Loading…
Loading…
1 critical material scored · binding chokepoint: Lithium (🇨🇳 CN 65% of refining) · 40 restrictive government measures on record
A verification pass re-checked this dossier's ownership/corporate-structure fields against their cited sources. It did not re-read the material_exposures claim the score, band and stress figures below are built on — treat those as not yet independently re-checked.
Tianqi Lithium produces 1 of the 1 scored material above (Lithium). For those, a supply restriction by the controlling country is a tailwind, not a headwind — the exposure is to disruption of a market this company supplies, not to a chokepoint it depends on. Every scored material here sits on its output side, so the Moderate · 54/100 band should be read as chokepoint salience, not as buyer vulnerability, and the Art. 24 input-side duties below are qualified accordingly.
Role from an explicit dossier role: tag or the producer-sector classifier behind the /minerals alternatives bench (one classifier on disk, generated 2026-10-07) — the same source the company page uses. A material the classifier has no entry for defaults to a buyer dependency, which can understate a producer's output side. Descriptive classification only: it enters no score.
The binding exposure is Lithium — 🇨🇳 CN controls 65% of global refining. On this company's production footprint that scores 54/100 (partially hedged; global 61). The register holds 40 restrictive government measures touching this company's materials — each traced to its primary source below.
Peer rank · Lithium Tianqi Lithium is the 423rd-most-exposed of the 487 named companies we track on 🇨🇳 CN's Lithium chokepoint; the most-exposed is Momentum Technologies (72/100). Ranked on the same footprint-adjusted buyer score as above — a relative read of an existing metric, not a new one.
Tianqi Lithium ranks 77th of 99 verified chemicals companies, tied with 3 others at 54.
Same sector_primary, ranked on the company supply-risk index. Restricted to hand-verified dossiers — 13 further chemicals companies are tracked but auto-onboarded, and excluded here because their exposure list is a sector template rather than company research. A peer scoring lower is the useful read: it usually means a different production geography or a qualified second source.
Company supply-risk index 54/100 — the binding chokepoint dominates, with a modest add for exposure breadth across 1 scored material. Buyer-relative (first-order): weighted by where the company produces (AU 55% · CN 45%, estimated split — no cited source states these exact shares), applied across all materials — it does not yet trace each input to its specific sourcing step.
Disclosed production sites
Named plants and what they make, from the company's disclosures. Descriptive detail — the buyer score above is still driven by country-level footprint weights, not per-site material intensity.
Tianqi Lithium (SZ: 002466; also HKEX: 9696) is one of the world's largest lithium chemical producers, vertically integrated from mine to finished compound. It holds a majority stake (via Tianqi Lithium Energy Australia, alongside IGO) in Talison Lithium, which operates Greenbushes in Western Australia — the world's largest hard-rock (spodumene) lithium mine — and converts spodumene concentrate into battery- and industrial-grade lithium carbonate, lithium hydroxide monohydrate, lithium chloride, and lithium metal at its own Chinese processing plants and at its Kwinana lithium hydroxide refinery in Western Australia. It also holds a roughly 22% stake in Chile's SQM, giving it a second leg into brine-based lithium supply.
Lithium is not a raw material Tianqi sources and consumes; it is the company's entire business — it mines spodumene at Greenbushes and converts it (plus SQM brine-sourced supply) into the battery- and industrial-grade lithium chemicals it sells. This makes Tianqi itself a node in the lithium supply chain that downstream battery and EV makers depend on, rather than a company exposed to someone else's lithium chokepoint. Its exposure is therefore price/policy exposure (Australian mining royalties/export terms, Chilean lithium-sector state policy via the SQM stake, and Chinese processing-capacity utilization) rather than input-scarcity exposure.
(merged from tianqi-lithium.md 2026-08-14). Greenbushes alone accounts for a large share of world hard-rock lithium concentrate output, and its ownership is split between one Chinese (Tianqi/IGO joint arrangement via Windfield) and one US (Albemarle) group — so output allocation and offtake terms are a recurring policy-relevant friction point between Chinese and Western battery supply chains. Downstream refining capacity is likewise split between Australia (Kwinana) and China (Sichuan), giving Tianqi a foot in both a Western-jurisdiction and a China-jurisdiction processing base. (This is the structural counterpart to the Albemarle JV-offtake reading trap documented under "Named counterparties" below.)
2026-08-14)*. Talison's Phase III chemical-grade spodumene expansion began commissioning in December 2025 (targeting 2.14Mt/yr total concentrate capacity) but a fire at the new plant in mid-2026 has disrupted the production ramp-up, illustrating that near-term supply growth from this chokepoint is not guaranteed on schedule.
The sector-derived stub previously listed phosphate, potash, vanadium, nickel, cobalt, silicon, antimony, and manganese as typical chemicals-sector inputs; none of these are verified in Tianqi's actual product line (battery/ industrial lithium carbonate, lithium hydroxide monohydrate, lithium metal, lithium chloride) and they have been dropped. The deleted tianqi-lithium.md independently reached the same conclusion against the mining-metals sector default (cobalt, copper, chromium, manganese, nickel, aluminium, niobium, vanadium), and additionally noted that Tianqi's processing byproduct streams (sodium sulfate, aluminium-silicon powder from tailings) are waste-stream byproducts, not a production line — correction 2026-08-20: aluminium IS a scored material generally, but this tailings powder is not a metallurgical aluminium exposure and stays excluded on that basis, not on scored status. Neither rejected list is reinstated here.
Source: Tianqi Lithium FY2025 annual report (天齐锂业股份有限公司 2025 年年度 报告全文), filed with Shenzhen Stock Exchange via cninfo 2026-03-28 — PDF. All figures below cover FY2025 (calendar 2025) and are the filing's own; they are not a current position. Total FY2025 operating revenue: RMB 10,346,364,563.27 (p.10).
Tianqi is on the CSRC Art. 20 semi-annual grid; the H1-2026 semi-annual report (天齐锂业股份有限公司 2026 年半年度报告全文, unaudited — "半年度财务报告是否已经审计: 否") filed 2026-08-28, three days ahead of the 31 Aug batch date — PDF, 188 pages, retrieved in full via scripts/py/cninfo_filings.py 002466 --semi-annual.
Finding: no top-5 customer/supplier concentration table in this filing at all. A full-text search for "前五名"/"前五大" plus every "销售总额"/"采购总额" phrasing the FY2025 table used found zero matches to a sales- or procurement-concentration table — the disclosure below is annual-report-only for this filer, not something the interim omitted by oversight; the interim's own related-party-transaction section (§14, below) is complete and detailed, so this is a scope difference in the CSRC format, not a retrieval failure. Practical effect: Albemarle / Northvolt / LG Chem below carry no interim update — their next possible refresh is the FY2026 annual report, due around March 2027, not this cycle.
The interim's own §14 关联方及关联交易 (related-party transactions) table — which is where several other dossiers in this corpus (e.g. Xiamen Tungsten) got their H1-2026 named-counterparty rows — has zero rows on the purchases side and exactly one on the sales side: 雅江县润丰矿业有限责任公司 (Yajiang Runfeng Mining), consulting services, CNY 1,117,957.93, no prior-period comparative. Recorded as a named_counterparties row above for completeness; it is immaterial in size and does not bear on the concentration structure below.
| Top-5 total | % of annual total | of which related-party | |
|---|---|---|---|
| Customers | RMB 6,957,553,173.27 | 67.25% | 0.00% |
| Suppliers | RMB 2,550,678,478.39 | 31.38% | 0.00% |
Customer concentration is extreme — two-thirds of revenue in five accounts — and the filing declares all of it arm's-length.
Named in the top-5 table (primary-source, FY2025):
| Rank | Customer | Sales (RMB) | % of annual sales |
|---|---|---|---|
| 1 | Albemarle Corporation (雅保) | 4,070,021,293.47 | 39.34% |
| 2 | 第二名 — withheld | 1,231,073,504.26 | 11.90% |
| 3 | 第三名 — withheld | 676,141,112.54 | 6.54% |
| 4 | 第四名 — withheld | 524,036,665.12 | 5.06% |
| 5 | 第五名 — withheld | 456,280,597.34 | 4.41% |
Ranks 2-5 are anonymised in the filing as literally "the second/third/fourth/ fifth". Per charter discipline the concentration structure is captured and the names are recorded as withheld; no inference was attempted.
Named in the major-sales-contracts table (§4, p.43; contract terms p.121):
Both contracts are explicitly marked 否 / 非关联方 (not a related-party transaction) with pricing 双方协商确定 (negotiated between the parties).
All five top suppliers are anonymised (第一名 … 第五名), at 11.99%, 7.75%, 5.01%, 3.48% and 3.15% of annual procurement. The A-share lane yields the supplier concentration structure for Tianqi but no supplier names. This is a disclosure limit, not an access failure — the filing was retrieved and read in full.
This dossier is the inverse of [Yunnan Tin](yunnan-tin.md), and the contrast is the point of the arm's-length correction. At Yunnan Tin the disclosed 32.80% top-5 collapsed to 16.51% arm's-length once the parent roll-up was removed — the headline overstated external concentration. Here the headline understates relatedness:
called "境外经营实体少数股东客户" — customers who are minority shareholders of overseas operating entities — and states its principal member is Albemarle Corporation, "which holds 49% of the Group's significant non-wholly-owned subsidiary Windfield Holdings Pty Ltd through RT Lithium".
(IGO Limited 49%) → Windfield Holdings (Albemarle 49%) → Talison Lithium → Greenbushes.
So the single largest customer, at 39.34% of revenue, is the 49% co-owner of the consolidated subsidiary that owns the mine. That is a JV offtake, not a market sale — Albemarle lifting its share of Greenbushes spodumene, which appears as third-party revenue only because Tianqi consolidates Windfield. The 0.00% is the filing's own classification under CAS (a co-venturer in a subsidiary is not automatically a related party of the parent), so it is reported here unmodified and related_party: false matches the primary source. But any concentration read that takes 67.25%-at-arm's-length at face value is wrong in the more dangerous direction.
The FY2025 revenue-by-region split (p.42) is mainland China RMB 9,475,013,711.20 vs overseas RMB 871,350,852.07 — overseas is 8.42% of revenue. Yet Albemarle, a NYSE-listed US company, alone accounts for 39.34%. The two disclosures can only be reconciled if the Albemarle sales are booked inside the mainland-China bucket, implying the regional split tracks the contracting/booking entity rather than customer domicile. The filing does not say this, and it is flagged here as an unresolved reading trap rather than a conclusion — anyone using Tianqi's regional revenue split as a proxy for customer geography should verify this first.
Retrieved with scripts/py/cninfo_filings.py 002466 --index 0; ticker resolved through cninfo's own topSearch to 天齐锂业 / orgId 9900014189 before fetching. Column mapping for the Northvolt row (which has blank cells) was read off the parallel LG Chem row in the same table.
ownership_note.From the company’s own mandatory filings — the CSRC top-5 customer/supplier concentration disclosure and the related-party tables. This is a disclosure, not a netting: a named supplier concentration is shown beside the exposure score and never adjusts it. Figures are the fiscal years labelled, not a current snapshot.
Inverse of the Yunnan Tin pattern: the filing reports 0.00% related-party sales, but rank-1 (39.34%) is Albemarle, the 49% co-owner of consolidated subsidiary Windfield/Talison — economic relatedness is understated, not overstated, by the headline.
Rank-1 customer, RMB 4,070,021,293.47. Filing classifies it non-related, yet its own receivables policy files it under the receivables category 境外经营实体少数股东客户 (overseas operating-entity minority-shareholder customers; AR 2025 p.167 footnote names Albemarle as the main such customer) — it holds 49% of Windfield via RT Lithium. The overseas operating entity is Windfield/Talison, whose only mine is Greenbushes (Western Australia), so these are Australian spodumene-concentrate sales; the filing does not say where Albemarle takes delivery or converts (Kemerton AU or China), so the route stops at AU>? — not a CN>US flow.
Lithium hydroxide long-term supply agreement signed 2019-09-24 with Tianqi Lithium Kwinana; RMB 963.33mn cumulative performance, no FY2025 performance, RMB 51.01mn overdue at period end. Counterparty is bankrupt (Northvolt Ett Expansion AB declared bankrupt 2024-10-08, parent Northvolt AB 2025-03-12) and its remaining assets were acquired by Lyten (completed 2026-02) — treat as a closed relationship, not an active one. Verified 2026-08-25. Route (wake-trade-pairs 2026-09-29): the FY2025 major-contracts table names the seller entity as Tianqi Lithium Kwinana Pty Ltd, i.e. the Kwinana WA hydroxide plant, not a Chinese plant; the filing states no shipping route or delivery point, and the contract had no FY2025 performance, so no 2024+ physical flow is evidenced. Origin AU by contracting entity only; destination unstated.
Tianqi Lithium AR2025, major sales contracts table: lithium hydroxide products, counterparty LG Chem, Ltd, RMB 20.99mn (2,098.52万元) performed in FY2025, RMB 297.16mn cumulative, receivables normal. Contracting seller is 成都天齐锂业有限公司 (Chengdu Tianqi, CN subsidiary), signed 2022-06-27 (announcement 2022-035), whereas the Northvolt contract is signed by Tianqi Lithium Kwinana. Neither the AR nor the 2022 announcement (Sina 2022-06-27) names the producing plant (Tianqi has hydroxide output in Sichuan/Chongqing CN and Kwinana AU) or the LG Chem delivery site, so the route is unstated.
H1-2026 semi-annual report §14 related-party transactions: consulting services sold, CNY 1,117,957.93, no prior-period comparative shown. Related party via same ultimate controller (Jiang Weiping). Immaterial in size (~0.01% of FY2025 revenue) but it is the ONLY related-party goods/services transaction disclosed this period — the purchases side of the same table has zero rows. Not a substitute for the Albemarle/Northvolt/LG Chem top-5 disclosure, which does not appear in this filing at all (see prose section).
Alternative track — a counterparty read from primary filings, never merged into the exposure score. Absence of a name is not absence of a relationship: A-share issuers anonymise counterparties they are not required to name. · section source filing ↗
Ranked by buyer-relative risk, highest first.
0 of 1 of your scored CRMA-strategic material breach the EU’s own Art. 5 65% single-third-country ceiling (global-production proxy).
| Material | Controlled by | You | Global | Band | Art. 5 | Input share | Substitute | Laws | Trend |
|---|---|---|---|---|---|---|---|---|---|
| Lithium | 🇨🇳 CN 65% refining | 54 | 61 | Moderate | within 65% | High | some | 40 | ▲ rising |
You = buyer-relative score (this company's disclosed footprint vs. the controller). Global = buyer-agnostic supply risk. Substitute = ease of swapping the material out (none = locked in). Input share = the material's disclosed magnitude in the company's input basket (HIGH/MED/LOW only where a public filing quantifies it; — = unrated). Descriptive effect-size, never scored.
Art. 5 = does the global top single-country share breach the EU's own CRMA Art. 5 diversification ceiling (no more than 65% of a strategic raw material from a single third country)? A conservative global-production PROXY for the EU-import denominator — descriptive only, sits beside the score, never merged into it (— = non-strategic material). Reg. (EU) 2024/1252 Art. 5 ↗
Per-material factor scoring on a 1–5 likelihood×impact scale, mapped to the Art. 24(2)(b) risk-factor framework. The headline score above is a portfolio RAG; this matrix is the assessment — it is where two companies with the same binding chokepoint diverge.
| Material | Geopolitical | Concentration | Price / market | Substitutability | Import reliance | Logistics · ESG · Supplier |
|---|---|---|---|---|---|---|
| Lithium | 4 | 3 | 5 | 3 | 3 | company input |
1 = very low … 5 = very high — a standard supply-risk likelihood×impact scale (the form a competent authority expects for the Art. 24(2)(b) factor analysis, not a CRMA-numbered scale). Public-source factors are pre-filled from the engine's primary sources (USGS concentration, IPTM government actions, EU import data); the three rightmost factor categories need company / Tier-1 supplier data and are flagged as input under Art. 24(3). Hover any cell for its evidence.
Every new filing and every amendment (rate change, scope change, repeal) touching this company's materials in the window above. Append ?since=YYYY-MM-DD to this URL for a custom start date.
No filings or amendments in this window — the register has been quiet on this company's materials.
Restrictive government measures on this company's materials, newest first — each links to its primary government source.
+ 25 more in the register.
The Art. 24(2)(c) vulnerability assessment, made explicit. For each leading exposure we model the move in this company's buyer-relative score under two distinct supply-disruption scenarios — the production footprint held fixed, only one lever moved at a time so each delta isolates one shock:
Under the 🇨🇳 CN shock, these disclosed plants carry the binding Lithium exposure:
Counterfactual: Indonesia extends the hilirisasi ore-ban template (2020 nickel → 2023 bauxite) to the next rung of battery-mineral exports — tightening upstream supply for cobalt intermediates, lithium feedstock and graphite alongside the existing nickel + aluminium regime. Direct-hit lines are basket issuers whose binding material is a battery-cell input (nickel, cobalt, lithium, graphite) — irrespective of controller, since the template-export is global supply-chain pressure not bilateral targeting.
The binding exposure this precedent lands on — Lithium — is a material Tianqi Lithium produces, so this is an output-market event for this company, not a supply vulnerability. No modelled stressed delta is shown: the buyer-relative stress models a rising cost of an input, which is the wrong direction for a supplier of the material, and we would rather show no number than a wrong-signed one. It is never netted against the consumer-side levers in §6.4 — those are reported separately.
role: tag or the producer-sector classifier (one classifier on disk, generated 2026-10-07) — for this company the basis is a disclosed dossier tag. It enters no score.🇨🇳 CN has issued 6 restrictive actions on Lithium since 2021 — cadence accelerating (mean gap 339d → 306d), severity flat (3.3 → 2.7).A descriptive trajectory of past official actions — not a forecast.
You hold exposure to 1 of these 13 materials (Lithium) — your binding Lithium exposure is one of them.
Demonstrated cadence: 🇨🇳 CN has widened its restricted-material list a median of 20.8 months apart across 4 distinct restriction dates since 2021 (n=3 intervals).
Response coupling: when 🇨🇳 CN restricts, our causal register records these counter-moves —
| Type | Scenario | Today | Stressed | Δ |
|---|---|---|---|---|
| Policy | Lithium — 🇨🇳 CN escalates lithium controls to a full export-licensing / ban regime | 54 | 58 | +4 |
| Concentration | Lithium — 🇨🇳 CN becomes the single source for lithium — the second source is lost (full 65%+ monopoly) | 54 | 70 | +16 |
A zero delta means that lever is already modelled at maximum on that material — today's score already prices it in. This is why the two scenarios are shown together: where a material's policy lever is already maxed (zero policy delta), the concentration shock still carries a real delta, and vice-versa. Each stressed score isolates its one lever; all other factors are held at current values.
No material crosses the significant-vulnerability threshold on the input side — every scored material here is one Tianqi Lithium produces, and Art. 24 addresses the use of a strategic raw material as an input. The Art. 24(4) mitigation duty is not triggered on the public-source evidence; the mitigations below are precautionary.
Stated threshold (so the conclusion is reproducible and auditable): buyer-relative band ≥ High AND substitutability hard/none AND ≥ 1 in-force restrictive measure on the material, assessed over the 0 materials this company buys (the 1 it produces are excluded from the test and listed above). The CRMA does not fix a numeric definition of “significant”; the company may adopt a stricter or looser threshold and should record it here.
Proposed, announced or draft regulation that is not yet in force but would touch this company's at-risk materials if it passes. Forward-looking early-warning — the likelihood shown is an honest band derived from the legislative stage, not a forecast or a fabricated probability. Kept separate from the enacted register above: nothing here is law yet.
Bills at introduction (pre-committee) in US historically become law ~5% of the time (n=37,132, GovTrack — 117th–118th Congresses) — a base rate for comparable bills, not a forecast for this one. source ↗
Likelihood band is derived deterministically from the legislative stage (announced → low; draft-published / in-consultation → moderate; passed-committee → elevated; passed-vote / awaiting-signature → high) — a reproducible, source-traceable proxy, not a probability estimate. Where shown, the modelled impact-if-passed re-uses the same buyer-relative stress engine as the enacted scenarios above: it holds this company's production footprint fixed and escalates the proposed measure to a full export-licensing / control regime — the conservative upper bound for a measure that may pass only as a partial cap. The delta is the move from today's score to that stressed score; companies with no modelled production footprint show no delta.
Forward-looking read on the binding chokepoint, from the recent trajectory of policy on these materials. Directional, not a forecast.
Every scored material here is one Tianqi Lithium produces, so the Art. 24(4) buyer levers — qualify an alternative supplier, re-source, substitute the input — do not apply to this company. The output-side items below are what a concentrated producer's risk office actually acts on. We render them rather than a generic diversification list because a prescription addressed to the wrong side of the market is worse than none.
Under the EU Critical Raw Materials Act (Reg. (EU) 2024/1252), a Member State identifies the large companies (Art. 2(29): >500 employees and >€150M net worldwide turnover) using strategic raw materials to manufacture a listed strategic technology (batteries, renewables, hydrogen, traction motors, heat pumps, aircraft, data-storage equipment, robotics, drones, satellites, advanced chips). Those companies must, at least every three years and to the extent the information is available to them (Art. 24(2)), assess their strategic-raw-material supply chain. Where suppliers do not provide the data on request, the assessment may rely on the Commission's monitoring dashboard (Art. 20(4)) or other publicly available information (Art. 24(3)) — which is the evidence base this report assembles. Board reporting (Art. 24(5)) is voluntary unless the Member State mandates it (Art. 24(6)).
| CRMA provision | Obligation | Where addressed |
|---|---|---|
| Art. 24(1) | Member State identifies the company as in-scope (uses an SRM to make a listed strategic technology). | Scope & applicability |
| Art. 24(2)(a) | Map where the strategic raw materials are extracted, processed and recycled. | Exposure register + Supply-risk factor analysis |
| Art. 24(2)(b) | Analyse the factors that might affect supply. | Supply-risk factor analysis (factor matrix) + The laws that threaten it |
| Art. 24(2)(c) | Assess vulnerabilities to supply disruptions. | Stress test + significant-vulnerability conclusion |
| Art. 24(3) | Where supplier data is unavailable, rely on Commission (Art. 20(4)) / public sources. | This report's basis — see Methodology & sources |
| Art. 24(4) | Where significant vulnerabilities are found, assess diversifying or substituting. | Significant-vulnerability conclusion + Priority mitigations |
| Art. 24(5)–(6) | Report results, sources, significant risks and mitigations to the board. | This document — board-ready, PDF-exportable |
This report pre-fills the Art. 24(3) public-source half of the assessment. The company-specific inputs — employee/turnover thresholds, bill-of-materials volumes, the tiered supplier map, and formal board adoption — remain the company's to complete; they are flagged as “company input” where they appear.
Article 24 applies only when both size thresholds are met and a Member State has identified the company as making a listed strategic technology with strategic raw materials.
| Threshold test | This assessment |
|---|---|
| Average employees (last FY) > 500 | company input |
| Net worldwide turnover (last FY) > €150M | company input |
| Uses a strategic raw material as an input | company input — all 1 scored SRM here is one this company produces, not buys; input use is not evidenced by this assessment |
| Manufactures a listed strategic technology | chemicals (confirm against Annex) |
| Formally identified by a Member State authority | company input |
Production-concentration figures: USGS Mineral Commodity Summaries 2026 + the production dataset behind each material page. Policy measures trace to the primary government sources below.
Each material's global supply-risk index blends five weighted factors: concentration of refining/processing (35%), active trade-control & policy pressure (25%), import reliance (15%), substitutability (15%), and price stress (10%). The buyer-relative score then scales the relational factors (concentration / policy / import) by this company's production-footprint alignment against each material's controlling country — bloc-neutral factors (substitutability, price) are left intact.
Caveats. The footprint is the company's assembly / manufacturing geography applied uniformly across all materials — a first-order proxy, not per-material input tracing. Scores are an analytical judgement on public data with a transparent weighting, not a market forecast or investment advice. Production shares reflect 2024-2025 figures and the policy position as of 2026-08-06; the register is continuously maintained and should be re-pulled against each new policy action.
MACROLENS · CICONIALABS · GEOPOLITICAL SUPPLY-RISK REPORT (EU CRMA ART. 20–25) · report generated 2026-10-07
Tip: the change log above defaults to the last 30 days. Append ?since=YYYY-MM-DD to this URL for a custom start date (e.g. ?since=2026-04-01).
This is a description of the actual automated pipeline (verifiable against this repo's own cron schedule), not a contractual commitment.