Moat
Moat — What Protects This Business
A moat is a durable, company-specific economic advantage that lets a business defend returns, margins, share, and customer relationships against competition. The question this page tries to answer in one line is: what, if anything, protects Kaspi from competition, how do we know it works, and what would make it fade?
1. Moat in One Page
Conclusion: narrow moat — wide inside Kazakhstan, unproven outside it. The Kazakhstan business has a real, measurable advantage: a three-engine super-app (Payments + Marketplace + Fintech) running on a single 15.7M-user base that equals ~83% of the adult population, funded by a 93% retail-tenge deposit franchise, and underwritten by behavioural data that lets Kaspi originate 99.9% of loans in under six seconds at a 2.2% Cost of Risk. That bundle has produced 51%+ ROE every year since FY2019 through a Covid shock, a 220 bps funding-cost step-up, regulatory ratchet, an activist short report, and the Hepsiburada consolidation. Outside Kazakhstan, the moat is a bet, not a proven asset — Hepsiburada engagement per consumer is roughly one-quarter of Kazakhstan's, and the Türkiye deposit licence (Rabobank A.Ş.) is unclosed.
The two weakest links are single-country geography (the moat does not export easily — Türkiye is paying integration cost without yet earning a deposit franchise) and Payments take-rate compression (1.18% → 1.10% FY24→FY25 as QR / B2B mix shifts faster than card-network revenue can compensate). If Cost of Risk drifts above 3%, the deposit-cost gap with Halyk closes, or Hepsiburada engagement stalls, the rating becomes "narrow and shrinking" rather than "narrow and proving."
Moat Rating
Evidence Strength (0–100)
Durability (0–100)
Weakest Link
Why this is not "wide moat" yet. The Kazakhstan economics resemble a wide moat — high ROE sustained across cycles, dominant brand recall, low Cost of Risk through rate hikes. But a wide-moat call requires the advantage to survive export or regime change, and Kaspi's first test of either (Türkiye) is mid-flight with engagement metrics roughly one-quarter of the Kazakhstan baseline. Until Hepsiburada engagement converges or a second country reaches Kazakhstan-class economics, narrow is the honest grade.
2. Sources of Advantage
Each candidate moat source below is explained in plain English on first mention, then mapped to a specific Kaspi disclosure or operating fact rather than to adjectives. A switching cost is the friction (workflow disruption, data migration, lost rewards, retraining) that keeps a customer from leaving even when a rival offers a marginally better deal. A network effect exists when the product becomes more valuable to each user as more users (or merchants, or counterparties) join. Data flywheel means underwriting or product decisions get measurably better with each additional transaction observed. Brand intangible protects the business only when it changes customer behaviour — share, pricing, or churn.
3. Evidence the Moat Works
A moat that doesn't show up in returns, share, retention or pricing is just storytelling. The table below lists eight pieces of evidence — six supporting the moat and two refuting parts of it — drawn from filings, the FY2025 20-F, KResearch / Kantar surveys, and Halyk Bank's published reports. Confidence reflects how independently verifiable the data point is; watch for distortion names the way the number could mislead.
The ROE step-down from 79% to 51% in FY2025 is the consolidation effect, not moat decay — equity base expanded materially with Hepsiburada and ₸326B of new Eurobond debt. On a Kazakhstan-only basis ROE remained in the high-50s. The durability test is that returns stayed in the 50%+ band even after consolidation diluted them.
4. Where the Moat Is Weak or Unproven
Honest moat analysis requires naming what could go wrong. Four pressures matter most over the next 24 months.
The moat conclusion depends on one fragile assumption. If Hepsiburada fails to reach Adj-EBITDA breakeven on the 2026 timeline and Halyk Super App grows faster than Kaspi's MAU base, the rating moves from "narrow moat" toward "narrow moat, shrinking." Both signals are visible in quarterly disclosures — there is no need to guess.
5. Moat vs Competitors
No listed company combines all three engines of the Kaspi bundle, so this comparison reads each peer for the moat dimension where it is closest. Halyk is the only in-country read. The integration column captures the meta-moat: how many engines run on one customer base.
The structural point: Kaspi scores at the top of five out of six moat dimensions inside Kazakhstan and at the bottom of the sixth (geographic TAM). MELI is the closest analog at the integrated-business level; Halyk is the closest analog at the deposit-franchise level; nobody combines both. The Türkiye build-out is the company's stated answer to its single weakness — and it has not yet earned an upgrade on the scorecard.
6. Durability Under Stress
A moat only matters if it survives stress. Below are eight stress cases drawn from the macro, regulatory, and competitive evidence in the rest of the deck, with what FY2024-25 already showed under partial versions of each.
The pattern across the eight stress cases is that the moat survives where the loop runs on one customer base (rate cycle, short report, Halyk catch-up, Big Tech wallets, cross-border marketplaces) and is unproven where Kaspi is exporting the playbook (Türkiye macro, by extension future second-country expansion). Founder concentration is the only stress case the deck cannot quantify in advance.
7. Where Kaspi.kz JSC Fits
The moat is not uniform across the bundle. The honest split is that Kazakhstan carries a wide-moat-quality engine, Türkiye carries no moat yet, and the three Kazakhstan engines themselves have different moat qualities. Reading the multiple discount correctly requires holding all three views at once.
The bull case rests on Türkiye crossing from "no moat" to "narrow moat" while Kazakhstan stays at "wide-quality." The bear case is that Türkiye stalls (HEPS engagement plateaus, Rabobank delayed) while Kazakhstan compresses (NIM, take-rate, regulatory ratchet). What the multiple is paying for today is the bear case; what the moat evidence supports is closer to the base case.
8. What to Watch
These are the six signals that test the moat directly over the next four quarters. Each is observable in 20-F filings, quarterly press releases, NBK/ARDFM publications, or Halyk's annual report.
The first moat signal to watch is Cost of Risk and NPL coverage rebuild — the underwriting moat is the leg most directly tied to the data-flywheel claim, and the FY2023 ICFR weakness in the ECL model means a small drift in either number changes the rating before any other signal does.