Type “transformer manufacturers in USA” into a search box and you get nine lists of companies. Almost none of them tell you the one thing that decides whether a supplier can actually serve your project: where the iron is made. A transformer can be assembled in Texas from a core cut in Asia, windings wound in Mexico, a tank built in India, and still carry a nameplate that reads like a domestic product. For a plain commercial job that may not matter at all. For a job with federal money in it, it is the difference between a compliant purchase and a False Claims Act problem.
So this guide does three things the other pages do not. It separates who owns the company from where the factory sits, because those are different questions with different answers. It puts real numbers on the four compliance lines that actually decide your supplier list in 2026: Build America Buy America, the Buy American Act, Section 232 tariffs, and the DOE efficiency clock running to 23 April 2029. And it gives you the arithmetic, per kVA and per week, so you can tell whether the quote in front of you is a bargain or a trap.
If you are choosing between a US factory and an overseas one on a specific machine, our US versus China pad-mount sourcing comparison goes through that decision unit by unit. This page is the wider map: who is in the American market, what each tier can build, and what the rules are.
Two numbers to hold while you read. Domestic mills supply only a fraction of American demand for grain-oriented electrical steel, and the National Electrical Manufacturers Association told DOE in writing that the United States has a single domestic producer of it. And a 1,000 kVA pad-mount quoted at $10,000 from a non-China origin does not land at $10,000. Under the Section 232 framework that took effect on 6 April 2026, it lands closer to $12,500 before a single freight or brokerage line is added.
“Made in USA” is three different claims
Before the list of names, the vocabulary. Suppliers use these interchangeably and buyers hear them as the same sentence. They are not.
Made in USA. The finished unit is manufactured in the United States. For a transformer, the practical test is where the core is cut, where the coils are wound, and where the tank is welded and the unit is filled, dried and tested.
Assembled in USA. Major subassemblies arrive from elsewhere and are put together domestically. Final test may well be done here. Core origin usually is not.
Designed in USA, engineered to American standards, compliant with US regulations. None of these is a claim about manufacturing location. The first is about engineering hours. The second means the unit is built to meet IEEE and ANSI requirements. The third means it can legally be sold here. All three are true of plenty of imported equipment.
The reason this matters is not patriotism. It is that three separate rulebooks reward or punish you on the basis of it, and each one draws the line in a different place.
| Rule | What it asks | The number |
|---|---|---|
| Build America, Buy America (2 CFR Part 184) | Manufactured product made in the US, with US-origin component cost above a threshold | 55% of total component cost |
| Buy American Act (FAR 52.225-1) | End product manufactured in the US with domestic component cost above a threshold | 65% for deliveries 2024 to 2028, 75% from 2029 |
| Section 232 metals tariffs | Where the steel, aluminium and copper were melted and poured, not where the unit was assembled | 85% US-origin metal content earns the reduced 10% rate |
| Trade Agreements Act | Country of origin by substantial transformation, above the dollar threshold | China, Russia and India are not designated countries |
Read the second and third rows together, because they are the ones that catch people. Buy American counts component cost. Section 232 counts metal origin. A transformer can fail one and pass the other.
One more trap worth naming up front. For domestic content purposes, “domestic” means the United States. Canada and Mexico do not count, USMCA notwithstanding. A Canadian-built unit from a well regarded brand is a foreign end product under Buy American and does not help you on a domestic content bonus. This trips up a surprising number of projects, because the brand is familiar and the freight is short.
Who is actually American, and who merely builds here
This is the table the other lists do not build. It separates the headquarters country from the factory country, because that split is what determines your answer under each of the four rules above.
US-headquartered, with US factories
| Company | Where | What it is known for |
|---|---|---|
| Virginia Transformer Corp | Roanoke and Troutville VA, Pocatello ID, Rincon GA, plus two plants in Chihuahua, Mexico | Describes itself as the largest US-owned power transformer manufacturer in North America. Six plants. Roanoke builds to 30 MVA base at 161 kV and 650 kV BIL, dry type to 10 MVA at 35 kV. Pocatello reaches 230 kV at 1050 kV BIL. Rincon was built for shell units to 1,400 MVA and 500 kV, with 47,000 MVA of stated plant capacity |
| Pennsylvania Transformer Technology | Canonsburg PA (a one million square foot, 18 building campus) and Raeford NC | Founded 1929, now a Quanta Services company. Rated 500 kVA to 750 MVA through 765 kV. Both US plants |
| ERMCO | Dyersburg TN | Employee-owned, oil-filled distribution transformers, a large share of utility pad-mount and pole-mount volume |
| Howard Industries | Laurel MS | Distribution and small power, strong in utility frame sizes |
| Central Moloney | US plants | Distribution transformers and transformer components |
| Southwest Electric Co | US | Liquid-filled from 25 kVA to 7,500 kVA, standard and custom |
| Olsun (American Transformer Manufacturers) | Illinois, founded 1950 | Dry type, marine duty, cast coil and encapsulated designs |
| Federal Pacific | Bristol VA | Liquid-filled distribution |
| MGM Transformer | Los Angeles CA, since 1952 | Dry type and liquid-filled |
| Power Partners / UTB | Athens GA | Single and three phase pad-mount for underground distribution; the company cites over nine million units produced |
| Acme Transformer | Bland VA | Distribution and control transformers |
| GE Vernova | US | Grid and power transformers, the historic American line |
Foreign-headquartered, with substantial US manufacturing
| Company | Parent country | US footprint |
|---|---|---|
| Hitachi Energy USA | Japan | Raleigh NC, grid and power transformers |
| Siemens Energy | Germany | Orlando FL, power and distribution |
| ABB Inc | Switzerland and Sweden | Cary NC, full range |
| Schneider Electric / Square D | France | Palatine IL and other sites |
| Eaton | Ireland-domiciled, run from Beachwood OH | Very large US electrical manufacturing base |
| WEG Transformers USA | Brazil | Washington MO, plus two Mexican plants; distribution to 10 MVA, power to 350 MVA and 161 kV per company literature |
| Hammond Power Solutions Americas | Canada | Describes itself as the largest dry-type manufacturer in North America, with US plants |
| Prolec GE | Mexico (Apodaca, Nuevo León), a joint venture | Rated to 1,200 MVA and 345 kV. GE Vernova has moved to take full control of the venture |
None of this makes the second group a worse choice. Hitachi Energy and Siemens Energy build excellent machines and employ thousands of Americans. But if your project carries a domestic content requirement, the badge on the truck is not the thing that gets audited. The bill of materials is.
Two more categories that belong on the list, because for a lot of buyers they are the actual answer:
Remanufacturers and surplus houses. Maddox, Sunbelt Transformer, Emerald Transformer and others hold inventory and rebuild units. A fully remanufactured distribution transformer, stripped, rewound where needed, retanked, refilled and tested to ANSI, typically lands at 50% to 70% of new cost in four to six weeks instead of 30 to 50. For storm response or a schedule critical expansion, this is often the only route that holds the date.
Licensed manufacturing. Some brands license designs to overseas factories. A familiar name on the cut sheet tells you nothing about which plant will build your order. Ask about the specific product line, not the brand.
For the global view that ignores nationality entirely, our power transformer manufacturers and brand overview covers the tier structure worldwide.
Who builds what: the capacity and voltage ladder
Not every American factory can build every machine, and knowing which band your job falls in saves a wasted round of enquiries.
| Band | Typical US builders | Notes |
|---|---|---|
| Dry type, 15 kVA to 5 MVA, 600 V to 35 kV | Hammond Power Solutions, Olsun, MGM, Acme, Federal Pacific, Virginia Transformer (Chihuahua) | The shortest lead times in the market. Cast resin and VPI |
| Distribution, 10 kVA to 5,000 kVA, up to 34.5 kV | ERMCO, Howard Industries, Central Moloney, Power Partners, Federal Pacific, WEG, Southwest Electric | Pad-mount, pole-mount, single and three phase. The volume heart of the market |
| Small power, 5 MVA to 30 MVA | Virginia Transformer, Pennsylvania Transformer, Southwest Electric, Hitachi Energy, ABB | Where the queue starts to bite |
| Medium power, 30 MVA to 100 MVA | Virginia Transformer (Rincon), Pennsylvania Transformer, Hitachi Energy, Siemens Energy, ABB, WEG | 60 to 90 week territory |
| Large power, above 100 MVA | Rincon (Virginia Transformer and Georgia Transformer), Pennsylvania Transformer, Hitachi Energy, Siemens Energy, ABB, Prolec GE | 110 weeks and up. One published estimate puts the domestically built share of units above 100 MVA at 18% as of 2019, a figure that has improved since but has not closed |
| Generator step-up and HVDC | A short list, mostly the global OEMs | The tightest segment in the world right now |
If you are sourcing anywhere in the distribution band, our distribution transformer manufacturers page lists the suppliers by that specific layer rather than by country.
The product range we build across these bands is set out in the product index, and completed work by application lives in the project list.
Lead times in 2026, by machine class
The lead time crisis is in its sixth year. It has eased for small units and stayed brutal for large ones. Here is where things actually sit.
| Machine class | 2026 lead time | In months | Pre-2020 norm |
|---|---|---|---|
| Dry type, under 1 MVA | 8 to 16 weeks | 1.8 to 3.7 | 4 to 8 weeks |
| Dry type, 1 to 5 MVA | 12 to 24 weeks | 2.8 to 5.5 | 8 to 12 weeks |
| Dry type, above 5 MVA | 20 to 32 weeks | 4.6 to 7.4 | 12 to 20 weeks |
| Distribution pad-mount and pole, under 2.5 MVA | 30 to 50 weeks | 6.9 to 11.5 | 8 to 12 weeks |
| Substation, 5 to 50 MVA | 75 to 110 weeks | 17.3 to 25.3 | 20 to 30 weeks |
| Mid-size power, 30 to 100 MVA | 60 to 90 weeks | 13.8 to 20.7 | about 30 weeks |
| Large power, above 100 MVA | 110 to 130 weeks, many quotes past 160 | 25.3 to 29.9 | 40 to 52 weeks |
| Generator step-up, average | about 144 weeks | 33.1 | 40 to 52 weeks |
| Tier 1 global OEMs | 48 to 60 months | 48 to 60 | not comparable |
There is genuine good news buried in there, and it comes from a source that has every reason to be accurate. In a comment letter filed with the Department of Energy, NEMA stated that distribution transformer lead times fell from over 100 weeks in 2023 to roughly 30 weeks in 2025 and 2026, a reduction of about 70%. That is the industry responding to a stable rule. It is also why NEMA has asked DOE not to reopen the standard.
Two practical consequences follow from that table.
The first is that custom designs, unusual voltages and extra testing add four to twelve weeks on top of everything above. Every engineering change after you have a slot is a change to your place in the queue.
The second is that the queue is tiered by supplier, not just by machine. A 110 to 220 kV unit from an export-oriented factory may quote 18 to 36 months while a Tier 1 OEM quotes 48 to 60 and is closed for 2028 delivery entirely. If your energization date is fixed, the tier you can access is a bigger variable than the machine you need.
Why nobody can go faster: one steel mill
Every lead time conversation ends at the same place. Grain-oriented electrical steel, GOES, is the silicon-iron alloy that makes a transformer core work, and it is the binding constraint on the whole North American supply chain.
NEMA put it plainly in its DOE filing: the United States currently has a single domestic producer of grain-oriented electrical steel, creating a structural supply constraint. Published estimates of what that mill covers range from about 12% to roughly one fifth of US demand. Either way the answer is “not most of it,” and the same steel has to serve distribution transformers, large power transformers, generator step-up units and the data centre buildout at the same time.
Everything else in the shortage story compounds on that base:
- Demand is up sharply and unevenly. Power transformer demand is reported up 116% since 2019, distribution transformer demand up 41%, and generator step-up demand up 274%.
- Price has followed. Power transformer unit prices are reported up 77% since 2019, GSU up 45%, and distribution classes up somewhere between 45% and 95% depending on the class.
- Wood Mackenzie models a supply deficit of roughly 30% for power transformers and 6% to 10% for distribution units.
- Over half of North America’s in-service distribution transformers are past their 33 year expected service life, so the replacement wave is still building.
Re-tooling a steel mill takes years and capital. Building a large transformer plant takes three to five years and a workforce that has to be trained. That is why the queue is structural rather than cyclical.
The four compliance lines that decide your supplier list
If your project has no federal money and no tax credit in it, skip to the next section. If it does, these four rules will do more to shape your bid list than any technical requirement.
Build America, Buy America
BABA applies to federal financial assistance for infrastructure: grants, loans and cooperative agreements, not just direct federal contracts. It covers transformers explicitly as manufactured products. The thresholds are:
| Category | Requirement |
|---|---|
| Manufactured products (transformers) | Made in the US, with US-origin component cost above 55% of total component cost |
| Iron and steel | Every manufacturing process from initial melting through coating must occur in the US. No percentage test |
| Construction materials | All manufacturing processes in the US |
Three things surprise buyers here.
There is no BABA certificate and no federal registry. Compliance is documented through a manufacturer self-certification letter, and it must be specific to your project and your order. A generic “we are BABA compliant” letter is not sufficient. The letter should name the project, describe the product, state the city and state of manufacture, and carry an authorised signature on company letterhead.
The 55% test is per bill of materials, not per company. A manufacturer can use European bushings, an Asian tap changer or Brazilian specialty steel on one model and not another. The same brand can be compliant on one line and not on the next.
Waivers exist but they cost time. Non-availability, unreasonable cost (where domestic sourcing raises total project cost by more than 25%), and public interest. All require public comment of at least 15 days, and DOE may take up to 90 calendar days to process. Plan for weeks to months.
Buy American Act
This one covers direct federal procurement, and it is a different regime with a different number. The domestic component cost threshold is 65% for deliveries in 2024 through 2028, rising to 75% for deliveries from 2029 onward. For multi-year contracts, the threshold in effect at the year of delivery governs, not the year of award. Foreign offers carry a price evaluation penalty of 20% for large businesses and 30% for small businesses.
Trade Agreements Act
Above the applicable dollar threshold, the TAA replaces the country-of-origin analysis. Products must be made in, or substantially transformed in, the US or a designated country. China is not a designated country. Mere assembly of imported components does not create a new article of commerce.
FEOC
Foreign Entity of Concern rules began with battery components and have broadened. In 2026, IRS guidance clarified the material assistance rules and carved main power transformers out of the Material Assistance Cost Ratio calculation. That carve-out applies to main power transformers only. Distribution transformers and other components stay in scope.
Across all four, the penalty structure is not a rounding error. False certifications carry treble damages and civil penalties in the range of $14,308 to $28,619 per false claim, and qui tam relators, who can be competitors or employees, are entitled to 15% to 30% of any recovery.
If you are moving equipment between the IEC and ANSI worlds, our IEC 60076 versus ANSI and IEEE comparison covers where the two standard families diverge.
Section 232: how the tariff actually lands on your quote
Here is the piece every other page skips, and it is worth real money.
On 6 April 2026 a proclamation restructured Section 232 tariffs on steel, aluminium and copper. The change that matters most: duties are now assessed on the full customs value of the article, not on the declared value of the metal inside it. Under the old system you could reduce exposure by reporting a low metal content. That door is closed.
The tier structure, as reported:
| Tier | Rate | Applies to |
|---|---|---|
| Steel and aluminium mill products | 50% | Primary metals, HTS 7206 to 7229 and 7301 to 7306 |
| Derivative articles, standard | 25% | Including HTS 8501 to 8504, which covers motors, generators and transformers, where metal is at least 15% of weight |
| Large transformers above 10,000 kVA (8504.23) | temporarily reduced | Runs through 31 December 2027, then steps back up to 25% |
| High US-content articles | 10% | Where at least 85% of the applicable metal content was melted, smelted or cast in the US |
| Low metal content | exempt | Where the applicable metal is 15% or less of total weight |
What that does to a quote. Take a 1,000 kVA liquid-filled unit at a quoted $10,000:
| Layer | Amount | Running total |
|---|---|---|
| Quoted price, ex works | $10,000 | $10,000 |
| Section 232 derivative at 25% | $2,500 | $12,500 |
| Section 301 where Chinese origin applies, at 25% | $2,500 | $15,000 |
Those are duty lines only. Freight, insurance, MPF and HMF come on top. A published tariff engine worked example for HTS 8504.22.00.40, a liquid dielectric transformer between 650 and 2,500 kVA, from a non-China origin, shows exactly this: general rate of duty Free, Section 232 derivative 25%, $2,500 on a $10,000 customs value, and about $47 in CBP fees on an ocean entry.
Three cautions before you put any of this in a budget.
The stack is specific. Which layers apply depends on the HTS line, the country of origin, whether metal content clears 15% by weight, and which annex or trade agreement rate produces the better outcome. Published all-in rates for HTS 8504 from Chinese origin span a very wide band. Do not treat one number as your rate. Get the classification and the origin determination in writing from your broker before you sign.
The 10% rate is a documentation claim, not a marketing claim. It depends on where the metal was melted or smelted, verified by mill certificates, not on where the unit was assembled. A product built entirely in the United States from imported steel does not qualify. A product built overseas from qualifying US-melted steel can.
The temporary reduction has a sunset. Units ordered now with 30-plus week lead times can clear customs near the end of 2027. Write a duty clause into the purchase order that says who owns a rate change between order and entry, and state whether the quoted price includes current Section 232 duties.
Verifying a domestic content claim without getting burned
The questions that actually work are specific ones, and good suppliers answer them every day.
- Where is the core manufactured, and where is the steel for it melted? This is usually the component that decides domestic content status.
- Where are the windings produced?
- What percentage of component cost is US-origin, for this model?
- Which plant, city and state will build my specific order?
- For FEOC-sensitive work, who is the ultimate beneficial owner of the manufacturing entity? A European parent can sit above a subsidiary with disqualifying ownership.
- Can you provide a country of origin certificate with a component breakdown, and a signed manufacturing attestation?
Four scenarios show why this is worth ten minutes:
The specification gap. A developer ordered from a brand with a well known US presence, assuming domestic manufacture. The specific product line they ordered was built at an overseas plant. They found out during compliance documentation and had to re-source to keep tax credit eligibility.
The assembly question. A solar project specified domestic content. The supplier quoted from its US facility, which did final assembly, while cores came from overseas. The units did not meet substantial transformation.
The ownership chain. A buyer assumed a European-branded supplier was clear on FEOC. A subsidiary in the chain had disqualifying ownership.
North American is not domestic. A wind developer sourced from a Canadian manufacturer and had to redo the domestic content calculation, because Canada is not domestic for this purpose.
None of these require suspicion. They require documentation, and documentation is far cheaper at quotation stage than reconstructed during an audit.
While you are collecting it, ask for the factory test report with measured no-load and load losses, not the design values. Our guide to transformer routine testing sets out what a unit should pass before it ships.
The 2029 efficiency clock, and your last safe order date
The Department of Energy published its final rule on distribution transformer efficiency standards on 22 April 2024 (89 FR 29834). The rule took effect 8 July 2024. Compliance with the amended standards is required on and after 23 April 2029.
Two details decide how this lands on you.
The date attaches to date of manufacture, not to your purchase order and not to delivery. A unit built on or after 23 April 2029 must meet the new floor. A unit built before it does not, and can be installed later.
The scope is specific. Per 10 CFR 431.192, the rule covers 60 Hz transformers with primary at or below 34.5 kV and secondary at or below 600 V: liquid-immersed from 10 to 5,000 kVA, dry-type from 15 to 5,000 kVA. Generator step-up units, anything above 5,000 kVA and anything above 34.5 kV sit outside this rule, so the clock below does not apply to them.
Working back from 23 April 2029, here is the last date you can place an order and still expect manufacture to complete in time, using today’s lead times:
| Machine class | Lead time | Last safe order date | Margin from today |
|---|---|---|---|
| Distribution pad-mount, 30 weeks | 30 weeks | 25 September 2028 | 713 days |
| Distribution pad-mount, 50 weeks | 50 weeks | 8 May 2028 | 573 days |
| Substation 5 to 50 MVA, 75 weeks | 75 weeks | 15 November 2027 | 398 days |
| Substation 5 to 50 MVA, 110 weeks | 110 weeks | 15 March 2027 | 153 days |
That last row is the one to look at. A substation class unit at the long end of the quoted band gives you roughly five months of slack from today. If your project has any scope uncertainty, that is not much.
What the standard actually changes, in losses, using the published current versus 2029 efficiency pairs at each class’s representative loading point:
| Class and rating | Current | 2029 | Allowed loss goes from | Annual saving | 30 year present value |
|---|---|---|---|---|---|
| Liquid, three phase, 500 kVA | 99.35% | 99.38% | 1,636 W to 1,560 W | 665 kWh | $826 |
| Liquid, three phase, 1,000 kVA | 99.43% | 99.46% | 2,866 W to 2,715 W | 1,329 kWh | $1,649 |
| Liquid, single phase, 500 kVA | 99.49% | 99.59% | 1,282 W to 1,029 W | 2,210 kWh | $2,743 |
| Low voltage dry type, three phase, 150 kVA | 98.83% | 99.06% | 622 W to 498 W | 1,080 kWh | $1,341 |
| Low voltage dry type, three phase, 500 kVA | 99.14% | 99.31% | 1,518 W to 1,216 W | 2,647 kWh | $3,285 |
The loss and present value columns are our arithmetic on the published efficiency pairs, at $0.10 per kWh over 8,760 hours with a 30 year, 7% present value factor of 12.409. The efficiency pairs themselves are the published figures.
The pattern is worth noticing: the rule leans hardest on dry type and on smaller single phase units, and goes comparatively lightly on the big three phase liquid units that utilities are struggling to source. DOE’s own analysis concluded the standard could be met with roughly 75% of distribution transformer electrical steel remaining grain-oriented rather than forcing a wholesale shift to amorphous. That was deliberate, and it is why the industry has not pushed back harder.
One more regulatory note. On 20 April 2026 a presidential determination found that grid infrastructure supply chains, including distribution transformers and electrical core steel, are essential to national defense. DOE followed with a request for information on 15 June 2026. If you are planning 2029 compliance now, watch that docket, because it is where the next change will appear first.
What things cost
Price per kVA is not linear, and the shape of the curve matters more than the level.
| Rating | Indicative price per kVA | Indicative unit price |
|---|---|---|
| 100 kVA | $22.0 | $2,200 |
| 200 kVA | $17.0 | $3,400 |
| 315 kVA | $13.7 | $4,316 |
| 500 kVA | $11.8 | $5,900 |
| 800 kVA | $10.9 | $8,720 |
| 1,000 kVA | $10.0 | $10,000 |
| 1,600 kVA | $7.2 | $11,520 |
Per-kVA cost falls by roughly 3.1 times between 100 kVA and 1,600 kVA, then flattens. The practical consequence is that moving up one kVA tier is often cheaper per kVA than negotiating the same unit down, and right-sizing against the real load profile routinely saves more than vendor haggling.
A caution on this table and every other price table you will find. It is indicative, from a secondary market source, and it carries no date, no Incoterm, no winding material and no efficiency tier. We publish it because the shape is useful and repeatable across sources, not because the level is quotable. No unsourced price range should ever go into a budget. What you can do with it is sanity-check the spread between two quotes you already hold.
Where the spread between two quotes on the same kVA comes from, in rough order of impact:
- Loss level and core grade. Amorphous and high-grade GOES against standard GOES. Amorphous and S20 designs are reported to add 8% to 15% to base pricing while cutting no-load loss sharply.
- Winding material. Copper against aluminium is one of the largest single line items.
- Impedance. Typically 4% to 6% in the distribution range, and it changes the copper and core.
- Tap changer type, dielectric fluid, enclosure rating.
- Test scope. Routine, type and special tests are not the same product. A stripped quote with a nameplate that will not survive a utility audit is not the same machine as a fully tested unit built for a 65 °C average winding rise at 40 °C ambient, even at the same kVA.
On total cost of ownership, dry type and cast resin cost 20% to 40% more up front and often win over 20 years, with reported savings of $15,000 to $30,000 per unit in energy and maintenance against a low-bid oil-immersed unit. No-load loss dominates, because a transformer is energized around the clock. A 30% lower no-load loss is reported to pay back the efficiency premium in five to eight years at typical industrial load factors.
Where buying outside the US still makes sense
It would be dishonest to pretend domestic is always the answer. Three situations favour an overseas factory, and being clear about them is more useful than a flag-waving conclusion.
Schedule. When a Tier 1 OEM quotes 48 to 60 months and the same class from an export-oriented factory quotes 18 to 36 months, and your energization date is fixed, the tier you can access decides the project.
Federal money is not involved. If no BABA, no BAA, no tax credit and no FEOC sensitivity applies, the compliance premium buys you nothing. Buy on delivered cost, lead time and test scope.
Standard catalogue units in volume. Where the design is settled and repeated, the per-unit economics of a high-volume factory are hard to argue with. Just be sure the comparison is apples to apples on losses, impedance, test scope and winding material before you believe the gap.
Those three cases, and the sourcing models that fit them, are worked through in more depth across our resource library. The short version is that the right answer moves with the project, not with the flag.
Where domestic wins, and it often does, the reasons are concrete rather than sentimental. No ocean freight or port exposure. No tariff reclassification risk. Warranty service a flight away rather than an ocean away. Simpler compliance if rules tighten later, which they have every year since 2024. And delivery certainty, which at today’s lead times is worth more than the price delta to most project schedules.
The honest middle is what a lot of serious buyers now run: dual source. Critical, long-lead, federally funded units go domestic. Standard catalogue units and bulk replacements go to qualified export factories. The mix shifts with the project, not with the flag.
The RFQ package
Send this, and you will get quotes you can actually compare. Every line exists because its absence causes a specific, recurring problem.
1. Rating: ______ kVA / MVA, ______ phase, ______ Hz
2. Primary: ______ kV, connection ______, BIL ______ kV
3. Secondary: ______ kV, connection ______, BIL ______ kV
4. Vector group: ______ Neutral: ______
5. Cooling class: ______ Temperature rise: ______ K
6. Impedance: ______% at ______ MVA base, tolerance ______%
7. No-load loss: ______ W (guaranteed, not design)
8. Load loss: ______ W at ______ deg C, at ______% load
9. Insulation: liquid / dry / cast resin / ester. Fluid: ______
10. Standards: IEEE C57.12.____ , NEMA, UL 1561 / 1562 / CSA C88
11. Efficiency compliance: DOE 10 CFR 431.196, 2029 edition YES / NO
12. Tests required: routine / type / special. List: ______
13. Factory test report with MEASURED losses: required YES / NO
14. Country of manufacture (city, state, country): ______
15. Country of melt and pour for core steel: ______
16. Domestic component cost %: ______ Self-certification letter: YES / NO
17. HTS classification proposed: ______
18. Section 232 treatment assumed in this quote: ______
19. Incoterm: ______ Duty clause: ______
20. Lead time: ______ weeks from approved drawings, not from PO
21. Validity of price: ______ days
22. Warranty: ______ monthsThree lines carry more weight than the rest. Line 7 and 8 ask for guaranteed losses, not design values, which is the only way a loss evaluation means anything. Line 15 asks where the steel was melted, which is the question that decides domestic content and tariff treatment. Line 20 pins lead time to approved drawings, because otherwise you are measuring from a date you do not control.
Bottom line
The nameplate is not the answer. Roughly half the companies on a typical “American transformer manufacturers” list are foreign-headquartered with US plants, and some are not even in the same product category. If your project carries a domestic content requirement, ask where the core is cut and where the steel was melted, and get it in a signed letter that names your project.
The compliance rules now carry numbers, and they move. BABA sits at 55% of component cost. Buy American sits at 65% until 2028 and 75% from 2029. Section 232 since 6 April 2026 is assessed on full customs value, at 25% for most transformer classifications, with a temporary reduction for units above 10,000 kVA through the end of 2027. False certifications carry treble damages and penalties per claim.
The clock is the thing you can still act on. DOE compliance is required on and after 23 April 2029, and it attaches to the date of manufacture. At 110 weeks of lead time, the last comfortable order date for a substation class unit is around March 2027. That is closer than it looks.
And the queue is not going back to 2019. One domestic GOES producer, demand up 116% for power transformers and 274% for generator step-up units, and a modelled 30% supply deficit. Plan on 18 month horizons for large units, lock the specification before you ask for a slot, and treat a second qualified source as insurance rather than disloyalty.
FAQ
Who is the largest transformer manufacturer in the United States? It depends on whether you mean ownership or output. Virginia Transformer describes itself as the largest US-owned power transformer manufacturer in North America, with six plants. Hammond Power Solutions describes itself as the largest dry-type manufacturer in North America, and is Canadian-headquartered. The global OEMs, Hitachi Energy, Siemens Energy, ABB and Eaton, have very large US operations but foreign parents.
Are transformers actually made in the USA? Yes, across every band from 15 kVA dry type to 750 MVA power transformers through 765 kV. The constraint is not capability, it is capacity and core steel. Published estimates put the domestically built share of power transformers above 100 MVA at 18% as of 2019, improved since but not closed.
What is the BABA domestic content requirement for transformers? Transformers are manufactured products under 2 CFR Part 184, so the unit must be manufactured in the US with US-origin component cost above 55% of total component cost. Iron and steel used in the project face a stricter rule with no percentage test: every process from initial melting through coating must occur in the US.
How is BABA different from the Buy American Act? Buy American covers direct federal procurement and sets the component cost threshold at 65% for 2024 to 2028 deliveries, rising to 75% from 2029. BABA covers federal financial assistance such as grants and loans, and sets it at 55%. Different thresholds, different documentation, different waiver procedures.
Do Canadian and Mexican transformers count as domestic? No. For domestic content purposes, domestic means the United States. USMCA does not change that. This is a common and expensive surprise on projects sourcing from Canadian or Mexican plants of familiar brands.
What tariff applies to imported transformers in 2026? Since 6 April 2026, Section 232 duties on steel, aluminium and copper derivatives are assessed on the full customs value rather than metal content. Most transformer classifications fall in the 25% derivative tier where metal is at least 15% of weight. Units above 10,000 kVA under HTS 8504.23 sit in a temporarily reduced band through 31 December 2027. A 10% rate is available where at least 85% of the applicable metal content was US melted or smelted. Chinese origin can add Section 301 layers. Confirm the specific stack with your broker.
How long does a transformer take to deliver in 2026? Dry type under 1 MVA runs 8 to 16 weeks. Distribution pad-mount runs 30 to 50 weeks. Substation class 5 to 50 MVA runs 75 to 110 weeks. Large power above 100 MVA runs 110 to 130 weeks, with many quotes past 160. Generator step-up units average about 144 weeks, and Tier 1 OEMs are quoting 48 to 60 months.
Why are transformer lead times still so long? The binding constraint is grain-oriented electrical steel. NEMA told DOE in writing that the United States has a single domestic producer of it. Layered on top: demand up 116% for power transformers and 274% for GSU units since 2019, a modelled 30% supply deficit for power transformers, and a fleet where over half of in-service distribution units are past their 33 year expected life.
What is the DOE 2029 transformer standard? DOE’s final rule of 22 April 2024 (89 FR 29834) requires compliance on and after 23 April 2029 for 60 Hz distribution transformers with primary at or below 34.5 kV and secondary at or below 600 V: liquid-immersed 10 to 5,000 kVA and dry-type 15 to 5,000 kVA. Compliance attaches to the date of manufacture. GSU units and anything above 5,000 kVA or 34.5 kV are outside the rule.
How much does a distribution transformer cost per kVA? Indicative figures run about $22 per kVA at 100 kVA, $10 at 1,000 kVA and $7.20 at 1,600 kVA, a fall of roughly 3.1 times across that range before flattening. Treat the shape as useful and the level as unquotable: published ranges carry no date, Incoterm, winding material or efficiency tier.
Is buying a remanufactured transformer sensible? Often yes. A fully remanufactured distribution unit, stripped, rewound where needed, retanked, refilled and tested to ANSI, is reported at 50% to 70% of new cost in four to six weeks against 30 to 50 weeks for new. For storm response or a schedule critical expansion it is frequently the only option that holds the date. Check the test report and the warranty.
How do I check whether a transformer is really made in America? Ask four questions and get the answers in writing: where is the core manufactured, where was the core steel melted and poured, where are the windings produced, and which specific plant will build your order. Then request a project-specific manufacturer self-certification letter. There is no government BABA certificate or registry, so the letter is the document an auditor will ask for.
