Mysteel's 13 November piece, "看得准、接得快、搭好台" ("seeing it right, responding fast, building the platform"), is not a market report. It is a conference sales pitch, published about five weeks before the 18–20 December 2026 gathering at Taihu, and its central device — a retrospective of forecasts the annual meeting supposedly got right first — is self-serving framing that should be read as such. Anyone who has watched trade-conference marketing will recognise the genre.

That does not make the article useless. Buried inside it are two threshold events dated 1 January 2026, a record export number for 2025 and one analyst call worth taking seriously. All four check out against primary sources, and the second half of 2026 has supplied a natural experiment on whether the market's bearish reading was right. On the evidence so far, it was.

The two rules, and a correction on each

Mysteel bundles together two unrelated regulatory changes that happened to share an effective date.

China's export licensing. The governing instrument is MOFCOM and GACC Announcement No. 79 of 2025, which adds "some steel products" to the 2025 export-licence goods catalogue. The official text is dated 9 December 2025 and takes effect 1 January 2026. Mysteel's article says 12 December; the official record does not support that date. The "approximately 300 HS codes" figure is also Mysteel's, not something visible on the notice page itself — the annex is issued separately — so treat the count as trade-press reporting rather than a documented fact.

What matters is not the tonnage coverage but the documentary requirement. Under the notice, an exporter must present the export contract and a product quality inspection certificate issued by the producer. That is a Chinese-language detail with real commercial consequences: the licence regime is aimed less at restricting volume than at displacing unverified product and broker-driven re-export of low-grade steel. It effectively forces a producer-of-record relationship and a paper trail onto transactions that previously ran through traders with no mill-side quality attestation. For small and mid-sized exporters and for the grey-channel trade into Southeast Asia, that is the binding constraint — not the licence form itself.

EU CBAM. Mysteel's phrasing — that from 1 January 2026 importers "must purchase and surrender certificates" — compresses the timeline misleadingly. The definitive regime does apply to imports from 1 January 2026, and importers above the single mass-based threshold of 50 t must hold authorised declarant status. But certificate purchase for 2026 imports only opens on the common central platform in February 2027, and the first declaration and surrender fall due by 30 September 2027. Firms that treated 1 January 2026 as a payment date rather than a data-and-registration date have mis-sequenced their compliance.

The cost signal is nonetheless already visible, because the Commission has been publishing quarterly certificate prices benchmarked to EU ETS auction averages:

Quarter of importCBAM certificate price (€/tCO2)Published
Q1 202675.367 Apr 2026
Q2 202675.286 Jul 2026
Q3 202682.325 Oct 2026
Q4 2026to be published4 Jan 2027

Source: European Commission. Pricing moves to weekly publication from 2027.

Applying those prices to a BF-BOF flat product with embedded emissions of roughly 1.8–2.1 tCO2 per tonne of steel gives a gross indicative exposure of about €135–170/t before any deduction for carbon already paid in China or for the phase-in of free allocation. That is an order-of-magnitude estimate based on typical emission factors, not a verified per-mill figure, but the implication is structural: at current Chinese export prices of roughly $694/t (2025 average, per Customs data), CBAM at those levels consumes a fifth to a quarter of the realised price band for conventional blast-furnace material. Mysteel relays Taiyuan Iron & Steel's claim that green-power short-process routes cut product carbon footprint by more than 75% and clear CBAM; that is a company assertion, unverified here, but the arithmetic explains why it is being made.

The record, and what happened after it

The 2025 export number is solid: finished steel exports of 119.02 Mt, up 7.5% year on year, above the 2015 peak of 112.4 Mt, per GACC data released 14 January 2026. December 2025 alone saw 11.3 Mt, up 16% year on year — a surge widely attributed to exporters pulling shipments forward ahead of the licensing start. The record was therefore partly a timing artefact, and the 2015 comparison flatters in another way too: it coincided with crude steel output falling below 1 billion tonnes to a seven-year low, so the export share of output rose sharply even as absolute tonnage set a record. Domestic demand weakness, not export competitiveness, is the dominant driver of the record — a distinction the Mysteel piece does not draw.

That is why the post-1 January data matter more than the headline. Wang Jianhua of Mysteel's own research arm argued at the December 2025 conference that licensing, overseas anti-dumping and CBAM would combine to cut 2026 exports by around 19 Mt. Reported outturns have been consistent with that direction, if not yet the full magnitude: Q1 2026 exports of 2,471.7 Mt-equivalent were down 9.9% year on year; January–July fell 4% to 650.1 Mt on Mysteel's own count; January–August were down 3.0% at 751.5 Mt. The deceleration in the year-on-year decline through mid-2026 reflects base effects and the diversion of trade into non-EU, non-traditional markets rather than any recovery in licensing-constrained channels. Plate has been the weak segment under anti-dumping pressure; long products and some emerging markets have held up better — exactly the product-mix and market-layout shift Wang predicted.

The trade-remedy backdrop reinforces the point Mysteel makes with its 2015 anecdote. As of December 2025, 395 steel anti-dumping and countervailing measures were in force or under investigation globally, 113 of them against China. The 2015 conference's "post-gusher" (后井喷) call was directionally right — exports fell for five consecutive years from 2016 to 2020 — but the current environment is harder: the constraint is no longer just trade defence but a compliance architecture of licensing, carbon accounting and destination-market rulebooks that a trader cannot arbitrage.

What the conference framing obscures

The retrospective claims credit for foreseeing CBAM as early as 2020, when a World Steel Association representative mentioned it in a conference speech. Predictions made in that format are cheap and the hits are remembered more than the misses; no one at these events is publishing a scored record. The more defensible claim is the "搭好台" (platform-building) one: Mysteel documents a genuine channel in which industry complaints about the 2021 export-rebate withdrawal led to a 47-code high-end product classification and three tariff-line subdivisions adopted by the State Council Tariff Commission for 2023. That is a real, traceable policy feedback loop and more interesting than the fortune-telling.

For industry readers, the practical takeaway is that the two rules of 1 January 2026 have shifted the basis of competition in Chinese steel exports from price and tonnage to documentation and carbon accounting. The 2025 record belongs to the old regime. The question the December conference is advertising — how to move from selling product to building overseas capacity, warehousing, standards and service — is the right one, and it is being forced by policy rather than chosen strategically. Whether 2026 exports settle near 100 Mt, as some sell-side estimates suggest, will depend less on demand in destination markets than on how quickly small exporters can produce the producer-side quality certificates that the licensing regime now demands, and how many can document verified embedded emissions below the default values that CBAM will otherwise apply.

Source & methodology +

MetalMate separates original-source evidence from summary and analysis. Verify the original source before procurement, legal or engineering decisions.

Article status
Published · Version 1
Source date
13 November 2026
MetalMate publication
14 November 2026
Verification
Automated evidence checks; not human reviewed

看得准、接得快、搭好台:这场大会究竟好在哪里? — Mysteel / 我的钢铁网
MetalMate provides an English summary; consult the original wording.