Steel Price Volatility and the CRU Index: What Framers Actually Pay (2026)
Most construction materials have a price. Steel has a formula.
That single difference reorganizes everything downstream of it — how you bid, when you commit, what your purchase order says, and, most importantly, what it means to check an invoice. When a supply agreement references an index, the invoice is supposed to differ from the original quote. Which raises a question almost nobody in the industry answers: how do you know the difference is right?
This is a companion to our steel framing software guide, going deeper on the commercial mechanics.
One disclosure: we build CostCrunch, which audits material invoices. We are explicit at the end about what we do and do not do with index-priced contracts, because this is an area where it would be easy to overclaim.
Where steel prices actually sit
Section 232 tariffs doubled from 25% to 50% in June 2025. By January 2026, steel mill products were up more than 20% year over year, and aluminum mill shapes up 33%.
Domestic hot-rolled coil crossed the $1,000 mark: CRU HRC at $1,002 per ton in March 2026, spot transactions at $1,040 in April 2026, around $947 mid-2026 by another source's measure, with projections toward $1,100 by the third quarter. Analysis attributes roughly 15% of the domestic increase to tariffs alone.
Those numbers will be stale quickly, which is the point. What matters more is the structure underneath them.
The supply side is the real story
Lead times on some products are approaching six to eight months — multiyear highs. Spot availability is limited, and some mills are capping how much steel customers can order under longer-term contracts.
Mills prioritize contract customers, which further shrinks what is available on the spot market for everybody else. The defining condition is constrained supply, not surging demand, and that distinction matters because demand-driven price rises correct when demand softens while supply-driven ones do not.
Analysts have described the tariff environment as giving US mills both pricing power and supply power simultaneously — raising prices while delivering lower quantities.
Two consequences that land directly on a subcontractor:
Smaller purchasers cannot route around it. Becoming a licensed importer requires bonding, customs brokerage, and tolerance for extended lead times. Below a certain scale that is not practical, so smaller firms absorb the increases.
Location matters more than it used to. Steel prices are not uniform across the country, and regional differentials are now large enough to change sourcing decisions.
How index pricing actually works
Here is the mechanic that makes steel purchasing different from every other trade.
The CRU US Midwest Hot-Rolled Coil Price Index is the dominant benchmark. It publishes weekly, it is the settlement price for CME US Midwest domestic hot-rolled coil futures and options, and it is referenced in over 95% of physical hot-rolled coil contracts in the United States.
So a supply agreement frequently does not contain a number. It contains:
- An index reference — which index, and which published series
- A date convention — the index value as of when? Order date, ship date, the month average, the week prior?
- An adder — a fixed amount over index, covering processing, freight, and margin
- Sometimes a collar — a floor and ceiling limiting how far the price can move
The date convention is where most of the ambiguity lives, and it is worth being pedantic about. "CRU Midwest HRC plus $180, priced at time of shipment" and "plus $180, priced at time of order" are materially different contracts in a rising market, and on a job with a six-month lead time the gap between them can be substantial.
Why this defeats normal invoice checking
Every standard control assumes a fixed expected price.
Comparing the invoice to the quote fails, because they are supposed to be different.
Three-way matching fails, because it confirms the invoice agrees with the purchase order and the delivery. If the purchase order says "index plus adder," the match is arithmetic on a formula that neither system evaluates.
A price-variance alert fails, because variance is the expected behavior. Set a threshold tight enough to catch a formula error and it fires on every legitimate movement; set it loose enough to be quiet and it catches nothing.
Checking an index-priced invoice properly requires the contract terms, the index value on the correct date under the correct convention, and about twenty minutes per line. On a job with dozens of releases, nobody does that, and the honest reason is that it is genuinely tedious rather than that anyone is careless.
The four mitigation strategies
What contractors actually do about the exposure, roughly in order of how commonly they are available.
1. Tiered pricing contracts linked to a steel index. The most common. It does not remove price risk; it converts price risk into formula risk and moves the exposure from the bid to the buy. That is usually the right trade, and it creates the verification problem described above.
2. Early purchase orders during schematic or design development. Issue the PO to lock a fabrication slot before the design is finished. This is genuinely effective on lead time and requires two things most projects do not have: owner commitment early enough to fund it, and a general contractor or construction manager who can administer early purchase contracts. When a project has both, it is the strongest tool available.
3. Locked-in pricing agreements. Where a mill or service center will offer one. Availability depends on your volume and relationship, and in a constrained-supply market these are harder to get precisely when you most want them.
4. CME or LME futures to cap input costs. Financially the cleanest hedge, and out of reach for most subcontractors — it requires treasury sophistication, margin management, and a board or owner comfortable with derivative positions. Larger fabricators do it; framing subs generally do not.
The one to be careful with is escalation clauses passed down to you without a corresponding clause passed up. If your supply contract escalates with the index and your subcontract with the GC is fixed, you have taken the entire exposure. That is a contract-review problem rather than a purchasing problem, and it is the most expensive mistake available in this trade.
What a steel framer should actually do
Write the date convention down. Not just the index and the adder — the exact convention, in the purchase order, in words. Ambiguity here is resolved in the seller's favor by default, because they are the ones producing the invoice.
Record the index value at commitment. Whatever the convention, capture what the index read when you agreed, so there is a baseline to reason from later. Without it, verification requires reconstructing history from published archives.
Separate the adder from the index in your records. The index moves and you cannot control it. The adder is negotiated and should be stable, so an adder that drifts is a different and more actionable problem than an index that rises.
Track landed cost, not the index. Freight, processing, and the regional differential all sit outside the index. Two suppliers on the same index with the same adder are not necessarily the same cost.
Aggregate before you negotiate. The strongest position in a renegotiation is knowing exactly what you bought last year, at what adder, on what convention, across which suppliers. That requires reading the invoices.
Where CostCrunch fits — and what it does not do
We build CostCrunch, so weigh this accordingly, and this is a topic where the limits matter more than the capability.
We do not validate index-formula contracts. If your agreement prices off CRU with a date convention and an adder, we do not ingest the contract, look up the index for the correct week, apply the convention, and confirm the arithmetic. We looked for a tool that does this and could not find one — which we are reporting as an unsuccessful search rather than as proof none exists. If a vendor tells you they do it, press hard on the specifics: which index series, which conventions, and what happens when the contract language is ambiguous.
What we do is narrower and still useful on index-priced material. Every line of every invoice is extracted at 99% accuracy and compared against your own purchase history and local market comparables, so a price that moved gets surfaced for a human to check rather than passing silently. On a commodity where movement is expected, that is the difference between a change nobody noticed and a change somebody looked at.
It is most useful on the parts of a steel framer's spend that are not index-priced — fasteners, accessories, track, hangers, and the general supply-house purchasing that surrounds the coil buy. Those behave like normal materials, and that is where the ordinary 4-8% overpayment pattern shows up. Across $125M+ in audited invoices from 500+ companies, that is $20,000 to $40,000 a year on $500K of material spend.
And it is useful for aggregation: what you actually paid, by supplier, by item, over time — which is the input to the renegotiation described above.
Frequently asked questions
What is the CRU index and how is it used in steel contracts?
The CRU US Midwest Hot-Rolled Coil Price Index is the dominant US steel benchmark. It publishes weekly, serves as the settlement price for CME US Midwest domestic hot-rolled coil futures and options, and is referenced in more than 95% of physical hot-rolled coil contracts in the United States. In practice this means a steel supply agreement often contains a formula rather than a price: an index reference, a date convention specifying which published value applies, an adder covering processing and freight, and sometimes a collar limiting movement.
Why is my steel invoice different from my quote?
If your contract is index-linked, it is supposed to be. The price is calculated from the index value at whatever date the contract specifies plus a negotiated adder, so a change in the index between quote and shipment produces a legitimately different invoice. The question is whether the formula was applied correctly — which index series, which date convention, and whether the adder matches the agreement. Verifying that requires the contract terms and the index value for the correct period, which is why it is so rarely checked.
How much have steel prices risen in 2026?
Steel mill products were up more than 20% year over year as of January 2026, with aluminum mill shapes up 33%. Domestic hot-rolled coil crossed $1,000 per ton, with CRU HRC at $1,002 in March 2026 and spot transactions at $1,040 in April, and projections toward $1,100 by the third quarter. Roughly 15% of the domestic increase is attributed to tariffs, after Section 232 tariffs doubled from 25% to 50% in June 2025. These figures move quickly and should be re-checked against current sources.
Why are steel lead times so long?
Constrained supply rather than surging demand. Lead times on some products are approaching six to eight months, which are multiyear highs, and some mills are capping how much steel customers can order under longer-term contracts. Mills prioritize contract customers, which further reduces spot availability for everyone else. Analysts have characterized the tariff environment as giving US mills both pricing power and supply power at once — raising prices while delivering lower quantities.
What is the difference between the index and the adder in a steel contract?
The index is the published market benchmark, such as CRU Midwest hot-rolled coil, which moves weekly and is outside anyone's control. The adder is the negotiated amount added on top, covering processing, freight, and margin, and it should be stable for the life of the agreement. Tracking them separately matters because an adder that drifts upward is a negotiable problem you can act on, while a rising index is a market condition you can only hedge or time.
How do contractors protect against steel price volatility?
Four approaches are in common use. Index-linked tiered pricing contracts, which convert price risk into formula risk and move exposure from bid to buy. Early purchase orders issued at schematic or design development to lock a fabrication slot, which is effective on lead time but requires owner commitment and a general contractor able to administer early purchase contracts. Locked-in pricing agreements where a mill or service center offers one. And CME or LME futures, which are the cleanest financial hedge but require treasury capability most subcontractors do not have.
Does three-way matching work on index-priced steel?
Not meaningfully. Three-way matching confirms that an invoice agrees with the purchase order and the receiving record, which works when the purchase order contains an expected price. When the purchase order contains a formula, the match becomes arithmetic on terms neither system evaluates, so the invoice can be internally consistent while the formula was applied on the wrong date convention or with the wrong adder. Price-variance alerting fails for a related reason: variance is the expected behavior, so a threshold tight enough to catch errors fires constantly.
What should be written into a steel purchase order?
At minimum the index reference including the specific published series, the date convention in explicit words — order date, ship date, prior week, or monthly average — the adder stated separately from the index, and any collar or cap. Also record the index value at the time of commitment, so there is a baseline for later verification without reconstructing it from archives. Ambiguity in the date convention tends to be resolved in the seller's favor by default, since they produce the invoice.
The uncomfortable summary is that an entire trade has moved to formula pricing without a corresponding control. Fixed-price purchasing has three-way matching, price benchmarking, and variance alerting. Formula pricing has none of those working properly, and the mitigation is a person with the contract, an index archive, and twenty minutes per line.
Until that changes, the practical defenses are contractual rather than technological: nail the date convention in writing, track the adder separately from the index, and aggregate what you actually paid so the next negotiation starts from evidence.
Try CostCrunch free on your own invoices and see what your non-indexed steel spend looks like line by line.
Last verified: August 19, 2026. Steel pricing, lead times, and tariff policy are moving quickly and the figures above will date fast — verify against current sources before relying on them. If something here is inaccurate, tell us and we'll correct it.