The Flat Global Steel Headline Hides a Violent Regional Reshuffle: Reading worldsteel's September 2026 Data

A 0.2% year-on-year change in global crude steel output looks like a non-event. It is not. The September 2026 worldsteel release, as reported by SteelOrbis, shows a global total of 142.4 Mt broadly flat against a relatively weak 2025 comparison — a stability that masks one of the wider dispersions in regional growth rates in recent memory. The Middle East is down 20.4%, Germany down 13.0%, Africa up 10.6% and Russia up 6.2% in the same month. For anyone planning procurement, trade flows or capacity utilisation into 2027, the regional detail matters far more than the headline.

The top line is weak arithmetic, not a strong market

The 0.7% decline for January–September (to 1,367.5 Mt across the 70 reporting countries, roughly 98% of world output) needs to be read against the 2025 series. World output fell 2.0% in calendar 2025, and the monthly 2026 sequence has been consistently negative: −6.5% in January, −2.2% in February, −4.2% in March, −1.9% in April, −0.3% in May, +1.7% in June, −0.3% in July, −1.2% in August and −0.2% in September. (These are the year-on-year changes reported in SteelOrbis's own monthly headlines over the period; I have not independently verified each month against worldsteel's releases.)

What that sequence tells us is that the year-on-year comparisons are progressively normalising because the 2025 base was weakening through the year, not because 2026 output is recovering. The three-month run rate (July–September: −0.3%, −1.2%, −0.2%) is essentially flat, which is the closest thing to a floor the market has found since early 2025. That is stabilisation at a low level, not a recovery.

Source: worldsteel, September 2026 crude steel production. Note the SteelOrbis report renders the CIS line as simply "CIS"; worldsteel's series is "Russia & other CIS + Ukraine", a wider definition that includes non-CIS Ukraine.

China's share is rising again — quietly

China produced 73.6 Mt in September, up 0.1% year on year. That is the first flat-to-positive monthly print for China in some time, and it lifts China's share of global output to roughly 51.7% (73.6/142.4) — my calculation — after a year in which Chinese output has repeatedly lagged.

The critical number is not September, though, but the cumulative: China's January–September output of 724.8 Mt is down 2.9%, implying a 2025 base of about 746.4 Mt and an absolute decline of roughly 21.6 Mt. Against that, the whole Asia & Oceania region is down only about 9.2 Mt year to date. The arithmetic is unambiguous: the rest of Asia and Oceania grew by roughly 12.4 Mt in the first nine months of 2026 while China shrank by 21.6 Mt. India alone accounts for about 6.9 Mt of that growth (130.0 Mt, +5.6%), with Viet Nam the standout at 23.2 Mt, up 29.1%.

That matters for 2027 because China's stronger September print sits against the backdrop of a cumulative Chinese contraction and an Indian and Southeast Asian expansion running at multiples of global growth. If Chinese output merely stabilises at September's rate rather than resuming decline, the global headline will turn positive on a weak Q4 2025 base — but the marginal tonne will continue to shift toward low-cost Asian supply outside China.

Germany: one month is not a trend

The most misreadable number in the release is Germany's −13.0%, down to 2.6 Mt. On its own it looks like a European industrial collapse. It is not inconsistent with a positive year.

Germany's January–September total is 26.5 Mt, up 4.0% year on year. Working backward: implied September 2025 base ≈ 2.99 Mt; implied January–September 2025 base ≈ 25.5 Mt; implied January–August 2026 ≈ 23.9 Mt versus implied January–August 2025 ≈ 22.5 Mt — roughly +6.2% year to date through August (my derivation from worldsteel's published totals). In other words, Germany's September was not weak in isolation but weak against an unusually strong September 2025. The correct reading is that German output is one of the few bright spots in the EU-27, which is down 0.8% year to date.

That said, a 13% single-month drop is too large to dismiss as base effect alone. The likely drivers — unconfirmed in the release — include mill maintenance and relining schedules, weak automotive and construction demand, elevated energy costs and the structural contraction of German carbon steelmaking capacity. The December and January prints will tell us whether September was a blip or the start of a second leg down.

The Middle East collapse is the real story

A 20.4% fall to 3.6 Mt is the largest single-month decline in the release outside the CIS. The implied September 2025 base is about 4.5 Mt (3.6/0.796) — meaning the region has lost roughly 0.9 Mt of monthly output. The year-to-date figure, 36.9 Mt at −9.2%, shows the collapse has been building through 2026 rather than appearing suddenly.

Worldsteel's regional taxonomy bundles Iran, the GCC, Iraq, Israel, Jordan, Lebanon and Syria into "Middle East". The composition matters: Iran is typically the largest producer in the group, with Saudi Arabia, the UAE and Oman also significant. Disruption concentrated in any one of those — power shortages and gas curtailments in Iran have historically been the principal driver — can dominate the regional aggregate. On the published data, it is impossible to attribute the decline precisely, but a ~20% regional fall is not compatible with a demand-side explanation alone. (This paragraph contains my analytical judgement, not facts stated in the release.)

For global markets the practical consequence is that the Middle East's shift from net exporter of semi-finished and rebar to a harder-to-supply position tightens availability in adjacent markets (North Africa, Turkey, South Asia) and supports billet and DRI premiums. Turkey's +0.8% to 3.2 Mt in September and +6.1% year to date to 29.8 Mt partly offsets this but is unlikely to fully replace lost Gulf and Iranian tonnes in regional trade.

The two-tier world in one table

Stripping the regions back to the ten largest producers shows the divergence more clearly than any commentary can:

(e) = worldsteel estimate. Ranking based on year-to-date aggregate.

Three observations follow. First, the US is quietly one of the strongest large producers in the world — 7.2 Mt in September (+4.3%) and 64.9 Mt year to date (+5.8%), supported by tariffs, reshoring-linked capital spending and a robust electric-arc-furnace sector. Second, Russia's September surge (+6.2%) reverses a cumulative decline of −3.6%, suggesting the September figure exaggerates underlying momentum and the year-to-date direction remains negative. Third, Viet Nam at +29.1% for nine months in a row is a genuine structural expansion, not a base effect, and it materially changes Southeast Asian supply balances and export competition in rebar and HRC.

What this means for the market

The release is a snapshot, not a forecast, but it carries four implications worth tracking into Q4 and 2027:

• Global output is flattening, not recovering. The three-month run rate sits near zero, but only because the 2025 base was weak. Any 2027 growth will come from emerging-market additions, not from cyclical recovery in mature markets.
• The regional rotation is real and accelerating. China, the EU, the Middle East and the CIS are net subtractors year to date; the US, India, Africa, Türkiye and other Europe are adders. Trade flows (semi-finished, DRI, HRC) will follow this rotation, with consequences for freight, spreads and trade-remedy activity.
• The European picture is better than its headline. Germany's September print looks dire but is a base effect against an exceptional September 2025. The EU-27 year to date is only −0.8%, and Germany is +4.0%. For European buyers this is a reason to hedge input supply rather than assume a repeat of 2024–25 weakness.
• Data caveats deserve attention. The reporting base moved from 69 to 70 countries from May 2026, and SteelOrbis's "CIS" label understates worldsteel's actual "Russia & other CIS + Ukraine" series. Russia's September figure is explicitly an estimate. Country-level year-to-date figures (which change interpretation materially) are omitted in the SteelOrbis write-up and should be read from worldsteel's own release before acting on any single regional number.

The bottom line is that September 2026 was a flat month in a flat first three quarters — but a flat aggregate built from violent regional swings. That volatility, not the headline, is what should inform inventory, procurement and trade strategy for the next twelve months.