In this edition of Ghost Bites:
- African Rainbow Minerals to invest heavily in two South African projects
- Anglo American needs a strong second half in copper
- Cashbuild’s volumes aren’t telling an encouraging story
- Fairvest gets ready to gear up its fiber investment – literally
- Kumba’s full-year guidance is unchanged despite a dip in H1
- Mr Price: alarm bells are ringing for South African consumers
- Pepkor’s FintechCo would be a big listed company in its own right – in theory, at least
- Sirius acquires another defence-themed business park in Germany
African Rainbow Minerals to invest heavily in two South African projects (JSE: ARM)
Mining requires bravery when it comes to multi-year capex
African Rainbow Minerals has approved the Bokoni Platinum Mines project that comes with an estimated capex bill of R15.2 billion. That’s quite the show of faith in both this mine and the broader PGM story!
This comes after the completion of the Definitive Feasibility Study in June 2026. The idea is to achieve production of 180 thousand tonnes per month (ktpm), comprising 60ktpm from the existing concentrator and 120ktpm from a new concentrator.
The refurbished existing concentrator is expected to be commissioned in the first half of the 2028 financial year. The new concentrator will only be commissioned in the second half of 2030. Steady state across the project is expected to be reached in 2032.
The expected internal rate of return is 28%. Although this will ultimately depend on PGM prices, that’s an encouraging return that has some margin for error. It also helps that the capex bill will be spread over 7 years and that the refurbished concentrator will be a positive contributor to cash flows while the new concentrator is built. If you’ve ever built a cash flow model, you’ll know how important the timing of cash flows is.
African Rainbow Minerals believes that the project can be funded mainly by existing cash resources and profits generated over time, with external debt funding “to the extent required”.
The company has also approved the recommencement of open-pit mining and nickel concentrate production at Nkomati Nickel Mine. This is a far more modest capex bill (only R753 million) with an expected IRR of 28.36%. It’s interesting to note the similar percentage returns of these two projects.
Ghost Bite: Mining requires brave application of capex. Despite the share price having lost 21% of its value over 12 months (and now trading close to 52-week lows), the company needs to commit to through-the-cycle investment.
Anglo American needs a strong second half in copper (JSE: AGL)
This is the metal that everyone is talking about
Anglo American has released a production report for the quarter ended June 2026. Before I carry on, please note that there’s a similar update on Kumba Iron Ore (JSE: KIO), a subsidiary of Anglo American, further down in Ghost Bites.
As you’re probably expecting, copper is still the belle of the ball in the mining sector. With Anglo expecting the copper-driven merger with Teck to be completed by March 2027, this is the commodity that everyone is watching.
Anglo delivered increased copper production at Collahuasi and Quellaveco on a quarter-on-quarter basis. The restart of the second plant at Los Bronces was also a positive contributor. But if you look on a year-on-year basis, copper production is perfectly flat. Guidance for 2026 is unchanged and weighted towards the second half, so there’s significant execution risk that Anglo will need to make sure they manage.
In iron ore, Anglo describes Kumba as a “stable” performance (probably a fair take). The same language is applied to Minas-Rio. Although premium iron ore production increased by 1% quarter-on-quarter, the year-on-year number is a decrease of 3%. Production guidance is unchanged for the year, while sales at Kumba will depend significantly on Transnet’s performance.
In manganese ore, production was up 20% quarter-on-quarter and 22% year-on-year. The previous year was impacted by the knock-on effects of a tropical cyclone in March 2024.
That takes us to the end of the list of commodities that Anglo plans to keep. We now move into the businesses they are getting out of as part of the broader corporate simplification.
In May, they agreed to sell their Steelmaking Coal business in Australia to Dhilmar for up to $3.875 billion in cash. They expect to complete this deal by the first quarter of 2027. Steelmaking coal production jumped by 32% quarter-on-quarter, but was down 1% year-on-year.
As for De Beers, I’ve seen a number of news headlines that Anglo is looking to sell to a consortium led by Gareth Penny, an ex-CEO of De Beers. Diamonds being bought by a Penny is nominative determinism of the highest order – I don’t think Anglo will get much for it.
Diamond production increased by 9% quarter-on-quarter and 88% year-on-year. Although Anglo’s official line is that this is due to the timing of maintenance and the grade of ore being mined, you’ll have to forgive my cynicism here. When your premium pricing model is dying (H1 prices fell 32%), you need to ramp up the volumes. Production guidance for the full year is unchanged.
Finally, we deal with the nickel business. Anglo has agreed to sell this to MMG Singapore Resources, with the deal currently going through European competition approval processes. Production dipped 4% quarter-on-quarter and 6% year-on-year.
Ghost Bite: Anglo’s share price has jumped more than 50% over 12 months. You can thank copper for this.
Cashbuild’s volumes aren’t telling an encouraging story (JSE: CSB)
South African consumers aren’t spending on their properties
Cashbuild, a local company in which I have a stake that I should’ve sold, has released a voluntary fourth quarter update.
Although revenue was up 6% for the quarter, the existing stores (defined as stores before July 2024) only managed 1% growth. The new stores contributed 5% to growth. This isn’t a particularly encouraging outcome in terms of the underlying strength of the business.
The quarterly growth is consistent with the full financial year, which also grew by 6%.
Another lens you can use at Cashbuild is comparable store revenue, which excludes the impact of mergers and acquisitions and store closures. With this approach, you’ll find growth of 3% for both Q4 and the financial year as well.
Inflation was light, coming in at just 1.5% at the end of June 2026 vs. June 2025. Existing stores suffered a decline in volumes. That’s concerning during a period of modest inflation and supposedly improving sentiment in South Africa. Thank goodness the SARB didn’t increase rates last week!
Cashbuild has been an unfortunate story where I absolutely should’ve taken profit at the end of 2024 thanks to the GNU exuberance. I thought I would stick to my knitting and hold it as a play on SA Inc and things getting better here. Well, the joke is on me, with the share price now at R117 vs. the late 2024 peak of over R227. If you’re keen to see that chart, I covered it in this YouTube video (from around the 2:20 mark).
Ghost Bite: Buy-and-hold isn’t always the smart idea that people would like you to believe. I am much better at buying market weakness than I am at selling market exuberance.
Fairvest gets ready to gear up its fiber investment – literally (JSE: FTA)
They are serious about the township fiber opportunity
Fairvest is more than just a traditional property company. The group has an unusual investment in the form of Onepath Investments, a fiber infrastructure company. To help the market understand more about this opportunity, Fairvest delivered an investor presentation focusing exclusively on Onepath.
Onepath is the “landlord” in this situation, with fibertime as the tenant. There’s a separate company (Refiber Digital Infrastructure) that acts as the capital and asset manager. By using this clever analogy, Fairvest lands the point that owning fiber network infrastructure might not be such a big strategic departure from owning property.
Of course, the real play here is to achieve connectivity for township users, as Fairvest holds a number of township-adjacent malls. Having additional data on the users in these areas could help them make better property investments.
My understanding is that fibertime’s model is based on a cost of R5 a day and a true pay-as-you-go model. In FY25, they had 285k users. By FY27, they are targeting 3.05 million users! This works out to an average of 3.3 users per home by FY27.
To achieve this, cumulative capex by FY27 would be R4.8 billion. The longer-term goal (2030) is to have 20 million users and 5 million homes, delivered with cumulative capex of R24.3 billion. As growth stories go, that’s an exciting one.
If you can believe it, the technology partner is Nokia. I’m surprised to see in one of the pictures in the presentation that the routers need battery backups. We all know that there are still 3310 owners out there who haven’t charged their phones since 2002!
The net yield on capital deployed is sitting at at 15.1% on an ungeared basis. Admittedly, this is a different risk profile to traditional property ownership (retail malls etc. yield high single digits). Still, this seems like a really attractive return. Perhaps most importantly, it’s high enough to easily be able to service any related debt. Fairvest plans to introduce gearing in the coming months.
Ghost Bite: REITs are yield-focused companies. When you’re getting mid-teens on an ungeared basis, the introduction of gearing can leverage this up to really juicy geared yields (as the cost of debt is way below the ungeared yield). This could be one to watch!
Kumba’s full-year guidance is unchanged despite a dip in H1 (JSE: KIO)
This is one of the toughest business models in the country
Kumba Iron Ore released a production and sales report for the six months to 30 June 2026. This gives investors important clues about the interim financial performance and how the company is tracking against full-year targets.
It’s unfortunate that production was down by 3% year-on-year thanks to a 16% drop at Kolomela. Although Sishen is a much larger mine, a 3% increase at that mine still wasn’t enough to offset the downward pressure at Kolomela. This directly impacts the level of on-mine stock, which has decreased from 5.7 Mt in December 2025 to 4.8 Mt at the end of June 2026.
Sales volumes fell by 1% during a period that included planned maintenance by Transnet. Cleverly, Kumba planned maintenance at its own mines to coincide with the Transnet shutdown. Transnet’s ability to achieve export throughput on behalf of the mines is the fact that impacts stock held at Saldanha Bay Port, which increased from 1.8 Mt to 2.2 Mt.
Despite the pressure in H1, the company feels good about still delivering full-year production and sales guidance. This tells you a lot about the resilience that gets baked into their annual guidance!
Cost guidance has been let unchanged for Sishen and Kolomela, but the company has noted upward inflationary pressures. This is pushing costs towards the upper end of the range at Sishen and the middle of the range at Kolomela.
Separately, Kumba announced a 20-year energy offtake agreement with Envusa Energy for the on-site supply of electricity to Sishen. Envusa Energy is a joint venture between EDF Power Solutions and Anglo American (JSE: AGL), Kumba’s controlling shareholder. Together with existing projects, this will take Kumba’s renewable energy penetration to around 45%.
Ghost Bite: This is a very hard business to run. Before we even consider the volatility of commodity prices, Kumba also needs to navigate infrastructure challenges just to get the stuff to port. Kumba’s total return is -6% over 12 months, -26% over 3 years and -36% over 5 years. Ouch.
Mr Price: alarm bells are ringing for South African consumers (JSE: MRP)
There are a number of worrying signs here
Mr Price has given the market a voluntary trading update for the 13 weeks ended 27 June 2026. Group sales were up 45.3%, but that’s obviously because of the recent acquisitive activity at NKD. Believe me, if Mr Price was growing at that rate organically, money would be falling out of the sky in South Africa. As we learnt in Cashbuild further up, that’s certainly not the case.
So, the first thing we need to do is strip out NKD, which then gives us growth of 3.2%. That’s more in line with what we would expect to see. It’s still a decent outcome, ahead of Mr Price’s quoted market growth of 0.8%. This implies that they’ve been winning market share.
But then we get to the bad news – comparable store sales were flat. This tells us that all the growth came from new stores in South Africa. The store footprint increased by a net 32 stores, with trading space up 3.8% on an annual weighted basis. That’s better than no growth at all, but it paints a worrying picture for the South African consumer.
Another indication that all isn’t well is that online sales were up 4.7%, while total store sales were up 3.1%. Mr Price puts minimal emphasis on online sales, as they focus more on being a bricks-and-mortar business. To see online outperforming store sales is a surprise.
Mr Price’s focus is on cash sales, which seem to be hard to get right now. Cash sales were up 3.1% and credit sales were up 3.8%. This is another indication of a struggling local consumer.
Looking at product categories, Homeware is another flashing red alarm for South African consumers. Growth was just 0.7%, with comparable sales down 3.3% and volumes up just 0.2%. Yuppiechef was the star of the show as usual, with double-digit sales growth and improved gross margin. The wealthy still have money here, but nobody else seems to.
Apparel, the anchor of the group with a 78.8% contribution, grew by 3.4%. Comparable store sales increased by 0.6%, so at least they were positive here.
Telecoms, the smallest contributor at just 3.9%, grew by 11.2% – this is becoming a more important area over time.
On the plus side, the group protected gross margin over this period, with margin up 40bps during a time in which competitors were highly promotional. South Africa is described as having a clean inventory position, which implies that gross margins aren’t under immediate threat.
Then we get to Europe, where NKD outperformed the relevant market benchmarks (the total apparel market and the value segment in Germany). 21 stores were closed and 23 opened, so there’s only a tiny increase in the footprint to 2,156 stores. NKD is also described as having a clean inventory position. This is about as much as Mr Price will tell us at the moment, so we need to wait for more detailed reports.
Ghost Bite: When value-focused fashion houses are struggling like this, while the top layer of South Africans continue to buy Yuppiechef like ice creams on a hot day, then you really have to ask hard questions about our interest rates. I’m very glad that the SARB didn’t hike.
Pepkor’s FintechCo would be a big listed company in its own right – in theory, at least (JSE: PPH)
As a shareholder, I like this deal
By now you know this news, but I’m including a note on Pepkor for the sake of completeness in this catch-up edition of Ghost Bites. It’s also worth reminding you that I bought shares in Pepkor a couple of months ago based on their underlying business and the upside optionality of the bank they are building. It’s a nice surprise to see even more momentum in the fintech business than I expected!
Pepkor is combining its Flash business with Shop2Shop to create “FintechCo”, a R21.3 billion business that Pepkor will have 57.1% in. This is calculated based on the value of Flash (R10.6 billion) and a cash subscription by Pepkor for new shares in FintechCo to the value of R1.57 billion.
As I pointed out on social media at the time of the deal, FintechCo would therefore be worth more than Truworths (JSE: TRU) or The Foschini Group (JSE: TFG). You have to be careful when comparing listed companies to what is essentially just a directors’ valuation, but the point is hopefully still made.
What will the new fintech do? The fintech ecosystem is about as complicated as things get, but effectively they will have extensive participation across the value chain linked to the informal market. This aligns beautifully with Pepkor’s value-focused strategy.
In terms of financials, we only have outdated full-year numbers to work with (September 2025 for Flash and June 2025 for Shop2Shop). The difference in growth rates is staggering though, with Shop2Shop having achieved a three-year revenue compound annual growth rate (CAGR) of 28% vs. 9% at Flash. In terms of profit after tax, Flash achieved R488 million and Shop2Shop was R385 million.
Aside from questions around the relative valuation, the market has also raised concerns around the conflict of interest in the deal. CEO Pieter Erasmus has a stake in Shop2Shop that predates his appointment at Pepkor. Although it’s not a related party deal under a technical application of the JSE rules, the company did the right thing by excluding Erasmus from all the deliberations at board level.
My view on this? Whilst conflicts of interest need to be carefully managed, I would be far more worried if this looked like an “odd” deal, or if a vast amount of cash was changing hands. In practice, this deal makes a world of sense for Pepkor through a strategic lens. It’s also a merger where most of the value is on paper rather than in cash, so that gives me further comfort.
Ghost Bite: There’s an intention to list FintechCo down the line, so there’s now another value unlock opportunity brewing inside Pepkor. As a shareholder, I like that. The share price is trading close to 52-week lows, so I’m very tempted to add to my current position.
Sirius acquires another defence-themed business park in Germany (JSE: SRE)
The execution of this strategy continues
Sirius Real Estate is certainly consistent when it comes to their acquisition strategy. They stick to assets in the UK and Germany, with the latter generally having a defence industry flavour.
The latest acquisition is a light-industrial business park in Fulda, north east of Frankfurt. They are paying €49.8 million for this asset based on an EPRA net initial yield of 7.8%.
The anchor tenant is a manufacturer of ballistic protection equipment. This is a good opportunity to remind you that these defence properties aren’t always highly specialised – it often comes down to other tenants in the area, or proximity to supply chains. As an analogy, think about how financial services firms tend to huddle together in a certain area.
The weighted average lease expiry is 5.1 years, so there doesn’t seem to be an immediate opportunity for Sirius to work some magic on the yield. Not everything they buy is a fixer-upper.
Ghost Bite: If you would like to understand the Sirius strategy in more detail, this podcast with the top execs at the end of 2025 is just as relevant today as it was then.
Nibbles:
- Director dealings:
- Unsurprisingly, various Datatec (JSE: DTC) directors used the scrip distribution alternative as a way to get their hands on a further R103 million worth of shares. This is an opportune time to remind you that the CEO still has a huge stake in the company that he founded.
- The CFO of Lewis Group (JSE: LEW) has sold shares worth over R3.1 million. Although this sale is to “rebalance his portfolio”, a sale is a sale.
- A director of Santova (JSE: SNV) sold shares worth R400k.
- A director of Trematon (JSE: TMT) bought shares worth R82.7k.
- A non-executive director of Finbond (JSE: FGL) bought shares worth R32k.
- A few Hudaco (JSE: HDC) directors received shares based on the automatic exercise of share options. If my understanding is correct, there was also one director who chose to exercise the options and then sold the entire amount for R23.5k.
- Thanks to the company’s previous announcement about the underlying fund, we already knew that Reinet (JSE: RNI) wouldn’t be telling an exciting story around NAV growth in the latest quarter. We now have the numbers for the holding company, which confirm that NAV per share increased by just 1% over the past three months. This was thanks to share buybacks. For a more detailed look at Reinet, you can check out what I wrote when they gave the update on the underlying fund.
- Vukile Property Fund (JSE: VKE) announced that Global Credit Ratings (GCR) has affirmed its credit ratings with a stable outlook. This is very important for a REIT, as obtaining well-priced debt is a key factor in achieving solid shareholder returns. The company also announced that Dr Renosi Mokate will step down as Lead Independent Director in September, to be replaced by James Formby. As a final update on Vukile, the company is also busy with a debt capital markets roadshow – a critical source of finance for property funds. The presentation for the roadshow is a helpful overview of the group.
- The final step in the succession plan at Dis-Chem (JSE: DCP) is upon us. After founding the business five decades ago, Ivan Saltzman is now retiring from the board with immediate effect. This is hot on the heels of the news of his son, Saul Saltzman, also resigning from the board. I can’t help but wonder if this decision was accelerated by the recent bad press around certain social media posts made by a different member of the Saltzman family. Companies with strong ties to its founders can be vulnerable to the “social outrage” that is a feature of the modern world. Either way, after 48 years with the company, Saltzman Senior has certainly earned his retirement.
- NEPI Rockcastle (JSE: NRP) has announced that they will host a capital markets day on 21 October 2026. This is well worth diarising, as the day will provide deep insights into both the portfolio and the broader property market in the Central and Eastern European region.
- If you are invested in ASP Isotopes (JSE: ISO) and you want to understand more about the helium assets in the group and the planned deal with Noble Africa, then the company has made the transcript available from the recent investor day.
- Mustek (JSE: MST) announced that Rectron, a wholly-owned subsidiary of the group, has suffered a cybersecurity attack. They became aware of it on 15 July and immediately followed the process around incident response and business continuity. As this stage, they haven’t given any indication of the scope and extent of the incident.
- I’m not terribly surprised that Sappi (JSE: SAP) shareholders voted strongly in favour of the proposed joint venture between Sappi and UPM in Europe. Only 1.42% of votes were cast against this transaction. I genuinely cannot imagine why anyone would vote against it, as it’s not like Sappi as many other great options available right now. We are talking about a share price that has lost 55% of its value this year!
- Shuka Minerals (JSE: SKA) announced that it has raised gross proceeds of £750k through a subscription for new shares at 4 pence per share – a 53.9% premium to the closing price on 21 July. The subscriber is Menel Energy and Resources, a company that has various mining projects in Zambia. The first tranche of £375k has already been received, with the second tranche expected to be received by 31 August 2026. A warrant will be granted to Menel to subscribe for up to 18.75 million new shares at 8 pence per share, exercisable until 8 July 2029. Remember, the longer the time period of an option, the more valuable it becomes. Separately, the company announced that it will issue 375,000 new ordinary shares at 4 pence per share in lieu of accrued fees owed to a former director.
- Those keeping an eye on the share register at Labat Africa (JSE: LAB) will be interested to know that Muziwakhe Ibrahim Ndhlovu now has 28.41% of the shares in the company. This is related to the recent acquisitive activity at Labat.
- Insimbi Industrial (JSE: ISB) has added some strong investment banking skills to its board with the appointment of Dean Crommelin as an independent non-executive director. It’s always interesting when small caps make appointments like these.
- After a very rocky road, Kibo Energy (JSE: KBO) has found itself without a market for its shares. The listing on the AIM in London will be cancelled with effect on 27 July, as the company has failed to implement a reverse takeover transaction. The listing on the JSE is currently suspended. With the primary listing being cancelled, it’s hard to see how the JSE listing won’t follow suit. This leaves shareholders without any way to trade their shares publicly, although they do still own their shares in the company. If Kibo ever manages to put together a deal, shareholders may still get something out of this. But there’s also a good chance that it all just fades away.
