel, alongside transformers, high-voltage transmission components, and substations, as essential to national defense. That’s a materials-supply-chain finding, not an efficiency-standard one; a separate policy track from the DOE rule, but pointed at the same underlying shortage.
The stated trigger was blunt: distribution transformer backlogs had stretched past 12 months by early 2026, roughly double the pre-2022 norm, driven by data center construction, EV charging buildout, and broader grid electrification outpacing US manufacturing capacity.
The determination authorizes the Secretary of Energy to use direct purchases, purchase commitments, loans, and loan guarantees to expand domestic production. Does it specify GOES over amorphous, or the reverse? No — it treats the shortage as a capacity problem across core materials generally.
Read together with DOE’s June review, the signal is consistent even if the mechanisms differ. Washington now treats both transformer core materials as strategic supply-chain concerns, not settled commodity inputs; a different posture than either material had even two years ago.
China’s Own Amorphous Push Runs on a Different Track
Does all of this mean amorphous is losing ground globally? Not quite. Outside the US, the push toward amorphous cores hasn’t slowed down at all.
China’s own grid-efficiency framework, through GB national standards and MIIT efficiency classifications, has been steadily tightening no-load loss requirements for new distribution transformer procurement; thresholds that increasingly favor amorphous designs over conventional silicon steel.
In September 2025, China Southern Power Grid connected what was reported as the world’s first 110kV amorphous alloy 3D-wound-core oil-immersed power transformer at the Hongchang Substation in Shantou, Guangdong. A demonstration project, not yet mainstream deployment, but a real technical milestone.
Global amorphous ribbon production is more geographically distributed than most buyers assume, in our experience. China alone is estimated to account for roughly 35-45% of global amorphous metal output, alongside Japan’s established Hitachi/Proterial base; this isn’t a market with one dominant supplier.
For buyers of GOES specifically, the practical read is this: China’s domestic grid is leaning into amorphous for the lightly-loaded distribution segment, the same segment DOE’s rule targets in the US. GOES keeps the position it’s always held, in power transformers and higher-load applications, where amorphous’s lower saturation flux density is a genuine engineering constraint, not a preference.
That segment split isn’t new, and it isn’t going away because of a comment period in Washington. We’ve written separately about why the physics keep amorphous from being a drop-in substitute for GOES — worth reading if you want the engineering case rather than the policy one.
What This Means for Buyers Right Now
So what should buyers actually do with all this? None of what’s happening in 2026 changes a transformer spec that’s already on your desk today. The 2024 rule’s 2029 deadline hasn’t moved, and DOE’s June review is a comment period, not a repeal.
First, don’t treat this as resolved in either direction. Utilities want more time and flexibility. Manufacturers want certainty. The sole domestic GOES producer now wants the mandate softened. Those aren’t aligned positions, and DOE hasn’t signaled which way it leans yet.
Second, watch capacity geography, not just capacity totals. A single US amorphous plant plus a new India facility coming online around September 2026 tells you more about near-term availability and lead times than any market-size forecast does.
Third, keep the segment split in mind when specifying material. Distribution transformers under continuous light load are where amorphous’s efficiency case is strongest. Power transformers and higher-flux applications are still squarely GOES territory; that engineering reality doesn’t move with a comment period.
Fourth, and this is something we’ve noticed directly: buyers have started asking about US transformer policy unprompted in RFQ conversations this year, even ones with no direct US exposure. That’s new, compared to even six months ago.
Trade-policy and capacity risk can move faster than physical supply does. A rule proposal, a comment period, or a determination can shift sourcing conversations within months, while new production capacity still takes years to build regardless of which material it’s making.
Where We Stand on This
We’re not a party to DOE’s rulemaking, and we don’t have inside visibility into how the June 2026 comment period resolves. Rules like this can end up loosened, left alone, or replaced with something neither side proposed; we’d rather say that plainly than guess, and anyone telling you they know for certain is guessing too.
What we do know, from producing CRGO Hi-B and standard-grade coils at production scale, is that GOES demand for power transformer cores hasn’t shown signs of softening because of a US comment period on distribution-transformer policy.
Those are different transformer classes serving different loads; the current material split between them reflects real engineering constraints more than either country’s regulatory mood in a given year.
If your procurement plans depend on how this resolves, the safer move is tracking both tracks — the DOE rulemaking and the Defense Production Act capacity push — rather than assuming either one settles the material question on its own.
FAQ
Did the US ban amorphous steel transformer cores in 2026?
No. If anything, the opposite direction is under discussion — DOE’s June 2026 review is examining whether to soften or delay the 2024 rule that required more amorphous adoption, not to mandate it further.
Is the 2029 compliance deadline for the DOE transformer rule still in effect?
Yes, as of this writing. DOE’s June 2026 comment period is a review process, not a repeal. The 2029 date stands unless and until DOE issues a new final rule changing it.
Why would a domestic steel producer oppose a rule that protects 75% of the market for its own product?
Cleveland-Cliffs’ 2026 position isn’t about the 75% GOES allowance itself — it’s about the remaining 25% amorphous requirement, which it argues pushes manufacturers toward material the US can’t yet fully produce domestically.
Does China favor amorphous or GOES for its own grid?
Both, depending on application. China’s GB standards are pushing amorphous adoption in lightly-loaded distribution transformers, while GOES remains the standard material for power transformers and higher-load applications — the same segment split seen in most markets.
Where is new amorphous metal production capacity actually being built in 2026?
Primarily in India. Proterial’s February 2026 announcement of a Metglas joint venture in Andhra Pradesh adds roughly 30,000 tonnes of planned annual capacity, targeted to commission around September 2026 — alongside the existing 45,000-tonne US plant in South Carolina and established Japanese production.
Should this affect how I specify core material for a new transformer order?
Not directly, unless your application sits in the disputed distribution-transformer efficiency tier. For power transformers and higher-flux designs, GOES remains the practical choice regardless of how the US rulemaking resolves.
If you’re specifying core material for a power transformer or distribution transformer order and want to talk through GOES grade selection for your application, reach out to our export team — happy to walk through specs before you lock a quote.
