U.S. Steel formally broke ground on its $1.9 billion direct reduced iron (DRI) plant at Big River Steel Works in Osceola, Arkansas, on 30 October 2026, setting the company on a roughly three-year build toward first production in 2029. The ceremony itself is largely symbolic: the company describes only "preliminary site preparation" as underway, with main construction expected to begin in the latter part of 2026.
The more consequential developments for anyone tracking this project came earlier in the year — the board-level sanction of the investment and, in June, the selection of Midrex Technologies as process technology supplier. Those two decisions define what the plant will be. They do not yet define who builds it.
What is actually being built
The facility is a 2.5 million tonne-per-year MIDREX Flex plant capable of producing both hot direct reduced iron (HDRI) and hot briquetted iron (HBI). Midrex announced its selection on 15 June 2026, confirming first production for 2029.
The design intent is unusual for North America. The plant is sited immediately adjacent to Big River 2, the ~$3 billion expansion that took the campus past 6 Mt/y of steelmaking capacity when it came online in 2024. That adjacency allows DRI to be charged hot into the EAFs, avoiding the cooling, briquetting, shipping and re-melting that characterise the conventional merchant DRI trade. U.S. Steel has consistently framed this as a first in the United States: a DRI plant co-located with electric arc furnace steelmaking.
The Flex designation matters commercially. It gives U.S. Steel the option to run the plant in HBI mode, and Midrex's release states explicitly that HBI output is intended for merchant sales and export — a hedge against the day when captive HDRI demand and merchant pricing diverge.
Upstream, the feedstock strategy is already in place. The plant is designed around direct-reduced-grade pellets from U.S. Steel's Keetac operations in Minnesota, capability the company invested in from 2022. This vertical link — Minnesota ore to Arkansas metallics — is the strategic core of the project: it converts a purchased- scrap exposure into an integrated, self-supplied iron unit cost.
| Parameter | Detail |
|---|---|
| Capacity | 2.5 Mt/y HDRI/HBI |
| Technology | MIDREX Flex (Midrex Technologies) |
| Capex | ~$1.9 billion |
| Feedstock | DR-grade pellets, Keetac (Minnesota) |
| Configuration | Hot charge to Big River 2 EAFs; optional HBI for merchant/export |
| Construction duration | ~30 months (per AIST, May 2026) |
| Target first production | 2029 (H2 2029 per AIST) |
| Permanent jobs | ~200, plus ~35 embedded contractor roles |
| Peak construction jobs | ~2,000 |
Permits that predate the announcement
One of the more instructive details is how long the regulatory runway has been. Arkansas DEQ records for the Osceola site (AFIN 47-01073, held under the legal entity Exploratory Ventures, LLC) show a Title V/PSD air permit covering the DRI process — iron-oxide handling, reformer, furnace dedusting, briquetting, HBI cooling — issued as far back as January 2022, authorising roughly 2.2 Mt/y of briquette output. The active permit, 2445-AOP-R5, was issued on 1 April 2026, just ahead of the public announcement, with a renewal application (2445-AOP-R6) already in review.
A separate NPDES construction application for DRI wastewater facilities was lodged in November 2025. In other words, the environmental and water permitting for this plant was substantially de-risked before the board signed the cheque — a pattern worth noting for anyone trying to gauge how advanced other announced U.S. DRI projects really are.
The procurement picture: technology awarded, construction open
The commercial distinction that matters most right now is between what is contracted and what is not.
Awarded: the process technology package to Midrex. This is a significant award — the core reduction shaft, reformer and associated process design sit with Midrex.
Not publicly established: the EPC or construction management structure for the balance of plant, and by extension the structural and mechanical steel packages that flow from it. No EPC contractor has been named in U.S. Steel's releases, its Q2 and Q3 2026 earnings materials, or Midrex's announcements. References to SMS group or Primetals Technologies in connection with Midrex projects elsewhere (notably a 2025 Tosyali Libya order) should not be read across to Big River; no evidence links either firm to this contract. Primetals is a MIDREX licensee and could plausibly be involved, but that is inference, not fact.
With ~2,000 peak construction jobs planned and main construction only now starting, the structural steel, plate, pipe and mechanical works for the plant site are logically still ahead rather than behind. MetalMate's assessment flags these as unverified and possibly open. Readers should treat that as a reasonable expectation rather than an established opportunity: until an EPC award is announced, the packaging strategy — whether a single lump-sum turnkey, a split process/balance-of-plant structure, or multiple direct packages to U.S. Steel — is unknown, and that structure determines who actually tenders the steel.
Why it matters beyond Arkansas
Three implications stand out.
First, the U.S. metallics balance. Domestic DRI capacity has been thin relative to EAF steelmaking growth, leaving U.S. mills dependent on imported pig iron and HBI. A 2.5 Mt/y plant is a material addition to national supply, and its HBI optionality gives U.S. Steel a tradable product as well as a captive feed.
Second, the Nippon Steel context. The project sits inside Nippon Steel's roughly $11 billion U.S. investment commitment following its acquisition of U.S. Steel. That parentage underwrites the capex and signals that the Big River platform is the chosen vehicle for the group's American upstream integration, rather than the older Mon Valley blast-furnace assets.
Third, the quality dimension. Hot-charged DRI is not just a cost play; it supports lower-residual, tighter-specification steel for automotive exposed panels and similar demanding end uses — the product tiers Big River 2 was built to chase.
What to watch
The next genuine signals are, in rough order of importance: the EPC/consortium award; confirmation of the final investment decision language (the April 2026 board approval functions as FID, but U.S. Steel has not used that term publicly in the material reviewed); the construction package breakdown; and any HBI offtake agreements, which would validate the merchant-sales thesis. Start-up guidance has already firmed to 2029, with AIST reporting H2 2029 specifically.
Until the EPC structure is public, the construction supply chain for this plant remains an inference rather than a procurement fact.
