Independent deep-dives on JSE-listed companies.

A bed-licence-moated, cash-generative SA hospital operator at an all-time nominal low, carrying the same NHI risk as a peer that has re-rated hard. The de-rating is a fixable, company-specific margin problem the market has priced as permanent decline.

A lease-honest sum-of-the-parts puts a high-quality retailer at roughly fair value — richly priced for the quality at the demanding end of a defensible discount rate, with the NKD/Europe optionality coming cheaply. A good company, fully priced.

A cyclical business with a defensive core, fairly valued after a full round-trip from R67 to R49 and back. The upside from here is a specific bet on a Gulf-of-Mexico fish; the downside is a structural decline in South Africa's local resource.

South Africa's largest casino operator, leading a land-based market that online betting is eating alive and out-executed on online by its direct peer, Sun International. Cheap for a reason, floored by its assets.

A world-class dissolving-pulp franchise and a ramping packaging pivot, being separated from a dying European graphic-paper business (the UPM JV), through a deep cyclical trough that is now bottoming — a cheap, asymmetric, levered bet on the DWP recovery, refereed by a March-2027 covenant test.

Four businesses plus a credit book, de-rated ~70% to ~R54. The fall is mostly a swing to a trough multiple on a cyclical earnings dip, with one contained structural write-down (Phase Eight). The SA core is a quality franchise on trough earnings, the balance sheet is safe, and the market prices the offshore at roughly zero. Conservative probability-weighted fair value ~R66.

A dual-engine software business: a durable ~97%-margin Check Verification annuity funding a ~20%-growing Fraud & Identity SaaS engine on the AI-fraud wave, with a genuine data-network moat (Check Fraud Defender consortium — a top-5 US bank + Fiserv reseller, ~70% of US checking accounts, ACV +73%). The 3Q26 print was strong and RAISED guidance ($195-200m rev / F&I $105-109m / 32-34% margin), materially de-risking the bear. We value it DCF-led (owner cash, not multiples): normalised owner FCF ~$32m with an operating-leverage inflection (margin ~15%->~22% as F&I scales) grounds base fair value ~$14; the SOTP (~$17) is the market-priced upper lens. At ~$19 (up from $16.81 pre-print) the price has run through both — the great print delivered and the re-rate happened. HOLD; accumulate only on a pullback to the cash-grounded ~$14 zone; trim into $20-22.

A maturing, ultra-cash-generative niche subscription priced for durable growth it hasn't proven. The paying base (7.6M Active CLEAR+ Members) grows ~6%; the "38M members" is non-paying; retention has declined for years and management has now discontinued the metric; and the Q1'26 "acceleration" was a one-time TSA-shutdown beat against a decelerating pre-shutdown guide. Real owner FCF (~$304m) + a fortress balance sheet make the floor solid, but the up-C leaks value (TRA ~$252m + tax distributions) on ~133M shares under an existential-if-gradual TSA-CAT tail. At ~$52.63 it trades ~18–20% above our conservative fair value (~$42–43). Hold / watch-list.

A cheap-on-FCF, net-cash, sticky bank-security specialist (>60% of the world's 100 largest banks) whose activist-era margin turnaround is DONE — it peaked FY24–25 and is now being deliberately spent back down on growth reinvestment (FY26 adj-EBITDA guided down to $64–68m from $77.6m). Organic growth is only ~7–8% ARR (the headline +14% is mostly M&A), and management is deploying essentially the entire net-cash floor into mobile-security M&A (Nok Nok / ThreatFabric / Build38). Reported EPS $1.88 is a tax mirage (~$33m Canada valuation-allowance release; normalized ~$1.00) — value on FCF (~$50m; owner FCF ~$39m), not GAAP EPS. Owner-FCF capitalization → base ~$16.5; weighted ~$16.4 vs ~$15.75. Roughly fair — cheap cash flow and a real moat, but no growth in the number and the floor is being spent. Hold / watch-list; accumulate below ~$12.50.

The highest-quality of the neglected small-caps — a ~20%-growth identity/data-analytics compounder with 77% recurring revenue, 95% gross retention, net cash, a genuine entity-resolution moat (idiCORE + FOREWARN on the CORE/IRON platform), and the cleanest AI-tailwind story in the queue (usage-based pricing benefits from customer automation). The DRIVER MODEL (owner FCF $18m->$78m as capitalised-dev normalises 13->7%; WACC 11.75%, net cash) computes base FV ~$46.87 / bear $30.69 / bull $61.57 — and the ~$60.86 price sits right at our BULL case, i.e. the tape is paying for perfection. Adjusted EBITDA is flattered (R&D 100% capitalised); value on FCF/owner-FCF. Watch / do-not-chase + PUT candidate on a hype-driven overshoot; a crack lowers BOTH price and FV, so the real long entry is below today's FV.

The deep-value contrarian of the queue: a unified-CXM platform that de-rated to a ~1.0× EV/sales stub because growth stalled (guided +1% FY27) and net retention collapsed (120%→103%). The "14% FCF yield" is a headline illusion — on OWNER FCF (FCF − ~$84m SBC), FY26 was the FIRST positive year (~$57.5m). The crux is MARGIN not revenue: the driver-based DCF (base ~$5.77) lives in whether owner-FCF margin normalizes toward ~9%. Value-trap mechanism: strong product (59% of Fortune 100) but outsourced/offshore post-sale support → troubled CCaaS implementations → churn (ties together the NRR collapse, the large-customer slide, and the declining Americas). At ~$5.69 it is fairly-valued-to-slightly-full; positive skew (bear ~$4.00 / bull ~$8.20) but MoS only appears sub-$4.50 — the accumulation trigger. Watch / do-not-buy-here.

The odd-one-out of the queue: a patent-LICENSING house (royalty annuity, has debt + pays a dividend), not operating software. A cash-gushing, deleveraging IP business — Media IP ~94% (pay-TV melting, OTT growing ~50% share) + a genuine, back-end-loaded AI call option (semiconductor hybrid bonding for HBM/advanced packaging, 2027-28). Q2-2026 (8-K, 3 Aug) DE-RISKED the thesis without changing the near-term guide: Google/YouTube TV renewal re-locked the marquee pay-TV annuity, non-Pay-TV recurring +54% YoY, and management RAISED the long-term outlook $500m→$600m (semi to $200m) — the AI option, previously our inference, now management-stated. On a disciplined owner-FCF DCF (β0.95 5Y-monthly not the AI-inflated 2yr-daily 1.37, WACC ~9.8%, net-debt-subtracted), base fair value is ~$15 (weighted ~$16) — and at ~$23.24 (post the −17% Q2 reaction) the price still sits ABOVE it. The DCF is unchanged; the de-risking shows up as a lower bear probability (35%→25%), not a re-cut forecast. WATCH; accumulate only sub-$15.50. The negative reaction to a good print is a momentum tell — an overshoot toward ~$15 would be the entry, on a now-better thesis.

A founder-owned (Trefler ~45%, votes = economics), net-cash, ~$490m-FCF, Rule-of-40 enterprise-software leader — AI decisioning (Customer Decision Hub) + workflow automation, unified as Pega Infinity, sold to large REGULATED enterprises where the work must be predictable, governable and CHANGEABLE ("Build for Change"). De-rated ~46% in 2026 on a genuine H1 ACV stumble (2Q26 ACV +8% cc vs a +15% guide) + a sector-wide AI software de-rate. The turnaround is real and owner-positive (op income -$109m FY22 -> +$263m FY25; FCF $201m -> $491m/28%; owner FCF $335m/19%), the balance sheet pristine (converts repaid in cash, buying back >100% of FCF), and the Appian ~$2bn verdict VACATED (retrial Jan-2027, no accrual). Verification found NO hidden structural catalyst — long-holder capitulation, low & falling short interest. Tape-blind owner-FCF DCF (beta 0.87, WACC 9.46%, net cash): base $53.62 / weighted $51.04 (bull $75.00 20% / base 45% / bear $34.05 35%). At ~$31 the market prices BELOW our BEAR; street ~$58 and TIKR ~$51 sit above our driver-derived base. The crux: is the stumble an air-pocket or structural — the 2Q26 call read as partly structural (mgmt "max confusion", recovery "very difficult") so we cut the base recovery + lifted the bear weight; Q3-2026 (2/3 of net-new ACV is back-half-weighted) is the binary tell. HOLD / accumulate on weakness (conviction cut from Buy, not broken); cheap even on the bear, but hold rather than press until Q3.

The quality anchor of the fintech batch: the SaaS + AI + data NETWORK that runs the US auto-claims economy — 27 of the top 30 auto insurers + ALL national repair MSOs + 14 of 15 OEMs, 98-99% gross retention, $2 trillion of proprietary claim data, and the rare vertical SaaS where AI DEEPENS the moat. It de-rated ~40% in a year (issued EvolutionIQ stock at $11.83, bought it back at ~$7, now $6.05) despite beat-and-raise results. The one real caveat — enormous SBC making reported FCF flatter the truth (owner FCF was only ~$79m of the $255m reported in FY25) — is being fixed: SBC guided 17%→13%→single digits by 2027, which roughly TRIPLES owner FCF to ~$226m by FY27. On a disciplined owner-FCF DCF (β0.8129, WACC ~9%, net-debt-subtracted), base fair value ~$8.00 (+32%); independent bear ~$3.28 (leverage-amplified, below the ~$4.11 market low), bull ~$13.00. A-tier quality "only just" — de-rated but leveraged, and conditional on the cost/SBC discipline (which 1Q26 evidences). BUY (modest)/Accumulate; the fat pitch is a flush toward $5.

The de-rated CFO-software compounder: the dominant INDEPENDENT financial-close / record-to-report platform (reconciliations, close, journals, intercompany, invoice-to-cash) — audit-grade software your auditors sign off THROUGH (95-97% enterprise renewal, ~70% of the Fortune 100). Growth decelerated 22.8%(FY22)->7.2%(FY25) and the market extrapolated ex-growth; underneath, NRR troughed (102->105), non-GAAP op margin inflected 6%->22%, and FY25 brought the first-ever buyback ($235m, >2x SBC). The best AI-durability framing in our software book — the audit/control LAYER for the agentic enterprise (every AI transaction hits the GL and must be reconciled/governed/audited), with real consumption monetisation a 2027 option. The honesty crux (Eric's catch): reported FCF ~$135m looks cheap but OWNER FCF (FCF − SBC) is only ~$42m as SBC runs ~13% of revenue — and, unlike CCC, there's no commitment to cut it. On a tape-blind owner-FCF DCF (β0.8785, WACC ~9%, net-debt-subtracted): base ~$36.00 (+17%), independent bear ~$18, bull ~$65. B+ quality (7.1) — sticky + optioned but capped by thin owner-FCF + terminal ERP/AI risk. BUY/Accumulate, but a WATCH at $30.72 — it needs a real dip; accumulate < $29, trim > $46.

The just-inflected digital-banking SaaS: a sticky, mission-critical single platform (retail + SMB + commercial digital banking + relationship pricing + risk & fraud + Innovation Studio marketplace) for US community/regional banks and credit unions — 1,200+ FIs, 115% subscription NRR, $2.8bn backlog, record bookings. It has just crossed its profitability inflection (cloud migration DONE Jan-2026; gross margin stepped 58%->62%; adj-EBITDA margin 23.5%->28% guide, raised twice in 2026; net cash after retiring the converts). The moat is wide (brutal switching costs, regulatory + ecosystem lock-in, 93% M&A-survivor win). But the finding is PRICE, not quality: on a tape-blind OWNER-FCF DCF (β1.1206 app-Blume — NOT Yahoo's 1.33 — WACC ~11.8% on a live 4.6% risk-free rate, net cash added) base fair value is ~$40, and the ~$60 tape sits ABOVE even our bull case (~$57). Reverse-DCF: $60 needs WACC ~8.7% on our base — the market is discounting a low-teens grower at a mega-cap-quality rate. The honesty crux: reported FCF $173m looks strong but OWNER FCF (FCF − SBC) is $86.4m as SBC runs ~11% of revenue — ~48x trailing. A- quality (7.8) — genuinely excellent — but a firm HOLD / WATCH purely on price. Accumulate only on a real de-rating; no action at $60.

The "contrarian de-rating" that turned out NOT to be a bargain. Flywire is a vertical cross-border payments + software network (Education/Travel/Healthcare/B2B) with a genuine, hard-to-replicate global network (240+ countries, <1% churn) — but two things cap it. (1) The honest cash number is ~zero: management's OWN adjusted FCF (ex client-float) was $62m FY25 (flat YoY), and owner-FCF (after ~$72m SBC) is NEGATIVE and has been every year. (2) Its largest, highest-margin vertical — international education (~68%) — faces a genuine MULTI-YEAR structural headwind (student-visa/immigration tightening across Canada/AU/UK/US + US student-loan policy + a $100k H-1B fee that also hits its OWN cost base), and the diversification meant to offset it is BOUGHT (Sertifi/Invoiced) and lower-margin. On a tape-blind owner-FCF DCF (β1.025 app-Blume NOT Yahoo 1.32, WACC ~12.24%, net-cash-added): base ~$14 / bear ~$5.5 / bull ~$23.7, weighted ~$13. At $16.54 that's ~15% ABOVE base fair — no margin of safety; the market is pricing the bull. HOLD/watch (B/B-, 6.0). Accumulate only on a flush toward ~$11 (≈ the 52wk low + mgmt's buyback avg) AND evidence the owner-FCF inflection is real.