Cleveland-Cliffs has committed $200 million to expand grain-oriented electrical steel (GOES) output at its Butler Works in Lyndora, Pennsylvania, a project that is less about adding tonnage than about how that tonnage gets made. The centrepiece is the replacement of two natural-gas-fired slab reheat furnaces with four electric induction furnaces — a change that simultaneously increases GOES volumes by up to 25% and removes combustion emissions from one of the most energy-intensive steps in electrical steel production.

The project sits at the intersection of three forces shaping the steel industry in the mid-2020s: the electrification of industrial heat, the policy-driven reshoring of critical materials for the power grid, and the strategic importance of a product where the United States has exactly one domestic producer.

What Butler Works actually makes

Butler Works is not an ordinary steel mill. It is the leading US producer of high-silicon grain-oriented electrical steel, the material at the core of power and distribution transformers. Cleveland-Cliffs describes itself as the only US producer of regular grain-oriented (RGO) steel, and the facility also makes TRAN-COR high-permeability grades alongside stainless and carbon steels. Total raw-steel capacity at the site is roughly 0.4 million net tons per year.

The company does not publish a GOES-specific annual tonnage for Butler. What is known is more indicative of market position than of absolute scale: a 2020 submission from WEG Transformers USA to the US Bureau of Industry and Security estimated that AK Steel — then the owner of the Butler asset, now part of Cleveland-Cliffs — supplied about 225,000 tons of a roughly 300,000-ton US GOES market. That is approximately 73% domestic market share, on a five-year-old estimate. The figure should be treated as an industry estimate rather than a current audited number, but the direction is clear: for transformer manufacturers in the United States, Butler Works is effectively the domestic supply.

That is why the capacity question matters more than the dollar figure. A 25% increase in GOES output at a facility holding the majority of US supply is a structural change in the domestic market for a material that underpins grid investment.

One nuance deserves care. Cleveland-Cliffs has, in earlier filings, described the broader electrical steel expansion as capable of increasing American-made GOES tonnage by approximately 30% to 40% by using under-utilised capacity. The October 2026 announcement states up to 25%. These are not necessarily contradictory — one may refer to a theoretical capacity ceiling, the other to expected output — but readers should not treat the 25% figure as a precisely defined tonnage commitment. It is an "up to" number, and the company has not disclosed the baseline to which it applies.

The technical change: electrifying the reheat step

The furnace replacement is the substantive engineering story. In a hot mill, steel slabs must be reheated to rolling temperature before they enter the rolling train. At Butler, that reheating has been done with natural-gas-fired furnaces. The project replaces two of them with four electric induction reheat furnaces, complemented by new electrical infrastructure — local reporting on Butler Township land-development review mentions a hot-mill expansion, auxiliary buildings and three new electrical substations.

Induction heating works by inducing eddy currents directly within the steel, rather than transferring heat from a combustion flame through the slab surface. In practice this offers tighter temperature control, faster heating, lower surface oxidation losses and better temperature uniformity through the slab thickness — qualities that matter disproportionately in GOES, where metallurgical consistency directly affects magnetic properties such as core loss and permeability. A transformer core made from steel with inconsistent texture performs worse and wastes more energy over its multi-decade service life.

That is the commercial logic underneath the sustainability framing: this is as much a quality and yield project as a decarbonisation project. Cleveland-Cliffs has said the furnaces will improve slab heating efficiency, reduce production losses and cut emissions from hot rolling. The company has also estimated the installations could result in up to a 100% reduction in direct greenhouse-gas emissions from the reheating step — a claim that is plausible in the narrow sense that eliminating combustion eliminates direct CO2 from that process, provided the electricity supply is accounted for separately. Scope 2 emissions and grid carbon intensity are a different ledger.

Funding: a DOE award, delivered in phases

This is not a purely commercial capital project. It originates in the Department of Energy's Industrial Demonstrations Program, administered by the Office of Clean Energy Demonstrations. In March 2024, Cleveland-Cliffs was selected for up to $575 million across two projects, with the Butler Works induction reheat furnace project eligible for up to $75 million of federal cost share.

By August 2024, DOE had finalised a Phase 1 award releasing more than $19 million to begin planning, permitting, design and development. The published Phase 1 figures are specific: a total project amount of $38,151,901 for the phase, with an OCED award amount of $19,074,900 — approximately a 50/50 split for that phase, over an eight-month period starting August 2024. A DOE project fact sheet describes the four-for-two furnace swap and anticipates more than 200 union building-trades construction jobs.

The DOE structure explains an apparent discrepancy in the numbers that readers will encounter. The $200 million is Cleveland-Cliffs' own investment; the $75 million is a maximum federal contribution. Local coverage of the project has cited a "$195 million expansion plan" — a figure that likely reflects a specific scope or an earlier planning-stage estimate rather than a conflict with the $200 million headline. The precise boundary between company-funded and DOE-supported scope is not publicly detailed.

Is there an EPC contractor?

No external engineering, procurement and construction contractor has been identified in any public material reviewed for this article, and that absence is itself informative. This project is being executed as an in-house capital project at an operating plant by a self-performing steelmaker, rather than as a greenfield development with a named EPC consortium. The procurement structure is therefore likely to be equipment-and-package-based — furnaces, electrical infrastructure, substations, civil works — contracted directly by Cleveland-Cliffs and its in-house engineering organisation, with union building trades supplying construction labour.

On the furnace equipment specifically, there is a strong but unconfirmed inference worth flagging as such. Ajax TOCCO Magnethermic is an induction heating and melting equipment manufacturer headquartered in Warren, Ohio. Cleveland-Cliffs' October 2026 press release states that the new induction reheat furnaces will be manufactured in Warren, Ohio. And the company's 2024 sustainability report carries a photograph of an induction furnace similar to the planned Butler equipment with an "Ajax TOCCO" photo credit. The geographic match, the technology match and the photo credit all point in the same direction — but no public project document explicitly names Ajax TOCCO as the Butler supplier. Treat it as a well-supported inference, not a confirmed contract award.

The substations are another open item. Three new electrical substations imply a substantial electrical balance-of-plant scope and significant interconnection requirements, but no supplier or utility partner has been publicly identified.

Timeline: two dates, and a real inconsistency

Here the public record diverges in a way that matters for anyone tracking the project.

SourceCompletion date stated
Cleveland-Cliffs press release (October 2026)Early 2028
Cleveland-Cliffs project page (DOE Butler Works)2029
Cleveland-Cliffs 2026 SEC filing2029
Butler Eagle local coverage (April 2026)Operational target July/August 2028

The most coherent reading is that these refer to different things: the "early 2028" target may cover the core furnace installation and GOES production increase, while the 2029 date in the sustainability project page and SEC filing may cover full project completion including all supporting infrastructure. That interpretation is my own analysis, not a company statement — Cleveland-Cliffs has not reconciled the two dates publicly. The practical consequence is that anyone modelling GOES availability from Butler should use a 2028–2029 window rather than a single date.

The chronology also shows this has been a long-running project rather than a sudden announcement:

  • March 2024 — DOE selects Cleveland-Cliffs for up to $575 million across two projects; Butler eligible for up to $75 million.
  • August 2024 — Phase 1 award finalised (>$19 million released); planning, permitting and design begin.
  • June 2025 — Pennsylvania DEP issues a Title V/state operating permit for Butler Works. This is a facility-wide air permit framework, not a project-specific construction permit, and it does not constitute NEPA or construction approval for the expansion.
  • April 2026 — Local reporting on Butler Township land-development review describes the expansion as moving forward, including hot-mill scope and three new substations.
  • October 2026 — Cleveland-Cliffs hosts Energy Secretary Chris Wright and other federal officials at Butler Works to showcase the project, restating the $200 million investment, the up-to-$75 million DOE award and the early-2028 target.

The October 2026 event is a political and communications milestone rather than a construction milestone. It signals that the project retains federal backing and remains a flagship for the administration's domestic energy and manufacturing agenda — but it is not evidence that procurement, contracting or construction is complete.

Why the electricity question is the interesting one

Converting a gas-fired reheat furnace to induction is a genuine process change, and it shifts the emissions and cost profile rather than eliminating the energy problem. A hot mill reheating slabs requires very large quantities of thermal energy. Moving that load from combustion to electricity means taking on a substantial new electrical demand — which is consistent with the three new substations, and which raises questions that the public record does not answer: what the interconnection capacity and timeline look like, what power contract or tariff applies, and how the operating cost compares with gas at prevailing prices.

The project also creates a neat circularity that is worth noting without overreading: Cleveland-Cliffs is expanding production of the steel used in transformers and grid equipment, and it is doing so by electrifying its own process. Grid investment increases demand for GOES; electrifying industrial heat increases demand for grid capacity. The company is on both sides of that loop.

For steel-industry observers, the more transferable lesson is about the reheat furnace itself. Slab reheating is one of the larger direct-emissions sources at an integrated or electric-arc hot-mill site, and it is exactly the kind of high-temperature duty that is difficult to decarbonise with the usual toolkit. If induction reheating works at Butler at GOES scale and quality requirements — and GOES is arguably the most quality-sensitive product on the mill — it becomes a reference case for other hot mills, both within Cleveland-Cliffs and across the industry. That is a meaningfully different proposition from a pilot or demonstration line.

What is still unknown

The gaps are substantive enough to state plainly:

  • Absolute capacity. No GOES-specific tonnage baseline or increment has been disclosed. The 25% figure cannot be converted to tons from public information.
  • Contractors and suppliers. No EPC, general contractor, furnace supplier, substation supplier or electrical contractor has been publicly confirmed. The Ajax TOCCO link remains a strong inference.
  • Procurement status. Whether major equipment packages have been awarded, are out for bid, or are being executed in-house is not public.
  • Scope boundary. The split between Cleveland-Cliffs' $200 million and the DOE's up-to-$75 million, and the scope each covers, is not detailed.
  • Reconciled schedule. The early-2028 versus 2029 discrepancy has not been explained by the company.
  • Power arrangements. Interconnection, supply and pricing for the new electric load are undisclosed.

The clearest conclusion is that this project has moved past the announcement stage. Phase 1 DOE funding was released in 2024, planning and permitting activity is visible, local land-development review has occurred, and the company is now hosting federal officials to showcase it. But it remains a project whose commercial details sit almost entirely inside Cleveland-Cliffs, and whose significance to the North American GOES market is easier to argue than to quantify from public sources.

DOE Office of Clean Energy Demonstrations Award Wednesdays graphic
The Butler Works project received Phase 1 funding through DOE's Office of Clean Energy Demonstrations in August 2024.Credit: Award Wednesdays | August 21, 2024 | Department of Energy · Public-domain / reusable source (verify licence) · Illustrative, not a photograph of this project · source