Steel Structure EPC Contractor or Something Else? Four Project Scenarios and What Actually Works in Each

I'm not an engineer. Let me get that out of the way first, because the question I keep getting pulled into — "steel or concrete, and do we need an EPC contractor for it?" — sounds like an engineering question. It isn't. Not really. It's a question about who carries the risk and whose signature is on the paperwork at the end.

I'm the office administrator at a 240-person manufacturer. I handle facilities and vendor contracts — roughly $1.6M a year across thirty-some suppliers — and I report to both operations and finance. In 2023 I got handed a building project I had no business being handed: a 14,000 sq ft clear-span storage building for our service fleet, which is basically a hangar with better lighting. Then in 2024 I sat in on vendor selection for a four-storey staff accommodation block where the design team had proposed a steel frame.

Nine months of quotes, two design revisions, and one fairly awkward finance meeting later. Here's the version of what I know that I'd actually tell a peer over coffee.

Why "steel or concrete" is the wrong first question

The first question isn't the material. It's: how much of this building do you want to buy from one company?

That framing matters because people collapse two separate decisions into one. Structure type is one decision. Delivery model is another — supply-only fabrication, a prefab design-build package, full EPC (design, fabrication, erection, sometimes foundations and MEP), or a conventional design-bid-build with a general contractor. You can pick steel and still buy it four different ways. The delivery model is what changes your exposure when something goes wrong at week thirty.

So instead of a single recommendation, here are the four situations I've actually been in or watched closely, and what I'd do in each:

• Scenario A — a big open box: hangar, warehouse, workshop. Simple geometry, few interfaces.
• Scenario B — residential. Repetitive rooms, heavy MEP, code-heavy.
• Scenario C — large, unusual, or deadline-critical.
• Scenario D — you already own a design and a general contractor.

Scenario A: A clear-span box under roughly 30,000 sq ft

Don't hire an EPC contractor for this. I'll say it plainly, because I watched a colleague do the opposite and pay for it. Buy a prefab package instead: fabricated frame, roof and wall panels, doors, delivered and erected, with the fabricator stamping the design for one clearly defined load case.

Now, aircraft hangar prices — everyone wants the number. We collected five quotes for a shell in the 12,000–14,000 sq ft range, meaning frame, cladding and erection. As of Q4 2025 they landed somewhere between $32 and $55 per square foot. That pricing was accurate as of Q4 2025. Steel, freight and door lead times move constantly, so verify current rates before you put a number in a budget.

The more useful thing I learned: the exclusions matter more than the price. Slab, doors, insulation, fire protection, MEP, permits, civil work, apron paving. A single large sliding door can run into six figures on its own, and it's almost never in the headline number. Span drives cost far more than tonnage does — once you push past roughly 35–40 metres clear, you're into truss or space-frame territory and the per-square-foot curve steepens fast.

Here's the part that surprised me. Our three cheapest quotes were within about 4% of each other per square foot. The $200k gap between the highest and lowest total came almost entirely from what one supplier had quietly left out. So compare scopes, not bottom lines. Ask every bidder for a written exclusion list, and treat a bidder who won't produce one as a bidder who has one.

The honest limitation: if your project touches aviation authority rules, fuel systems, foam suppression, or apron pavements, a prefab package supplier will hand back half your scope. That's not a knock on them — it's simply not what they sell. (Should mention: our building never needed any of that, which is exactly why the package route worked for us.)

Scenario B: Steel residential buildings

This is where I'd push back hardest, because the industry conversation around "prefab construction" tends to make steel sound inevitable. In our case it wasn't.

For a four-storey residential block with repetitive units, the steel option was not the cheapest, and it wasn't clearly faster either. On total installed cost it came out within low single digits of cast-in-place, and the local concrete crews were simply quicker because they'd done the same building forty times. Steel frame went up faster. Everything downstream of it did not.

Where steel does win residential: constrained urban sites with no laydown area, programmes where getting weather-tight early is the whole ballgame, seismic regions with well-established ductile detailing, buildings that need an open ground floor or long spans, and vertical extensions where a lighter frame means less foundation work.

And compare the right numbers. Not dollars per ton against dollars per cubic metre of concrete — that comparison is meaningless. Compare total installed: frame plus floor system (composite deck versus precast), plus fire protection, plus acoustics, plus finishes, plus foundation savings, plus programme. With steel you're buying either boards or intumescent paint to hit a fire rating, and both cost money and time. Acoustics in residential steel needs deliberate attention. That's not a defect, it's a line item — but it's a line item people forget.

One more fork: light-gauge steel and hot-rolled steel are almost different industries. For three storeys, cold-formed C-sections come from a different supplier base with different lead times than a hot-rolled fabricator. Make sure you're comparing like with like. At least, that's been my experience on the two residential packages I've been close to.

Scenario C: Large, complex, or under a hard deadline

This is where a steel structure EPC contractor earns the premium. Long spans, unusual geometry, heavy crane loads, high seismic or wind demand, a date that cannot move — that combination is where one throat to choke is worth real money.

The upside was one contract, one schedule, one warranty holder. The risk was losing visibility into a company that now controls design, fabrication and erection simultaneously. I kept asking myself: is the coordination saving worth potentially discovering a problem at month seven instead of month two? For us it was.

What to check before you sign, in buyer language rather than engineering language:

• Execution class. EN 1090-2 splits steel execution into EXC1 through EXC4. Buildings are typically EXC2; large-span, dynamic or cyclically loaded structures push into EXC3; EXC4 is special-case territory. Ask for the certificate and read the scope — it's issued per factory and per product type, not as a company-wide badge.
• Welding quality. ISO 3834-2 is the usual expectation for structural fabrication, with welder qualification to ISO 9606 and non-destructive testing to a stated standard and percentage. If a fabricator can't tell you their NDT percentage, that's your answer.
• Design basis. Which code, which edition, which wind and snow values, which seismic parameters. "We always do it this way" is not a load case.
• Interfaces. Who holds design liability? Who erects? Who submits for permits? Who signs completion? Who owns the shop-drawing approval cycle and how long does it take?
• Corrosion protection. ISO 12944 durability class, stated surface prep, stated dry film thickness. Vague coatings are how buildings look tired at year eight.

I have mixed feelings about the one-stop model. Part of me resents paying a coordination margin on work I could theoretically split. Another part of me remembers the alternative — five interfaces, four of which had never spoken to each other. I'll take the margin.

Scenario D: You already have a design and a general contractor

Then don't hire an EPC contractor. Split fabrication from erection and buy both directly. You keep control of the schedule, you keep the shop drawings in your own review loop, and in our case the split package came in below the EPC number that was on the table. I want to say around 8% lower, but I should add that we already owned the stamped design, so it isn't an apples-to-apples comparison.

The catch is capability. If nobody on your side can read a shop drawing and hold a tolerance, splitting the package just moves the risk to a gap between two contracts that no one owns. Erection-only subcontractors working under a GC who doesn't understand steel is how you get four months of finger-pointing and a rework bill. That said, if you do have that person, this route is the most cost-effective of the four.

How to tell which one you're in

Five questions. Answer them before you call anybody.

1. Do you own a stamped design? No — you're in A or C, depending on size.
2. Is the clear span over roughly 35–40 metres, or are there cranes and heavy dynamic loads? Yes — you're in C. Don't shop this one on price alone.
3. Does the building have hundreds of small rooms and heavy MEP? Yes — you're in B, and you owe it to your budget to price concrete honestly before committing to steel.
4. Is the site outside the contractor's certification territory? EN 1090 is a European harmonised standard; it does not automatically satisfy a local authority elsewhere. Ask specifically which local approvals apply. This catches more overseas buyers than anything else I've seen.
5. Who signs off to the authority, and who holds the structural warranty in year five? If the answer is fuzzy, you're in C by default — whether you planned to be or not.

The one thing I'd do differently

Almost everyone I've watched go through this picks the structure type and the supplier first, then works out the delivery model at the end because it's the least exciting decision. That's backwards. The delivery model is the decision that determines who's responsible when the cladding arrives in the wrong colour and the completion date slips by six weeks.

Pick the model first. Then the material. Then the company.