After years of platinum miners cutting back, African Rainbow Minerals just signed off roughly R16 billion to reopen a mothballed platinum mine and restart a nickel mine. Counter-cyclical genius, or a seven-year bet on prices nobody controls?
By The Money Maze | Listen to the episode: Spotify
There’s a type of oke at every braai who buys a bakkie the week everyone else is selling theirs. Either he knows something, or he’s about to spend a lot of weekends under the bonnet.
On Thursday 23 July 2026, African Rainbow Minerals, Patrice Motsepe’s diversified miner, became that oke. Its board approved two projects on SENS: the Bokoni platinum development in Limpopo, at R15.2 billion, and the restart of the Nkomati nickel mine in Mpumalanga, at R753 million. Together, call it R16 billion.
So let’s look under the bonnet.
What ARM actually approved at Bokoni
ARM bought Bokoni in September 2022. It’s a big, high-grade platinum group metals orebody on the north-eastern limb of the Bushveld, and ARM says it’s the second-largest PGM resource base in South Africa.
The first plan was a big, fully mechanised mine. Then platinum prices fell over, costs went up, and at the end of the 2025 financial year ARM suspended mining and milling at Bokoni while it went back to the drawing board. That’s the mothballing. The mine wasn’t dead. It was on the couch with the curtains closed.
The new plan is phased. Step one: refurbish the existing 60,000 tonne-a-month concentrator and start producing in the first half of ARM’s 2028 financial year. Step two: build a new 120,000 tonne-a-month concentrator, commissioned in the second half of 2030. Total capacity: 180,000 tonnes a month, or about 350,000 to 400,000 PGM ounces a year at steady state, expected in 2032.
The R15.2 billion includes a 15% contingency and gets spent over seven years, from 2027 to 2033.
What a brownfield restart means
Jargon of the day: brownfield.
A greenfield project starts from veld. Nothing there but grass, a few cattle and a consultant’s PowerPoint. A brownfield project starts with stuff already in the ground: shafts, declines, a plant, roads, power lines.
Think of it as the difference between building a house and renovating your gran’s place in Benoni. The renovation still costs a fortune and you will find something horrible behind the kitchen cupboards. But at least the plumbing is mostly there.
ARM leans hard on this. It says Bokoni’s existing declines, concentrator and surface infrastructure “materially de-risk” the project, and that starting with the old plant first reduces execution risk. Translation: we’re not betting the farm on day one.
Nkomati: the smaller, quicker restart
Nkomati is South Africa’s only primary nickel producer, and it has been on care and maintenance. ARM looked at keeping it in mothballs and decided to restart the open pit instead.
The numbers are tidy. Capex of R753 million, a 13-year life of mine, payback in about 5.3 years and around R616 million a year of free cash flow at steady state. Plant refurbishment started in July 2026 and mining was scheduled to restart in October. It also comes with an off-take agreement with Swedish miner Boliden, which buys the concentrate, though that deal still had some conditions to clear.
NPV and IRR, in braai terms
The SENS throws two acronyms at you, so let’s deal with them.
NPV, net present value, is what all the project’s future cash is worth in today’s money, after paying for the build. Bokoni’s NPV is R5.9 billion. IRR, internal rate of return, is the yearly return the project is expected to earn on the money put in. Bokoni’s is 28%, with payback in 6.3 years.
Here’s the catch. Those numbers assume a long-term PGM basket price of R31,161 an ounce in real terms. Change the price assumption and the NPV moves like a cheap gazebo in a Cape Town southeaster.
How ARM plans to pay for it
ARM says Bokoni will be funded mainly by ARM itself, by cash from Bokoni once the first plant is running, and by external debt “to the extent required”. On the investor call, management said only about R2.3 billion is spent before first revenue, and peak funding drops to about R10.4 billion because early production pays for part of the rest.
There’s no mention of asking shareholders for more money. For anyone holding ARI on EasyEquities, that’s the line you read twice.
ARM bull case
Brownfield, not greenfield, so a lot of the expensive stuff already exists. A huge, high-grade orebody, with only about 13% of the measured and indicated resource used up in the 19-year plan. A phased build, where the first plant helps pay for the second. ARM expects Bokoni to sit below the middle of the global PGM cost curve.
And the timing argument: ARM reckons years of underinvestment and ageing shafts will push platinum into deficit. If you build while everyone else is closing, you might be the one selling when prices recover.
ARM bear case
Seven years is a long time to bet on a metal price. The whole case rests on PGM prices that ARM doesn’t control, and on electric cars not eating into platinum demand faster than expected. ARM admits battery vehicles are a factor.
Mining projects have a habit of costing more and taking longer, which is why the 15% contingency exists. Debt may be needed. And the Bokoni plan has already been redesigned once, after the first version met real-world prices.
Neither case says buy or sell. Both say: with a mine restart, the price deck matters more than the press release.
Key numbers
- Bokoni capex: R15.2bn over seven years. So what: about R2.3bn before first revenue, peak funding about R10.4bn.
- Bokoni NPV and IRR: R5.9bn and 28%. So what: based on a R31,161 real basket price assumption.
- Nkomati capex: R753m. So what: payback in about 5.3 years, around R616m a year of free cash flow at steady state.
- Steady-state output: 350,000 to 400,000 PGM ounces a year from 2032. So what: a meaningful lift to ARM’s platinum business.
Takeaway: Buying when everyone is selling only looks clever later. Until then it looks exactly like spending R16 billion.
Want the full braai version? Listen on Spotify. And if you’d like the Maze in your inbox, join the free Maze Mail list.
More from The Money Maze
- Kumba Iron Ore (JSE: KIO): Why the Dividend Fell to R7.90
- Harmony Gold’s R4.4bn Payout: 11 Years of Guidance Met
- Anglo’s Diamond Meltdown, or Just a Very Expensive Mood Swing?
- Mantengu’s R315 Million Loss and the Dubai Takeover
Sources
- Board approval of the Bokoni development project and Nkomati operational restart, African Rainbow Minerals on SENS, 23 Jul 2026.
- ARM board approves restart of Bokoni platinum, Nkomati nickel projects, Mining Weekly, 23 Jul 2026.
- Investor conference call: Bokoni development project and Nkomati operational restart, ARM, 31 Jul 2026.
- ARM investor update transcript, StockAnalysis, Jul 2026.
This is not financial advice.