Playing the wrong game

Tsogo Sun is the largest casino operator in South Africa, and that is exactly the problem. It runs about 15 casinos, holds roughly 39% of national casino gaming revenue, owns the bricks under Montecasino and Suncoast, throws off close to R3.5bn of adjusted EBITDA a year, and trades on four times that number. On the surface it looks like a cheap, cash-rich, property-backed cash cow. Underneath, it is the market leader of a business that is shrinking.

South African gambling has moved online. Betting, most of it on mobile, has become roughly 70% of national gross gambling revenue, while casinos have fallen to about 22% and are declining in real terms. Tsogo's casino gaming win peaked around FY2024 and has been slipping since. The company has an online arm, but it is late, sub-scale, spread across two brands, and it only offers sports betting. Its direct competitor, Sun International, built a proper online-casino product (SunBet) that now earns twelve times as much and is re-rating the whole company. Tsogo did not, and it isn't.

So the investment case is not "is it cheap." It is "is the market right that this business has peaked." We think it largely is. At Tsogo's own audited discount rate and a realistic, modestly-negative growth path, the company is worth about R8.00 a share. That is a touch above where it trades, but only a touch, and the downside is floored not far below by an asset base that management is itself impairing. It is a fairly-valued, structurally-challenged leader that has been out-executed on the one thing that matters. Hold.

The bull pitch writes itself. Tsogo Sun trades at about 4× EV/EBITDA, a P/E near 5, a dividend yield close to 6%, and it owns landmark property. It generates strong cash, it has cut net debt from R11bn at the depths of Covid to under R6.5bn, and it has started buying back stock. For a business with this cash generation, four times earnings sounds like a gift.

The catch is what four times is paying for. Run the arithmetic backwards: at a normal discount rate, the current price implies the market expects Tsogo's cash flows to shrink by roughly 1.5% a year in nominal terms, close to 6% a year in real terms, forever. That is not a pessimist's fantasy. It is a straight-line read of what betting is doing to casinos, and of a management team that has watched the shift happen and not responded to it.

The question is not whether Tsogo Sun is cheap. It is whether the market is right that the business has peaked.

Our verdict is Hold, base-case fair value about R8.00. The company is cheap for a reason, and the reason is real. There is a floor under the price from the property and the balance sheet, so this is not a short. But there is a better-run competitor in the same industry trading at a deserved premium, and if we wanted casino exposure we would rather own that. The one thing that would change the call is evidence that the new online team can actually build a competitive product. Until then, the base case is slow decline, correctly priced.

Tsogo Sun is really four businesses stacked on a property portfolio.

Casinos are the engine, about 87% of group EBITDA. About fifteen licensed casinos, led by Montecasino in Fourways, Suncoast on the Durban beachfront, Gold Reef City and Silverstar in Johannesburg, and a spread of regional properties. These are high-margin, cash-generative, and protected by a finite number of provincial casino licences. They are also mature. Gauteng, the biggest region and about 60% of casino EBITDA, is holding roughly flat. Everything else is eroding: KwaZulu-Natal fell 9% last year, the Western Cape and the smaller precincts are down mid-single digits. This is not one broken property that can be fixed. It is broad, gentle, structural decline with one resilient anchor.

Alternative gaming is the growth leg, about 16% of EBITDA. Vukani runs limited-payout machines in bars, restaurants and bookmaker sites, a genuinely resilient convenience-gaming business. Galaxy Bingo runs electronic bingo terminals, which are struggling under long leases and site-relocation red tape. LPMs grow slowly; bingo shrinks.

Online betting, run under the playTSOGO and bet.co.za brands, is tiny and, for now, a rounding error. More on it below, because it is the whole story.

Property. Tsogo owns its casino complexes rather than leasing them. On paper this is the safety net, and the land under Montecasino and Suncoast is genuinely valuable. In practice, as we will show, it is not the hidden treasure it first appears.

Above all of this sits Hosken Consolidated Investments (HCI), the controlling shareholder, run by Johnny Copelyn. HCI's own annual letter to shareholders is, refreshingly, one of the most honest assessments of Tsogo you will read anywhere, and we lean on it below.

Every argument about Tsogo Sun eventually comes back to one chart: how South Africans gamble.

Using National Gambling Board data, national gross gambling revenue for FY2025 breaks down roughly as below. Betting is about 70% of the total, of which online alone is close to 60%. Casinos are about 22%. Rewind a few years and casinos were the largest mode. Betting has more than tripled in three years and blew past casinos around FY2023. South Africa now has online-gambling penetration higher than either the United States or Australia.

The casino number is not just losing share, it is shrinking. National casino gaming revenue went from about R17.3bn in FY2023 to R16.6bn in FY2025, a nominal decline, and once you adjust for inflation it is down roughly 14% in real terms over two years. The operators are responding the way you would expect: the number of casino slot machines, tables and gambling positions in the country all fell last year. They are pulling capacity out of a declining market.

Why is this happening? Partly convenience and demographics, the same shift to mobile that has hit every consumer category. But there is a specifically South African accelerant: regulatory arbitrage. Online casino gaming is technically illegal here. Only betting is licensed, through provincial bookmaker licences. But bookmakers have used those licences to offer slots-like and casino-style online products that land-based casinos legally cannot, and they do it with lighter tax and far lower capital intensity. So the disruptors are not just more convenient. They are structurally advantaged on cost and product. A pending Remote Gambling Bill could legalise online casino properly, which would either let Tsogo compete on a level field or fragment the market further. Nobody knows which, and Parliament has parked it.

The honest way to hold this: casinos are not going to zero. They sell a night out, an experience, food and entertainment that a phone cannot fully replicate, and they keep a licence-protected niche. But the wallet is moving, the pie is shrinking in real terms, and the only question that matters is whether Tsogo has a hedge.

It does not, and the cleanest way to see it is to put Tsogo next to Sun International, because the two of them ran the same experiment and got opposite results.

Sun International leaned in. Its online arm, SunBet, grew from R77m of revenue in FY2018 to about R2,049m in FY2025, a twenty-six-fold increase, and now earns roughly R612m of EBITDA, about 18% of the whole group, still growing 60% to 75% a year with 766,000 active players. Crucially, SunBet is a full online-casino product: slots, live tables, sports and horse racing, on one brand tied to Sun's loyalty base.

Tsogo's online business earned about R50m of EBITDA off R313m of net gaming revenue. That is roughly one-twelfth of SunBet by profit. And the gap is not only size. Tsogo's platforms only offer sports betting. They are absent from online slots, which is the larger, higher-margin, higher-frequency segment, the part of the market that made Stake and the rest explode globally. And Tsogo runs two brands, playTSOGO and bet.co.za, which splits its marketing spend, its data and its player liquidity at exactly the moment when scale and customer-acquisition efficiency are the entire game.

Here is why that matters for the valuation rather than just the narrative. SunBet proves the disruption is beatable. A South African casino operator, starting from the same land-based base at the same time, built a real online business. So Tsogo's failure to do so is not bad luck or an impossible market. It is a management and execution verdict. The company has appointed new senior online leadership, which is the right move, but they are late, sub-scale, in the wrong product segment, and running two brands, against entrenched and better-funded competitors. Turning that around is possible. It is not the base case.

The market has already sorted the two. Sun International trades at about 5× EV/EBITDA and re-rated 35% during 2026. Tsogo trades at about 4×, roughly flat. That 20% discount is not a mispricing waiting to close. It is more likely the market correctly paying less for the company with no online hedge, double the leverage and half the return on equity.

Give management credit where it is due: the balance sheet has been repaired. Net interest-bearing debt fell from over R11bn during Covid to about R6.5bn, and to R6.3bn by April 2026. HCI has stated the group should carry no more than R6bn of debt, or 1.8× EBITDA, by FY2027, and is willing to cut the dividend to get there. Finance costs are falling as debt comes down. This is a company being run for the balance sheet, and prudently so.

The property is where the bull case usually plants its flag, and where it should be more careful. Tsogo's own accounts split property, plant and equipment into land and buildings of about R6.0bn net, and plant and equipment (the gaming machines and fit-out) of about R1.9bn, which is two-thirds depreciated. On top sits about R0.8bn of investment property carried at fair value.

Two things puncture the hidden-value story. First, the land and buildings are being written down, not up: net carrying value fell from R6.4bn in FY2023 to R6.0bn in FY2025, including a R166m impairment last year, because the casinos' value in use is falling. When a company impairs its own property, it is telling you the carrying figure is if anything optimistic, not conservative. Second, casino buildings are specialised, single-use assets. Their market value is essentially their value as casinos, which the operating cash flows already capture. You cannot both value the casino as a going concern and add the building again as a separate floor without double-counting. Only the underlying land in prime locations like Fourways and the Durban beachfront carries genuine independent value, and you cannot extract that without closing the casino on top of it.

Netted out, book equity is about R5.66 a share. A modest mark-up on the prime land lifts a realistic floor to perhaps R6.50 to R6.75. The current price of R7.57 already trades above book NAV. So the property caps the downside, but it does not make the stock cheap, and it is not the pot of gold the "buy the bricks, get the operations free" pitch implies.

HCI's control is the defining governance fact, and on balance it is a positive. Copelyn's letters are candid to a fault. On Tsogo's online problem, HCI wrote plainly that its "failure to develop a successful on-line offering to date has significantly threatened this position," and that "realistically it seems unlikely Tsogo will dominate the on-line space," hoping only that it "establishes itself as the provider of a medium-sized on-line offering over time." That is the controlling shareholder telling you the same thing this report is telling you.

Capital allocation is disciplined and shareholder-friendly at the margin. The company is deleveraging, it began a buyback in October 2025 (R438m, 62m shares, all cancelled), and it is pruning non-core assets: it sold its City Lodge stake, and it is disposing of its two smallest casinos, taking the estate from 15 to 13. Proceeds go to debt reduction and buybacks. This is sensible harvesting of a mature business.

Remuneration is moderately aligned but not best practice. The short-term incentive is tied to adjusted EBITDA (the CEO's award fell from 75% to 55% this year, and the total pool dropped 24%, so there is some pay-for-performance). The long-term incentive is a discretionary phantom-share appreciation scheme: cash-settled, no dilution, but struck at a 10% discount to the volume-weighted price and with no return-on-capital or relative-performance hurdles. It rewards absolute share-price gains off a discounted base rather than genuine outperformance. A mild flag, not a red one.

We value Tsogo three ways and they converge.

Sum-of-the-parts. Value casinos at about 4× to reflect the structural decline, LPMs at about 5.5× for their resilience, the online-and-bingo leg at about 4.5×, subtract the central-cost drag, subtract net debt, and add the small disposal proceeds. That gives an equity value of roughly R8.2bn, or about R8.50 a share.

The discount rate, and why a naive DCF gets this wrong

Discounted cash flow, done honestly. This is where a naive model goes badly wrong, so it is worth being precise. Group unlevered free cash flow is about R2.1bn. The two inputs that matter are the discount rate and the terminal growth rate, and both have to be right.

On the discount rate, Tsogo's own impairment accounts use a pre-tax rate of about 16.8%. Converted to the post-tax basis a DCF needs, that is roughly 13.5%. A textbook CAPM using the stock's measured beta gives a much lower figure, around 10.5%, but that is wrong here for two reasons: the plain formula omits the small-stock premium (about 3%) that applies to an illiquid mid-cap, and the measured beta is statistically meaningless (its regression against the market explains only 7% of the price movement, because Tsogo's risk is idiosyncratic and structural, not market-driven). The company's own 13.5% post-tax rate is the honest number.

On growth, the business has peaked, so terminal growth is not the +4% management uses or the +CPI a lazy model would assume. It is negative. A base case of about −1% nominal (roughly −6% real) reflects casinos shrinking while LPMs and online only partly offset. Put 13.5% and −1% together and the DCF returns about R8.55 a share, matching the sum-of-the-parts. At that discount rate, today's price of R7.57 implies roughly −1.5% terminal growth, so the market is pricing a fraction more decline than our base. That is the whole disagreement.

The bear is not open-ended: at roughly −3% decline you land near book NAV, so the business is worth its already-impaired asset value and no less. The bull requires the company to succeed at exactly what it has so far failed at, and to win a High Court fight with Sun International over the planned Somerset West relocation, on top of an online turnaround. It is a call option, not a base case, and its bright side comes with a re-rating kicker if it lands.

Weighing these, the expected value sits around R8, close to where the stock trades. The skew is mildly positive because the downside is floored and the upside is a genuine option, which is why this is a Hold and not a Sell. But an investor who wants South African casino exposure should ask why they would own the structurally-inferior operator when the better-run one, Sun International, is available at a deserved and defensible premium.

Verdict: Hold. Base-case fair value ~R8.00. Cheap for a reason, floored by its assets, and out-executed on the only game that is growing. The catalyst to watch is not in the numbers yet: it is whether Tsogo's new online team ships a single, competitive, full-casino product. If they do, this report gets rewritten. If they do not, the market's slow-decline verdict is the right one.