Huainan's Flat Week: Why a Stable 1/3 Coking Coal Quote Says More About Supply Than Demand

The price data itself is unremarkable: five consecutive sessions (12–16 October 2026) in which Mysteel's Huainan assessment held 1/3 coking coal at 2,090 yuan/t for Huaihe No.1 clean coal and 1,770 yuan/t for Huaihe No.2, both car-load (车板价), cash and tax-inclusive. Flat quotes for a week are normal in a market where the big Anhui producers set monthly or long-term contract prices rather than letting daily spot discovery do the work.

The more informative signal sits in the same week's national survey: Mysteel's 523-mine sample put coking coal capacity utilisation at 71.4%, up 5.2 points week-on-week, with output recovering in Shanxi's Linfen and Lüliang. Against a hot metal run-rate of 233.66 Mt/day — down a marginal 0.18 Mt/day — that is a supply-led week, not a demand-led one.

What the numbers actually say

Three things stand out from simple arithmetic on those figures.

The utilisation swing is large. If 71.4% is 5.2 points above the prior week, the starting point was 66.2% — a relative increase of roughly 7.9% in a single week across a 523-mine panel. Moves of that size are usually the unwinding of a disruption rather than organic demand pull, which fits Mysteel's attribution to recovering Linfen and Lüliang output.

Hot metal is flat, not falling. A 0.18 Mt/day decline on 233.84 Mt/day is 0.08%. On a 30.4-day month that run-rate implies roughly 710 Mt/month of hot metal — high by historical standards, but no longer expanding. Buyers are covering minimum requirements, not building inventory.

Coke is where the margin pain sits. A second round of coke price cuts (提降) landing in the same week as coking coal prices held flat compresses the coke plant's spread from both ends. That is the mechanism that eventually forces coking coal prices down: coke plants cut runs or bid coal lower. Neither has happened yet in Huainan.

Why Huainan holds while the national market chops

The Huainan quote is not a spot-clearing price in the ordinary sense. It reflects the pricing behaviour of large Anhui state-owned producers — Huaihe Energy's mining assets being the dominant local source — whose clean coal grades are sold predominantly on contract to East China steel and coke customers. Contract pricing moves infrequently and in steps; daily Mysteel assessments of those grades therefore read "stable" for long stretches by construction.

That makes the weekly stability a statement about contract discipline, not about a tight market. The relevant question for readers is not why Huainan was flat but how long the contract price can hold while the surrounding spot complex softens.

The level is the story

Huainan's 2,090 yuan/t for No.1 clean coal sits high in the national complex. Comparing against the same page's 16 October reference prices:

Two of these comparisons matter more than the others.

First, 1/3 coking coal is a mid-tier feed: it co-blends with premium hard coking coal and normally clears at a discount to it. A No.1 clean coal at 2,090 against Lüliang premium low-sulphur at 2,350 leaves only an 11% gap — narrow for these grades. Either Huainan's clean coal carries quality or logistics advantages that justify the premium, or the contract price is lagging a spot market that has already softened. In a week when national supply recovered sharply, the second explanation deserves weight.

Second, the Shandong comparison is the one to treat cautiously. The Jining figure of 1,450 carries no grade label on the source page, and Shandong's 1/3 coking coal has historically been quoted in a different band from its gas coal. A 44% gap between two East China 1/3 coal benchmarks is too wide to accept without checking the underlying specification and delivery basis.

A data caveat worth flagging

The same Mysteel topic page carries archived East China daily commentaries (日评) dated 16, 17, 18, 21, 22 and 23 October, plus November and December dates, all quoting Huainan 1/3 coking coal at 1,710/1,540/1,570 yuan/t for Huaihe No.1/No.2/No.3 — not 2,090/1,770. A 22 July commentary quotes 1,080/980/1,020.

These are almost certainly series from different years served on an undated archive page, or a mix of long-term contract and spot assessments. Readers should not read them as a 2026 price trajectory. One clue that a contract reset has occurred is an undated Mysteel note on the same page reporting that Pingdingshan (Henan) and Huainan/Huaibei (Anhui) producers raised long-term contract prices by 50 yuan/t. Confirming which basis the 2,090/1,770 quote represents — contract or spot, and for which delivery month — is essential before drawing any trend conclusion.

What to watch

Mysteel's own guidance for late October is a holding pattern: continued bargaining, prices broadly steady, with hot metal output and winter restocking (冬储) as the swing factors.

The analytical case for caution rests on three linked observations. Supply has just taken a large step up from a disrupted base. Coke margins are being squeezed by a second round of cuts, which historically transmits to coking coal within weeks. And hot metal, while high, has stopped rising. A flat contract price can coexist with all three for a while — but not indefinitely if the spot complex keeps softening around it.

The counterweight is winter restocking. Coke plants and steel mills building inventory ahead of the heating season can absorb the incremental Shanxi tonnes without forcing contract prices down. Whether that happens depends less on Huainan than on how quickly northern output continues to recover and whether steel margins hold through November.

References

• Mysteel (我的钢铁网), 近期淮南炼焦煤价格 topic page, including daily Huainan assessments for 12–16 October 2026 and the synopsis of Mysteel周报：国内炼焦煤市场概览（10.12-10.16）: https://www.mysteel.com/hot/1582048.html

Note on evidence: the analysis above draws on a single Mysteel topic page. Percentage changes, monthly run-rate conversions and cross-benchmark comparisons are the author's calculations from figures quoted on that page and should be treated as indicative. The 5.2-point utilisation change is read as a percentage-point move (66.2% → 71.4%); Mysteel's Chinese wording permits a relative reading, which would materially change the implied output step-up.