Does "Zimbabwe-origin" lithium diversify away from China?

Trade-flow companion to the price wedge (R72). Most rows in this corpus ask whether a consumer-side control is leaking (China cuts a buyer off; does the molecule sneak back under a laundered flag). This is a producer-side beneficiation case — the nickel/bauxite family — and it asks the inverse question that a country-of-origin exposure model gets most wrong: when a resource-nationalist ban forces processing onshore, does the new host-country origin actually move economic control away from the controlling country, or does it just re-flag the same owner's output? For Zimbabwean lithium the answer is the latter, in its purest form yet. This is a DUAL-SCORE / alternative-track signal — never folded into any Tier-1 exposure score. Research, not investment advice; ownership is traced through public listings and company announcements, and no per-shipment relabelling is asserted.

Verdict

On 25 February 2026 Zimbabwe's Ministry of Mines suspended exports of all raw minerals, including lithium-bearing spodumene concentrate — removing the concentrate carve-out that had survived the December-2022 raw-ore ban (SI 213/2022), and bringing forward by ~11 months a full concentrate-export ban that had been telegraphed for 1 January 2027. It sits on top of a fiscal escalator (Finance Act No. 7 of 2025: 10% VAT on unbeneficiated concentrate, 0% on lithium sulphate, 0% on carbonate, effective 1 Jan 2026) that makes raw export structurally uneconomic. Zimbabwe is the world's #5–6 lithium producer (~6–7% of global contained-LCE supply) and the largest in Africa, and its producer base is overwhelmingly Chinese-owned.

The volume-implausibility test that catches consumer-side leaks (a ~0%-capacity country surges under a phantom flag) is rejected by construction here: bulk spodumene concentrate cannot re-flag (the bauxite / tin / chromium rule), and Zimbabwe's landlocked exports move through a genuine logistics corridor (the ~1,000 km Gwanda→Beitbridge→Maputo rail route, activated 2026), not an origin-relabel. The real signal is mode B — capacity relocation under the controlling country's ownership — and Zimbabwe is its sharpest instance:

The "ban" is a soft ministerial letter, not a gazetted SI, and was restructured within five weeks (2 April 2026) into a producer-quota regime with a 10% interim levy — so unlike Indonesia's clean ore→0 severance, Zimbabwean spodumene kept flowing to China through 2026. The binding cut is the hard 1-Jan-2027 concentrate-export ban plus the fiscal pull toward sulphate. What the control actually does is convert the export line from raw concentrate to onshore-refined lithium sulphate — and that is where control launders into a new flag.

The corporate pipe (the captive-owner tell)

The strongest signal is not an implausible volume — it is who owns the new "Zimbabwe-origin" refinery. It is the same Chinese groups that own the mines and the Chinese converter capacity the concentrate used to feed:

Ownership and plant figures: company announcements (Huayou 20-Jul-2026 Arcadia commissioning; Sinomine Bikita), MINING.com, The Zimbabwe Independent, plus the public listings of each parent. Presented as the disclosed, worked example of the mechanism — a legal vertical-integration structure, NOT a hidden namesake shell and NOT a per-shipment accusation.

This is the pattern the nickel and bauxite cases proved on Indonesia, now on a third commodity and a third continent — but with a twist that makes it the purest "genuine origin masks control" case in the corpus. In Indonesia the hilirisasi ore ban forced foreign (Chinese) smelters to build capacity they had not previously owned onshore; the Indonesian state/SOEs at least took equity in some of it. In Zimbabwe the mines were already ~wholly Chinese-owned, and those same listed groups own the Sichuan/Jiangxi converters that the spodumene fed. So the "relocation" is a single vertically-integrated Chinese group moving its own midstream step from China to its own Zimbabwean mine site — the feedstock owner, the "lost" home-country converter, and the new host-country refiner are the same entity. The margin Zimbabwe "captures" from a Chinese miner is handed to the same Chinese miner's Zimbabwean subsidiary. Zimbabwe captures tax and jobs; it does not capture the asset or the molecule. Call it captive mode-B — beneficiation that produces a genuine new country-of-origin on a product whose economic controller never changed hands.

The sulphate conversion itself is a real industrial transformation (a mode-C overlay): it legally confers Zimbabwean origin and the 0% export-VAT treatment, so the molecules are truly transformed — the volume-implausibility test shifts from "who can produce this" to "who owns the transformation."

Why it matters for the buyer

1. A new "Zimbabwe origin" flag can be fully re-Sinicised one corporate layer down. A risk team that reads "our battery-grade lithium now comes from Zimbabwe, diversified away from China" is reading a country flag on a product that is Chinese-controlled end-to-end — mine (Huayou/Sinomine/Chengxin/Yahua), refinery (their onshore sulphate plants), and offtake (their own converter networks). The country-of-origin diversification is illusory; the corporate-ownership concentration is unchanged. Any exposure model keyed on stated country-of-origin is compromised the same way it is for a laundered consumer-side flag — the correction here is the ownership ledger, not an implausible-origin table. 2. Resource nationalism does not automatically de-concentrate control. The corpus's EM-upstream-capture thread (Indonesia nickel/bauxite, DRC cobalt) is usually read as producers clawing rents back from the dominant processor. Zimbabwe is the counter-lesson: when the dominant processor already owns the resource, a beneficiation mandate can transfer tax + employment to the host while leaving the strategic dependency intact and merely re-badged. 3. It confirms the chokepoint is binding — via price and the 2027 cliff, not a leak. Zimbabwe's ~15% of China's spodumene imports and the forced pivot to sulphate tighten the seaborne concentrate market (China lithium price rose on the suspension); the hard 1-Jan-2027 concentrate ban is the dated gate. As with tin and phosphate, the honest signal is the direct-line conversion + the price channel, read alongside the ownership tell — not a phantom flag.

Method & honesty rails

• Trade data: China customs (GACC) mirror for the Zimbabwe→China spodumene

share; Zimbabwe government / Chamber-of-Mines export-earnings statements for the value and post-ban continuation figures. The clean free primary for the mirror is UN Comtrade, reporter = China, HS 2530.90 (spodumene falls in "mineral substances nes") and the lithium-chemical lines (HS 2836.91 carbonate, 2825.20 oxide/hydroxide) as the sulphate ramp shows up. Late months lag; the soft-ban/quota transition means 2026 flow figures are provisional.

• Alternative-track only: this never touches buyerRelativeScore or the base

exposure — it sits beside them, exactly like the China–West price wedge.

• Inference, not accusation: mode-B relocation is evidenced by public

ownership + plant announcements, not asserted per shipment. The common-ownership tell is Y but DISCLOSED (listed parents + named Zimbabwean subsidiaries) — the palladium-Zug / palm-oil pattern, a legal vertical-integration structure, not a laundering shell. Mode-A relabelling is rejected (bulk concentrate can't re-flag; the Maputo corridor is genuine logistics).

• Gate: GATE 0 — (a) the binding stage is partly prospective (hard

export ban 1-Jan-2027; 2026 is a soft-quota transition with spodumene still flowing); (b) the volume-implausibility test is rejected by construction; (c) the ownership/relocation tell is real but disclosed, not a hidden pipe. All figures trace to named public sources; no per-shipment relabel is proven.

• Generalises: the detector's lesson — captive-owner beneficiation —

extends to any producer-side beneficiation mandate where the controlling country's firms already own the host's resource (watch DRC/Chinese cobalt and copper beneficiation, Indonesian Chinese-owned HPAL, and Zimbabwe's own chrome and PGM lines if the ban extends to them). The honest read: a beneficiation ban that leaves the mines foreign-owned re-badges the same owner's output with the host's flag — the diversification is on the label, not on the cap table.

Addendum 2026-09-25 (wake-circumvention, Namibia case): "no phantom flag" holds for concentrate only. On the earlier crushed-ore leg (after SI 213/2022, Dec-2022), China's HS 253090 imports labelled South Africa (no lithium mine output per USGS MCS 2025) rose 63,500 t (2022) → 972,832 t (2024); CRU (via Oxpeckers, 7-Apr-2025) ties the surge to Zimbabwean ore via Beitbridge→Durban. See 2023-lithium-namibia-ore-export-ban-circumvention-transshipment.md, Table 2.