A Note on Why Your Steel Costs Just Became Unpredictable

I spent the last three weeks on calls with mills, distributors, and operations directors at shops that consume 500 to 5,000 tons of steel per month. The conversation always turns the same direction: everybody is confused about what steel should cost right now, and nobody wants to be wrong on a three-year contract.

The surface story is simple. Global crude steel production hit 1.93 billion metric tons in 2025, up from 1.86 billion the year before. China still dominates output, but capacity utilization rates are the number that matters. Most mills worldwide are running at 70 to 75 percent of nameplate capacity. That should mean cheap steel. Instead, prices have been volatile and stubborn, refusing to collapse the way financial models predicted.

Here is what changed. China's domestic demand softened in late 2024 and never fully recovered. Real estate construction, which consumed roughly 35 percent of China's steel output historically, is still contracting. The government flooded capacity with stimulus, but mills were already making steel for an economy that did not need it. They exported the overflow. European mills, already running at 65 to 70 percent utilization, felt the pressure immediately. Several major European producers announced maintenance shutdowns in Q2 2026, which sounds like capacity reduction but is actually just mothballing unprofitable furnaces. They will restart them if prices move up, but that takes three to six weeks.

North American mills have taken a different path. US crude steel production held steady around 85 million metric tons annually through 2025 and into early 2026. What shifted is the product mix. Flat-rolled and structural grades, which command higher margins and matter most to fabricators and automotive suppliers, are being produced at lower volumes. Mills optimized for hot coil and rebar because those grades move faster and require less working capital. If you need 1/4-inch HSLA plate for truck frame fabrication, you are now competing for a smaller slice of available tonnage, and mills know it.

This creates real friction for operations directors writing RFQs. A sheet metal shop that historically locked in annual contracts with a single mill at fixed price plus quarterly adjustments now cannot get a quote that runs more than 90 days without escalation clauses. Distributors, stuck in the middle, are hedge-buying to lock in inventory at current levels, which pushes their cash float but protects their customers from sudden repricing. If you are a small to mid-size user without direct mill relationships, your distributor costs just went up 2 to 4 percent because they are paying to hold your margin.

The real issue is capacity expansion lag. New capacity in India, Vietnam, and Turkey is coming online, but ramping a modern mill to full utilization takes 18 to 24 months after first production. Meanwhile, demand recovery in developed markets is slower than historical norms. Automotive production is still below 2019 levels in most regions. Construction is uneven: some sectors are moving, others are flat. This mismatch between available supply and actual demand creates the perfect environment for price volatility and for mills to squeeze margins on customers who move fast and need certainty.

If you are in contracting mode now, negotiate shorter commitment windows. Three-month rolling prices with volume discounts will cost you more per ton, but you avoid the risk of being locked into supply at 85 percent of today's price while market conditions shift. If you can absorb the working capital cost, buy forward selectively on grades that are hard to source and that you know you will use. Skip the long-term contracts. Mills have zero incentive to hold prices stable when demand signals are this mixed, and you have zero visibility into what your orders will actually require six months from now.

This is the steel market as it actually exists right now: abundant globally, unevenly distributed regionally, and priced for optionality rather than certainty. Plan accordingly.