Gemfields Wrote Down Its Ruby Mine by $125 Million — Then Restated Last Year's Numbers Too

Gemfields Wrote Down Its Ruby Mine by $125 Million — Then Restated Last Year's Numbers Too

For a company whose core assets are literal rocks, Gemfields' interim results for the six months to 30 June 2026 read less like a mining report and more like an accounting exorcism. The headline numbers looked almost respectable — $106 million of revenue, $40.7 million of EBITDA, a $17.4 million free cash inflow and adjusted EPS of $0.06 — but those were flattered by an auction-calendar quirk, with the deferred December 2025 ruby sale dropping into February 2026 and giving MRM two auctions in the half instead of the usual one. Peel that away and you are left with a $125.2 million non-cash impairment of the Montepuez ruby mine and a restatement of the prior year's accounts. David Lovett, the interim CEO-cum-CFO, did not pretend otherwise: “It's fair to say that the first half was difficult and the reported loss is clearly disappointing, principally due to the challenges at MRM.” — David Lovett, Interim CEO and CFO · 2026-09-30

The impairment drivers were threefold — one of the wettest rainy seasons on record, a PP2 processing ramp that ran below the 400 tonnes-per-hour target for most of the half, and, most damningly, premium ruby recoveries that came in "significantly below operational targets." Premium ruby output fell 13% year-on-year even as ore produced rose 133%. Size is not the problem at MRM; grade is.

The restatement is the real red flag

An impairment is a judgment call; a restatement is an admission. Buried in Becki Tate's section was the disclosure that the group had revisited its life-of-mine model and found errors that also existed in the December 2025 model.

The restatement arose following the identification of some inconsistencies while reviewing the 30th of June 2026 life-of-mine model that were also found to exist in the 31st of December 2025 model upon review.

The practical effect: the prior-year impairment charge was lifted from $35 million to $65 million, and MRM's carrying value is now down to $80.3 million. This is the kind of disclosure that makes investors wonder what else in the model had been leaning optimistic — and it lands in the same half that a new interim CEO took the helm after Sean Gilbertson's departure, which Lovett would only describe as "a mutual decision between Sean and the Board."

The grade math behind a going-concern warning

The call was peppered with a phrase that never appeared as a headline but framed everything: material uncertainty. Asked what grade would let Gemfields escape it, Lovett gave the number that matters.

If we could get the grade up above 0.03, that would certainly be a much more positive financial model, which would help in terms of our going concern.

Here is the squeeze. Grades are currently running around 0.025 premium carats per tonne at Mugloto and roughly 0.04 at Maninge Nice — but Maninge Nice stones fetch about 20% less per carat, so a blended grade north of 0.03 is not a given. Cash at the centre sits at $34 million against a stated $5 million minimum requirement, and net debt stands at $44 million before auction receivables. On whether MRM can yet cover its own management and auction fees, Lovett did not reach for a euphemism: “In short, no.” — Heinrich Richter, Moderator / Investor Relations · 2026-09-30 The company still expects MRM to service its debt next year, but refinancing talks are, in its own words, "very early stage."

What management quietly stopped saying

The most telling keyword shift is not something management introduced — it is something they retired. Maninge Nice was the top-ranked theme two quarters ago, when analysts pressed repeatedly on whether the company would normalize it against Mugloto. This quarter it has tumbled to a footnote, and the question has effectively been answered by redefinition: secondary sapphires are now excluded from the premium ruby characterization, and history has not been restated "because there is, quite frankly, not sufficient detail to recalculate those categories." That is a cleaner number, but it also shrinks the denominator of what counts as premium ruby — worth remembering when the recovery charts are read.

Two other themes have gone quiet. Contract mining was raised twice and answered with a hedge — a "theoretical decision" by year-end, no contractor instructed. And Nairoto, the gold asset Sean Gilbertson flatly declared dead in March (“the reality is Nairoto, therefore, is not a project that we're going to bring back into business” — Sean Gilbertson, CEO · 2026-03-26), is now a live "sale process" with "no concrete sale options." The one theme management wants you to hear is competitive insulation: on rubies there are "no big players" out of Mozambique, while emeralds face a Zambian neighbour and Brazilian and Colombian supply. That big players framing is doing a lot of work holding up the long-term thesis.

The one genuinely improving thread

Not everything is a write-down. Kagem, the Zambian emerald mine, is the quiet hero: roughly 125,000 carats of premium emerald to August, ahead of most prior years, plus about 7.8 million carats of the higher-volume emerald grade. A September commercial-quality emerald auction already secured almost $30 million, three more auctions are scheduled for 2026, and the first Trade Select auction — a new format to widen the mix and push sapphires — brought in $23 million. At MRM, PP2 has run consistently between 410 and 450 tonnes per hour since mid-June, lifting combined annualized throughput to roughly 3 to 3.2 million tonnes; July through September were described as "healthier." Management is responding with the three words it wants investors to internalize — operational stability, cost control and balance sheet resilience — plus expanded bulk sampling to rebuild grade confidence before a 2027 resource update.

The isolation trade

What makes Gemfields an oddity today is how completely it sits outside the market's dominant narrative. Global keyword flows are saturated with AI data centres, memory pricing, co-packaged optics and Bitcoin miners — not one gemstone or luxury-commodity theme registers. Gemfields is not riding a wave, and no wave is going to save it. This is an idiosyncratic turnaround, and the price of the equity now hinges on two things the company cannot fully control: whether bulk sampling finally stabilizes premium ruby grade above the 0.03 line Lovett named, and whether Mozambique's tax authority keeps dribbling VAT refunds back after the first cash receipt since October 2024. Everything else — the cost discipline, the auction cadence, the Kagem strength — is support. The model, and now the accounts themselves, have already proven they can be wrong once.