StoneCo Ltd. (NASDAQ: STNE) — Coiled Spring or Falling Knife on Rate-Flattered Earnings

An independent equity research note Date: June 8, 2026 Price reference: ~$10.6 (52-wk range $9.45–$19.95) · Market cap ~$2.6B · ~229–273M shares Reporting: Foreign private issuer (Cayman-incorporated, Brazil-operating); IFRS in Brazilian Reais (R$); files 20-F / 6-K · CIK 0001745431 · FX ~5.16 USD/BRL · Sector: Financials — Fintech (Payments + Banking + Credit)

⚡ Claude’s Take

This block is the author’s own subjective opinion and general information only — not investment advice. The analysis that follows is deliberately position-free and carries no price target; this block is the single exception.

Verdict: Speculative BUY / accumulate-on-weakness — a buyback-compounder at a distressed multiple, but a high-beta falling knife. Size it small. Conviction: low-to-medium. Tag: “Cheap on contested earnings, fair on normalized earnings — coiled spring or falling knife, and you only know which afterward.”

StoneCo screens like deep value: ~7x clean earnings, ~1.3x tangible book, ~20% ROE, ~6th-percentile price/sales versus its own history, returning roughly a fifth of its market cap a year through buybacks plus a freshly-initiated dividend and a $2.53/share special dividend from the Linx sale. The buyback math alone is compelling — repurchasing ~19% of the cap at ~7x earnings mechanically delivers high-single-digit per-share growth even if absolute profit goes nowhere (Q1 2026 proved it: adjusted EPS +15.4% on +3.5% net income). If you believe Brazil’s rate cycle normalizes gently and the credit book behaves, you are paid ~10% a year to wait for a re-rate that the ~12% short interest and a 65%-upside analyst target say is plausible.

But I will not pretend this is a high-conviction compounder, and the honest framing is the opposite of the bulls’ “payments-growth-story.” About 71% of revenue is SELIC-rate-sensitive financial income — today’s ~20% ROE is earned at a ~15% (now easing) rate peak, and management itself cut 2027 gross-profit guidance ~29% and withdrew KPI guidance, conceding the normalization. Embedded expectations already price through-cycle ROE of only ~13–15%, so the multiple is low because the “E” is genuinely contested, not because the market is asleep. Layer on a structural PIX threat to the card/prepayment profit pool, a fast-growing credit book with rising NPLs (the 2021 blow-up is not ancient history), the worst user-growth in the Banco Central ranking as Nubank and Mercado Pago pull away, an October 2026 election with real BRL/fiscal tail risk — and the fact that Berkshire, the canonical value buyer, fully exited at the end of 2023. My accumulate zone is below ~$11 / under ~1.2x tangible book; base-case fair value ~$15–18 if through-cycle ROE holds the mid-teens and Brazil partially re-rates; downside to ~$7–8 on a credit-plus-election shock. This is a position for risk-tolerant EM investors who can size it as one of many small contrarian bets, not a core holding.

What flips me more bullish: two to three quarters of stable-to-rising take rate and contained cost-of-risk as SELIC falls (proving the earnings aren’t merely rate carry) → the buyback engine plus a Brazil re-rate compound together. What flips me bearish: cost-of-risk pushing toward 20%+ / NPLs climbing, OR sequential take-rate erosion as Pix Parcelado scales — either confirms the value-trap read and the buyback is then destroying value into a melting “E.” On balance the price already discounts a lot of bad news and the capital return is real, which earns a small constructive tilt — but the quality and durability are not here, so conviction stays low-to-medium.

1. Executive Summary

StoneCo is a Brazilian financial-technology company serving micro, small and medium businesses (MSMB) with a bundle of card-acquiring (payments), digital banking, and merchant credit, having just divested most of its software (Linx) segment. After a brutal multi-year de-rating — from a 2021 peak above $90 (pre-events) to ~$10.6, near its 52-week low — the stock trades at roughly 7x clean earnings, 1.3x tangible book, and ~1.2x sales (the ~6th percentile of its own ten-year history), with ~12% short interest and a ~$17–18 average analyst target implying ~65% upside. The setup is a textbook contested-value EM financial.

The business is profitable and, on the surface, returning to form: FY2025 revenue R$14.2B (+11%, ~$2.7B), operating income R$7.3B (+24%), IFRS net income R$2.3B (~$0.45B) after a 2024 IFRS loss caused by a ~R$3.56B goodwill impairment of the Linx acquisition. ROE is ~20%, and management is returning capital aggressively — ~R$2.9B of buybacks in 2025 (~19% of market cap), a newly-initiated regular dividend, and a $2.53/share special dividend funded by the Linx-to-TOTVS sale.

The central analytical fact is that ~71% of revenue is SELIC-rate-sensitive financial income (prepayment of card receivables, merchant credit, and deposit float). Today’s earnings are flattered by Brazil’s ~15% policy rate, which has begun easing (14.50% as of April 2026; market consensus anchors ~10.5% by 2027). Management itself cut 2027 adjusted-gross-profit guidance ~29% (R$10.2B → R$7.2–8.3B) and discontinued KPI guidance — an explicit admission that the through-cycle earnings power is materially below the rate-peak run-rate. Embedded-expectations math says the current price already prices through-cycle ROE of ~13–15% (a 25–35% earnings decline) — i.e., the market is not naively extrapolating peak earnings.

The bear case is structural, not merely cyclical: PIX (Brazil’s central-bank instant-payment rail), now extended to installments (Pix Parcelado) and contactless, is a slow-moving attack on the high-margin card-MDR and receivables-prepayment profit pool; competition from far-better-capitalized super-apps (Nubank, Mercado Pago) is winning the customer-relationship war — the Banco Central client ranking shows StoneCo with the sharpest user decline while rivals add millions; the relaunched credit book (~R$2.9B) carries rising NPLs (>90d 5.2%) and ~17% cost of risk into a softening SMB macro, echoing the 2021 credit blow-up; and the October 2026 election injects BRL/fiscal/equity-risk-premium volatility. Berkshire Hathaway’s full Q4-2023 exit underlines the skepticism.

The bull case rests on the buyback compounding at a distressed multiple (~10% annual return even with flat earnings), a known-and-guided rate headwind that the price appears to over-discount, a clean tangible book now that goodwill is nearly written off, and Brazil re-rating optionality. Net: a structurally mediocre and deteriorating industry, a narrow and eroding moat, rate-flattered and contested earnings, intelligent current capital allocation against a scarred M&A record, and a multiple that is cheap on reported earnings but only fair on normalized earnings. No recommendation or price target appears below this section.

2. Business Overview

What StoneCo does. StoneCo is a payments-led financial-services platform for Brazilian MSMB merchants, with four interlocking products. (FACT — FY2024 20-F; FY2025 6-K earnings release.)

• Payments / acquiring (the engine). Stone signs merchants to accept card payments, capturing total payment volume (TPV) and earning a take rate (merchant discount rate, MDR) plus a spread from prepaying merchants their card receivables. FY2025 total TPV was R$560.9B (+8.7%); MSMB TPV R$503.4B (+11%). The MSMB take rate reached a record 2.58% in 2024.
• Digital banking. Stone/TON accounts, cards and deposits. Retail deposits reached R$11.1B in Q4-2025 (+27% YoY) — the cheap funding base for prepayment and credit. The cross-sell metric ARPAC was ~R$278/month, but down ~11% YoY (mix shift toward smaller/TON clients).
• Credit. Relaunched cautiously after the 2021 collapse; portfolio ~R$2.9B at Q4-2025 (working capital + cards), still small relative to the franchise but the main forward growth-and-risk lever.
• Software (Linx) — divested. StoneCo sold the core of Linx (retail/ERP software) to TOTVS (enterprise value R$3.05B, ~R$3.19B including net cash; ~79% of software revenue), closing February 27, 2026. Software was a small, declining share of profit; the exit simplifies the story.

How it makes money — and the critical nuance. StoneCo’s “revenue” is dominated by financial income, not processing fees. Financial income (prepayment + credit + deposit float) runs roughly R$10B annualized, ~71% of FY2025 revenue, while true transaction/MDR revenue is actually shrinking (management explicitly reclassifies economics from MDR into financial income via bundled pricing). This means headline “revenue growth” overstates underlying franchise growth and is heavily geared to Brazil’s interest-rate cycle. (FACT — quarterly disclosures; INTERPRETATION — the business is best modeled as a specialty-lender/acquirer hybrid, not a pure payments compounder.)

Customers and recurring nature. Core customers are long-tail Brazilian merchants (micro via the TON brand; small/medium via Stone). Revenue is largely recurring/transactional (merchants reorder acceptance, banking and credit monthly), but the profit is rate-sensitive. Active payment clients ~4.6M. (FACT — 6-K.)

Verdict (Business Overview): A focused MSMB fintech with a coherent payments-banking-credit bundle, but one whose economics rest overwhelmingly on a rate-sensitive financial-income line rather than durable processing fees. Operationally an “ecosystem”; economically a balance-sheet-geared lender/acquirer.

3. Industry Dynamics

Market structure. Brazil’s card-acquiring industry was historically a Visanet/Redecard duopoly (now Cielo and Rede), cracked open from ~2010 by Stone, PagSeguro, GetNet, Mercado Pago and SafraPay. Cielo still leads (~R$880B TPV / ~25% share); Rede ~22% (Itaú). The market is large and growing on card penetration, but structurally crowded — more than five credible players — and increasingly regulated. (FACT — industry profiles; Reuters.)

PIX — the structural wild card. PIX, the central bank’s instant-payment rail (launched 2020), is reshaping profit pools. Two extensions sharpen the threat: Pix Parcelado (installments without a credit card, live end-October 2025), explicitly positioned as a credit-card alternative, and Pix contactless/NFC (2025). The USTR even opened a Section 301 investigation, with a June-2026 determination calling Brazil’s PIX practices “unreasonable” and proposing a 25% duty — underscoring how disruptive the rail is. (FACT — Payment Expert; USTR; Federal Register.)

For StoneCo specifically, PIX is two-sided but net-negative at the margin. Stone monetizes PIX QR-code acceptance (MSMB PIX QR TPV R$78.3B in 2025, +56%), and management says PIX QR “offers comparable monetization to debit.” So today PIX largely displaces low-margin debit while Stone still earns a fee. The risk is structural: as Pix Parcelado scales, it attacks the high-margin credit MDR + receivables-prepayment spread — the actual profit core — and account-to-account payments can increasingly route outside any acquirer, recaptured (if at all) only at debit-tier take rates. (INTERPRETATION — the single most important long-run industry variable after rates.)

Capital cycle (Marathon lens). Brazilian fintech is in a bust→early-recovery phase: fintech funding collapsed, the 2022–23 MDR price war cooled into “more rational pricing,” and consolidation is underway (Cielo taken private by Banco do Brasil/Bradesco in 2024). Normally, contracting supply precedes recovering returns. But the cycle is broken by state action — Banco Central is permanently capping the card profit pool via PIX, and super-app entrants (Nubank, Mercado Pago) are funded by equity markets, not card economics. The classic “supply contracts → returns recover” mechanism is partly disabled. (INTERPRETATION — a capital-cycle breakdown case: a policymaker deliberately distorting the cycle.)

Verdict (Industry Dynamics): Structurally MEDIOCRE and deteriorating. Good volume growth (card penetration, TPV), but a regulator-pressured, multi-player market where MDR trends down and a state-sponsored rail (PIX) is engineered to compress the very profit pool Stone depends on. Volume grows; durable card profit does not.

4. Competitive Position

Where Stone stands. Stone is a strong MSMB-acquiring specialist that is gaining narrow card-acquiring share but losing the broader platform war.

• Card share: Stone’s MSMB share of total card industry (ABECS) rose 9.4% (2023) → 9.7% (2024) — modest gains in its niche. (FACT — 20-F.)
• Profit parity lost: On acquiring profit alone (Itaú BBA, 2024), Mercado Pago (R$1.93B) overtook Stone (R$1.90B), with PagBank close behind. (FACT — NeoFeed.)
• The damning datapoint: A Banco Central client ranking (Q4-2025→Q1-2026) shows Nubank (+2.7M to ~114.7M) and Mercado Pago (+2.5M to 71.3M) leading growth while StoneCo recorded the sharpest DECLINE in users. (FACT — Brazil Stock Guide.)
• Incumbents stronger: Cielo (now private, BB/Bradesco) and Rede (Itaú) can compete patiently on aggressive MDR and zero-fee receivables, integrated with the largest banks.

The moat — pressure-tested. The genuinely defensible asset is Stone’s hyper-local hub distribution — physical sales/service hubs and agents delivering same-day onboarding and on-the-ground service to long-tail merchants the big banks underserve. This is hard and expensive to replicate at scale, and it ties directly to the financial outcomes that would deteriorate without it: the record 2.58% take rate and ~25% (segment) ROE. Bundling (banking + payments + credit; ~40% multi-product merchants) and the growing deposit base add switching costs and cheap funding. (FACT — 20-F; 6-K.)

Greenwald classification: a localized economies-of-scale + customer-captivity moat — strongest exactly where Greenwald says, in small/local/niche markets where fixed cost-per-merchant of physical service is the barrier. But it fails the share-stability test at the platform level (Stone is losing clients while super-apps add millions), the accuracy/price edge is gone (Nubank and Mercado Pago are better-capitalized, lower-funding-cost, and own the consumer relationship), and ARPAC is falling 11%. It is a distribution moat, not a network-effect or proprietary-technology moat — defensible in the MSMB acquiring niche, not a fortress against the ecosystems.

The Linx post-mortem. StoneCo paid ~R$6.7B for Linx (2021) chasing software-payments cross-sell synergies that proved illusory — a textbook Greenwald M&A failure. It impaired ~R$3.56B (2024) and sold the core to TOTVS (2026) for roughly half the purchase price. The divestiture is the right call (asset contraction, special dividend, simplified story — Marathon-positive at the company level) but crystallizes ~R$3.5B of destroyed value and is a permanent mark against management’s capital allocation. (FACT — 6-Ks; INTERPRETATION.)

Verdict (Competitive Position): A NARROW, localized scale-plus-captivity moat in MSMB acquiring, eroding at the edges. Real and tied to genuine margins, but not durable against PIX disintermediation or against Nubank/Mercado Pago owning the merchant’s primary financial relationship. A defensible niche, not a wide moat.

5. Growth History and Forward Opportunities

History — growth, a blow-up, and a rate-flattered recovery. Revenue (R$M): 2020 3,167 → 2021 4,576 → 2022 9,016 → 2023 11,364 → 2024 12,739 → 2025 14,154. The trajectory looks strong, but two caveats dominate: (1) the 2021 credit blow-up (a faulty receivables-registry rollout drove disastrous credit losses and a R$1.36B net loss, forcing a credit-product shutdown); and (2) much of the 2022–2025 revenue surge is financial income inflated by Brazil’s rate hikes, not pure franchise growth. (FACT — financials.)

The quality problem in the growth. TPV growth is decelerating (MSMB TPV +11% FY2025 but only +5.3% in Q4-2025); ARPAC is −11% YoY; and the headline 2.58% take rate is rising optically because economics shift from MDR into prepayment/financial income, not because of pricing power. The growth is increasingly volume at declining unit economics, geared to rates. (FACT/INTERPRETATION.)

Forward opportunities.

1. Banking / deposits — the genuine bright spot: deposits R$11.1B (+27%), a growing, cheap, sticky funding source that improves net interest margin and engagement.
2. Credit — the highest-spread, highest-beta lever; ~R$2.9B and >2x YoY. Done well, it re-rates growth quality up; done badly, it is 2021 again.
3. PIX monetization — Pix QR volumes growing fast (+56%), recapturing some of what cards lose, but at debit-tier margins.
4. Capital return as a “growth” lever — at a distressed multiple, the buyback drives per-share growth independent of the top line (see Capital Allocation and Valuation).

Verdict (Growth): MIXED-to-LOW quality. TPV and ROE look healthy, but decelerating volume, falling ARPAC, the 29% cut to 2027 gross-profit guidance, share losses to super-apps, and PIX eating the high-margin layer mean the growth is increasingly low-quality. The credit relaunch is the swing factor in both directions.

6. Financial Quality

Summary (R$ millions; ~US$ at FX 5.16–5.4):

(FACT — EDGAR/IFRS via filings and data feeds; reconcile final FY2025 lines to the 20-F.)

The rate-dependence frame (most important). A large majority of revenue and most incremental profit is net interest / financial income geared to SELIC. The financial-expense (funding-cost) line more than doubled 2023→2025 with rates, so gross profit must be read net of it. With SELIC easing, management guides financial-income margin to compress — the explicit basis for the 29% cut to 2027 gross-profit guidance. (FACT — guidance; INTERPRETATION — peak-rate earnings are not the run-rate.)

Quality of earnings. The 2024 IFRS loss was almost entirely the ~R$3,558M non-cash Linx goodwill impairment (Q4-2024 printed −R$2,922M). Stripping it, the franchise was profitable throughout. Encouragingly, the FY2025 adjusted-vs-IFRS gap is small and defensible (~R$2.61B adjusted vs R$2.32B IFRS — chiefly intangible amortization and SBC), unlike many fintechs that adjust away real costs. The credibility issue is historical (2021 credit; 2024 impairment) and prospective (rate normalization), not aggressive add-backs. The honest “clean” continuing-ops P/E is ~7x — not the ~4x headline, which is inflated by now-discontinued Linx earnings. (FACT/INTERPRETATION.)

The credit book — the asymmetric risk. Portfolio ~R$2.9B (>2x YoY); NPL >90d 5.21% and rising (15–90d 4.43%, blamed on higher-ticket “specialized desk” clients); cost of risk ~17%; coverage ~264%. The book is still small (~4.7% of assets), so a shock is contained today, but it is growing fast into a weakening SMB macro with rising delinquencies — exactly the 2021 setup. At ~17% cost of risk against a ~37%/yr gross yield, the unit can flip to a loss quickly if NPLs run. (FACT — 6-K/transcript; INTERPRETATION.)

Balance sheet & funding. Equity ~R$11.8B (flat 2024→2025 as buybacks offset earnings); total assets R$62.3B; cash ~R$4.8B. Stone operates a regulated bank/SCD (Banco Stone), so the ~R$8.8B of “long-term debt” is funding for the receivables/credit book, not corporate leverage — a fintech-typical conflation. The funding mix is shifting toward low-cost retail deposits (a structural positive). No solvency concern; exact Basel/regulatory-capital ratio not extracted (open question). (FACT/OPEN QUESTION.)

Returns. ROE TTM ~20%, but flattered by the SELIC peak and a shrinking (buyback-reduced) equity base. Through-cycle ROE is plausibly mid-teens, not 20%+. Tangible-equity returns look better now that goodwill is nearly gone (R$671M). (FACT/ASSUMPTION.)

Verdict (Financial Quality): Economics are rate-flattered, not cleanly scale-driven. Real operating leverage and a genuine deposit-funding advantage exist, but a large share of the “improvement” is the rate tailwind, TPV growth has stalled, and the credit book is a live tail risk. Earnings quality is currently clean but cyclically/rate-inflated.

7. Capital Allocation

Aggressive, well-priced capital return — against a scarred M&A record.

• Buybacks: R$2,928M in 2025 (~$567M, ~19% of market cap), retiring ~40M shares; share count fell 319M → 302M → 273M (2023→2025; yfinance shows ~229M more recently). R$2.0B more authorized for 2026. (FACT.)
• Dividends: a regular dividend initiated (~2% yield, ~27% payout) plus a $2.53/share special dividend (~R$3.08B) from the Linx proceeds, declared April 2026 and paid May 2026. (FACT — 6-K.)
• The buyback logic: repurchasing at ~1x tangible book / ~7x earnings is rationally accretive if ~mid-teens-plus ROE is sustainable — a ~14% earnings yield redeployed into share retirement. Q1-2026 demonstrated the engine live: adjusted EPS +15.4% on +3.5% net income — ~12 points of per-share growth from the buyback alone. (FACT.)
• The M&A black mark: the ~R$3.5B Linx value destruction (bought ~R$6.7B, impaired ~R$3.56B, sold the core for roughly half) is a serious demerit on the same management’s record.

Incentives & governance. CEO Pedro Zinner (non-founder, since 2023); André Street chairman; CFO Mateus Scherer. The pivot from empire-building (Linx) to disciplined capital return is the correct correction, executed at attractive prices. Caution: management cut the core equity-ratio hurdle (20%→17% per updated BCB methodology), which frees buyback capacity but modestly raises balance-sheet risk into a softening credit cycle. Insider ownership is low (~2.8%). (FACT — transcript; MarketScreener; INTERPRETATION.)

The Berkshire signal. Berkshire Hathaway entered at the 2018 IPO (~11% peak), trimmed through 2021–2023, and fully exited in Q4 2023 — a notable negative from the canonical value investor, coinciding with the post-Linx/credit-crisis period. (FACT — Seeking Alpha.)

Verdict (Capital Allocation): Mixed-to-positive, improving. The current buyback-plus-dividend program is intelligent and well-priced — genuinely the strongest leg of the bull case. But the Linx destruction and low insider ownership temper the grade. Direction of travel favorable; track record scarred.

8. Changes and Headwinds — Last Two Years

SELIC normalization (the dominant driver). SELIC peaked at 15% and is easing (14.50% as of April 2026); market consensus anchors ~10.5% by 2027. Lower rates compress the gross yield on Stone’s float/prepayment/deposit book (a headwind to ~71%-of-revenue financial income) but also lower funding cost and support SMB credit demand and volumes (a partial offset). On a balanced book the funding-cost offset cushions but does not fully neutralize the yield drag on equity- and deposit-funded float — which is precisely why management guides 2026 gross profit down before a 2027 reacceleration. (FACT/INTERPRETATION.)

2026 election. The October general election (Lula vs. the Bolsonaro camp) is a BRL-volatility and equity-risk-premium event, amplified by pre-election fiscal spending. A market-friendly outcome could compress Brazil’s ERP and re-rate a high-beta ~7x financial with ~12% short interest; fiscal deterioration weakens the BRL (hitting the USD-listed share via translation) and threatens the very rate-cut path Stone’s guidance assumes. (FACT/INTERPRETATION — top-down, apolitical.)

The 2027 guidance cut & credibility. Cutting the 2027 adjusted-gross-profit floor ~29% and withdrawing KPI guidance is partly mechanical (removing Linx) but partly a downgrade of the core trajectory (rate/take-rate/mix). Two-plus years after a flagship Investor Day, it is a real credibility dent — management has spent credibility it will need to rebuild. (FACT/INTERPRETATION.)

Verdict (Changes & Headwinds): Net WEAKENING over the past 18 months — but much is in the price. Structural negatives (rate normalization, PIX, election/fiscal risk, rising credit NPLs, the guidance reset) dominate the narrative; offsetting positives (deposit growth, Linx simplification, aggressive cheap capital return) are real. At ~7x / ~1.3x book near a 52-week low with ~12% short interest, the market already prices a meaningfully negative scenario.

9. Risk Analysis (Risk Matrix)

Catastrophic-loss assessment. Solvency risk is low — the company is profitable, has a regulated bank with adequate capital, net positive equity, and is generating cash. The realistic severe case is a value-trap drawdown (earnings normalize down 25–35%, multiple stays low, BRL weakens, credit costs rise) — a further 30–40% loss from a price already near lows — not insolvency. A 2021-style credit accident plus an election shock is the plausible tail. (INTERPRETATION.)

10. Valuation Discussion (Embedded Expectations)

No price target and no recommendation in this section.

Method note. Stone’s bank-like balance sheet (credit book, prepayment receivables, debt-funded float) makes EV meaningless — favor P/E, P/TBV, P/S. Goodwill is nearly written off (R$671M), so P/TBV ≈ P/B on a “clean” book. (FACT.)

Comparables (favor P/E, P/B, P/S; ROE-adjusted):

(Sources: stockanalysis.com peer pages; access Jun-2026. The ~4x P/E sometimes quoted for STNE is inflated by discontinued Linx earnings — ~7x is the honest clean figure.)

Where STNE screens. Against its only true comp, PagSeguro, STNE is fair-to-slightly-rich on a P/B-to-ROE basis (1.3x book on ~20% ROE vs PAGS ~1.0x on ~15%). The dramatic “cheapness” is only versus its own history (P/S ~6th percentile) and versus the higher-ROE, higher-growth NU/MELI/DLO cohort — which deserve their premium. Cheap on contested earnings; fair on normalized earnings. (INTERPRETATION.)

Embedded-expectations math. Using the justified-multiple identity (P/B = (ROE − g)/(r − g)): at ~1.3x book on ~20% ROE with an EM cost of equity ~16–17% and terminal growth ~6–7%, the price is consistent with ROE holding ~19–20%. But if the whole Brazilian discount-rate complex falls as SELIC normalizes toward ~10.5% (lowering r toward ~14%), the same 1.3x book implies a through-cycle ROE of only ~14–15%. Conclusion: at ~7x clean earnings / 1.3x book, the market is underwriting a through-cycle ROE of ~13–15% — a ~25–35% decline from today’s ~20% — closely matching management’s own 29% cut to 2027 gross-profit guidance. The price is not extrapolating peak earnings; the cheap multiple reflects a contested, normalizing “E.” (INTERPRETATION/ASSUMPTION.)

The buyback accretion (central to the bull case). Repurchasing ~19% of the cap at ~7x earnings (~14% earnings yield) plus ~8–10%/yr share-count shrinkage delivers roughly 8–10% annual per-share return even with flat absolute earnings and no re-rating, demonstrated live in Q1-2026 (adjusted EPS +15.4% on +3.5% net income). At ~1x book, buybacks neither materially create nor destroy book value, so BVPS compounds at roughly the retained-ROE rate while the count falls. The risk: buying into a falling “E” destroys value if earnings normalize faster than the count shrinks. (FACT/INTERPRETATION.)

Scenario analysis (3-yr; embedded-expectations multiple ranges, NOT price targets):

Verdict (Valuation): The market prices normalization, not peak — so this is a bet on the degree of normalization and whether buybacks compound book faster than ROE decays. The bear outcome is largely already in the ~6th-percentile multiple; base/bull optionality (rate relief, election re-rate, contained credit) is arguably cheap, but it is optionality, not a safe margin. No price target stated.

11. Variant Perception

Consensus view. Sell-side is constructive — roughly 10 buy / 5 hold / 1 sell, average target ~$17–18 vs ~$10.6 spot (~65% implied upside) — pricing a manageable rate normalization plus buyback compounding (the base/bull). Yet ~12% short interest and Berkshire’s Q4-2023 exit signal a sharp institutional split. The analyst-target gap is the variant perception: targets are book-accretion/normalized-earnings DCFs; the market price embeds the normalization plus an EM/fiscal/FX discount the models under-weight. (FACT/INTERPRETATION.)

Strongest bull case. Deep value at ~6th-percentile P/S; a buyback compounder at ~1x book / ~7x earnings delivering ~10% with no re-rate; a known and guided rate headwind the price over-discounts (SELIC already easing); clean tangible book post-Linx; and Brazil re-rating optionality (election/fiscal relief) as a free call.

Strongest bear case. Rate-inflated peak earnings (~71% rate-sensitive; the 29% guidance cut is management conceding the cliff); a fast-growing credit book with rising NPLs and a poor history (2021); PIX structurally eroding the high-margin take rate; a value trap where the cheap multiple is efficient for a melting-ROE business; and EM/FX/election risk — with Berkshire, the canonical value buyer, already gone.

The 3–5 assumptions that matter most: (1) through-cycle ROE (13–15% bear vs 17%+ bull) — the whole thesis; (2) SELIC terminal level and speed; (3) credit cost-of-risk trajectory; (4) PIX impact on take rate; (5) buyback persistence at depressed multiples.

Falsification tests. Bull falsified if 2026–27 cost of risk rises toward 20%+/NPLs climb, OR take rate declines sequentially (PIX biting), OR buybacks slow. Bear falsified if ROE holds ≥16–17% through the rate trough AND take rate is flat/up despite PIX — proving earnings weren’t merely rate carry.

Our variant read. The honest synthesis is that STNE is cheap on reported/peak earnings and roughly fair on normalized earnings — a falling-knife-or-coiled-spring, not an obvious bargain. The edge, if any, is that the price discounts a lot of structural bad news while the buyback delivers a real ~10% even if nothing improves; the risk is that the same buyback compounds value destruction if the credit cycle and PIX bite faster than rates relieve.

12. Fact vs. Interpretation Table

13. Open Questions

1. Exact net-income sensitivity to SELIC (DV01-equivalent per 100bp) — not disclosed; the rate drag is directional only.
2. Blended take rate on PIX QR vs. card-credit and its trajectory as Pix Parcelado scales — the central industry variable.
3. Why ARPAC is falling ~11% YoY — pure TON/mix or genuine pricing erosion?
4. Credit underwriting quality post-relaunch (vintage NPLs, coverage adequacy) vs. the 2021 blow-up.
5. Banco Stone regulatory-capital ratio and headroom after the 20%→17% hurdle cut.
6. True current share count (yfinance ~229M vs balance-sheet ~273M) and class structure.
7. Stone–PagBank consolidation — recurring speculation, but no confirmed talks; would it restore pricing power or invite CADE scrutiny?

14. What Must Be True (Bull and Bear)

Bull case — what must be true:

• Through-cycle ROE holds the mid-teens or better as SELIC normalizes (earnings are not merely rate carry).
• Credit cost-of-risk stabilizes (~12–14%) rather than spiking; no 2021 redux.
• Take rate holds flat-to-up despite PIX; banking/deposit cross-sell offsets MDR erosion.
• Management keeps buying back stock at <~1.3x book, compounding per-share value; a partial Brazil re-rate (election/fiscal relief) is the free call.
• Falsification test: cost of risk toward 20%+/rising NPLs, OR sequential take-rate erosion, OR a slowed buyback — any breaks the bull case.

Bear case — what must be true:

• SELIC normalization compresses ~71%-of-revenue financial income faster than volume/funding offsets recover it; ROE sags to low-teens.
• The credit book mis-seasons into a soft SMB macro (NPLs >7%), forcing provisions — 2021 again.
• Pix Parcelado structurally migrates installment spend off card rails, permanently lowering the take rate.
• The cheap multiple proves efficient; the buyback compounds value destruction into a melting “E.”
• Falsification test: ROE holding ≥16–17% through the rate trough with flat/up take rate despite PIX — this breaks the bear case.

15. Source Appendix

See the Source Appendix (Appendix B) and Diligence Questionnaire (Appendix A) below. Primary sources: SEC EDGAR (StoneCo Form 20-F FY2024 and FY2025; 6-K earnings releases Q4-2024 through Q1-2026; CIK 0001745431); StoneCo investor relations (earnings releases, presentations, transcripts); Banco Central do Brasil (SELIC, PIX, client rankings); USTR/Federal Register (Section 301); and third-party data (stockanalysis.com peer pages, Trading Economics, MedTech/fintech trade press: MedTech Dive, NeoFeed, Brazil Journal, Rio Times, Payment Expert).

This is an independent analyst note for general information only and is not investment advice. No buy/sell recommendation and no price target appears in the analysis sections; the labeled “Claude’s Take” block above is an explicitly fenced exception.

APPENDIX A — Standard Diligence Questionnaire

Supplemental to the analysis above. Fact / Interpretation / Assumption labels applied where it matters. Frameworks: Greenwald “Competition Demystified” and Marathon “Capital Returns.” Where a question doesn’t map to a fintech/bank model, the correct sector analog is given. Figures IFRS/BRL; FX ~5.16 USD/BRL.

General

What thoughtful questions have other investors asked about this company? Five dominate: (1) Are today’s ~20% ROE and ~7x earnings sustainable, or are they a SELIC-rate-peak illusion? — the core debate, with embedded expectations already pricing a ~25–35% earnings normalization (INTERPRETATION). (2) Is PIX (especially Pix Parcelado) a structural killer of the card/prepayment profit pool, or a manageable, partly-recaptured shift? (3) Will the relaunched credit book repeat 2021? (4) Is the aggressive buyback at ~1x book intelligent compounding or capital destruction into a falling “E”? (5) Why did Berkshire — the canonical value investor — fully exit at end-2023? A sixth: can Stone defend its MSMB niche while Nubank/Mercado Pago win the customer relationship?

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Closer to a rate-driven cyclical high. ~71% of revenue is SELIC-sensitive financial income; SELIC at a ~15% peak (now easing) has inflated financial income. Management’s own 29% cut to 2027 gross-profit guidance signals earnings normalize down from here (FACT/INTERPRETATION).

Driven by the external environment or internal actions? Both. External: the SELIC cycle and Brazilian consumption drive the financial-income engine. Internal: the buyback (per-share growth), the credit relaunch, and the Linx exit are management choices. The recent profit recovery is more rate-and-buyback than franchise reacceleration (INTERPRETATION).

How stable are revenues? Moderately — recurring merchant payments/banking provide a base, but the profit is geared to rates and (increasingly) credit, both volatile. Headline revenue growth overstates franchise growth because economics are reclassified from MDR into rate-sensitive financial income (FACT).

Outlook for products/services? Payments: low-teens TPV, decelerating. Banking/deposits: the bright spot (+27%). Credit: high-growth, high-risk. Software: divested. Net: low-to-mid-single-digit underlying franchise growth, rate-dependent profit (INTERPRETATION).

How big will this market be — growing, shrinking, domestic or international? Brazilian card TPV is large and growing on penetration, but PIX caps the profit pool. Almost entirely domestic Brazil; no material international expansion. Growing volume, compressing per-unit economics (FACT/INTERPRETATION).

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More — super-apps (Nubank, Mercado Pago) are better-capitalized and winning users; incumbents (Cielo/Rede) are bank-backed; PIX is a state-built low-cost rail. Stone is losing the platform war while gaining narrow card share (FACT).

How profitable is the business (ROIC, ROE)? ROE ~20% (rate-flattered; through-cycle ~13–15%). ROIC is hard to clean given the bank-like balance sheet; ROE on tangible equity is the better lens and looks reasonable now that goodwill is nearly written off (FACT/ASSUMPTION).

How profitable is the industry — competitors, barriers to entry? A profitable-but-pressured multi-player market (Cielo ~25%, Rede ~22%, plus Stone, PagBank, Mercado Pago, GetNet, SafraPay). Barriers (licensing, scale, distribution, registradora) are real but lower than they look given well-funded super-app entry (INTERPRETATION).

Can the business be easily understood? Moderately — payments are simple, but the rate-geared financial-income engine and bank-like balance sheet require care. Many treat it as a “payments compounder,” which mis-frames the rate dependence (INTERPRETATION).

Can it be undermined by foreign low-cost labor? Not directly — it is a domestic, regulated financial-services business. The real threat is domestic super-apps and a state payment rail, not offshore labor (INTERPRETATION).

Do brands matter? Somewhat — Stone (SMB) and TON (micro) have brand recognition and a respected service reputation, but it is not a decisive moat versus Nubank’s consumer brand (INTERPRETATION).

What is the nature of competition? Take rate (MDR), receivables-prepayment pricing, credit terms, banking/app experience, and distribution/service. Increasingly price- and ecosystem-led (FACT).

Customers’ switching costs? Moderate via the payments-banking-credit bundle, deposits, and hub service relationships (~40% multi-product), but low at the micro/TON end and erodable as super-apps bundle more cheaply. A localized scale+captivity (Greenwald) moat, narrow (INTERPRETATION).

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The hub distribution network, merchant relationships, and the banking licence/deposit franchise are intangible value not fully capitalized. Conversely, goodwill has been over-recognized historically (Linx) and is now nearly written off (FACT).

Off-balance-sheet liabilities? Standard for a fintech with a credit book — watch receivables-financing structures (FIDCs) and prepayment funding. The ~R$8.8B “long-term debt” is operating funding for interest-earning assets, not corporate leverage (FACT).

How conservative is the accounting? Reasonably — IFRS, and the FY2025 adjusted-vs-IFRS gap is small and defensible (unlike many fintechs). The credibility issues are historical (2021 credit, 2024 impairment) and prospective (rate normalization), not aggressive add-backs (FACT/INTERPRETATION).

How CapEx-hungry is the business? Modestly — capex ~R$0.7–1.3B/yr; the capital intensity is in the balance sheet (funding the credit/prepayment book and regulatory capital), not physical capex (FACT).

Capital Allocation & Management

How much free cash does it generate, and how is it used? Strong earnings (R$2.3B net 2025) and aggressive return: ~R$2.9B buybacks (2025), a regular dividend, and a $2.53/share special dividend. Philosophy: return excess capital absent value-accretive opportunities — a correction from the empire-building Linx era (FACT).

Significant acquisitions recently? The Linx acquisition (~R$6.7B, 2021) was a value-destroyer (~R$3.5B impaired/lost), now divested to TOTVS. The current posture is divest-and-return, not acquire (FACT).

Buying back shares? Yes, aggressively — ~19% of market cap in 2025 at ~1x book/~7x earnings; share count down 319M→273M. Rationally accretive if mid-teens-plus ROE holds (FACT/INTERPRETATION).

Issuing large amounts of new shares to insiders? SBC is modest and more than offset by buybacks; the share count is falling, not rising (FACT).

Compensation / motivations of management? CEO Pedro Zinner (non-founder, 2023); André Street chairman; insider ownership low (~2.8%). Incentive detail in the 20-F; the pivot to capital return suggests shareholder-aligned current priorities, against a scarred M&A record (FACT/INTERPRETATION).

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer? Neither MLP nor K-1. STNE is a Class A ordinary share of a Cayman-incorporated foreign private issuer listed directly on NASDAQ (not a traditional ADR); files 20-F/6-K, reports IFRS in BRL. USD-listed price carries BRL/FX translation risk (FACT).

Dividend policy? A regular dividend was initiated (~2% yield, ~27% payout) plus a one-time $2.53/share special dividend from Linx proceeds (FACT).

How profitable is the business? ~20% ROE, ~17% net margin (rate-flattered); through-cycle profitability is the contested variable (FACT/INTERPRETATION).

Is net income diverging from cash from operations? For a bank-like model, OCF is distorted by changes in the receivables/credit book and is less meaningful than for an industrial. Earnings quality is currently clean; the key divergence to watch is credit provisions vs. cash losses if NPLs rise (INTERPRETATION).

Risks & Downside

What factors would cause the stock to decline? Faster SELIC normalization compressing financial income; a credit accident (rising NPLs/cost of risk); accelerating PIX disintermediation; share loss to super-apps; an adverse election/fiscal/BRL shock; a slowed buyback (INTERPRETATION).

Risk of a catastrophic loss? Low in the solvency sense — profitable, regulated bank with adequate capital, net equity, cash-generative. The realistic severe case is a value-trap drawdown (earnings normalize down, multiple stays low, BRL weakens), not insolvency (INTERPRETATION).

Chance of a total loss? Very low — a profitable franchise with a regulated balance sheet and aggressive capital return makes permanent total loss highly improbable absent an extreme systemic/sovereign event (INTERPRETATION).

Recent News & Events

Has the business environment changed recently? Yes: (1) SELIC began easing (15%→14.50%), starting the financial-income normalization; (2) the 2027 guidance was cut ~29% and KPI guidance withdrawn; (3) Linx was sold to TOTVS (closed Feb 2026), funding a special dividend; (4) PIX extended to installments (Pix Parcelado) and contactless; (5) the USTR Section 301 PIX probe; (6) the October 2026 election looms (FACT).

Significant acquisitions? No new ones — the story is the divestiture of Linx (FACT).

Change in accounting policies? None material flagged; the major item is the goodwill impairment/divestiture of Linx and reclassification of it to discontinued operations (FACT).

Recent changes — new markets, facilities, management? CEO transition to Pedro Zinner (2023); “one-Stone” holding-structure simplification; banking/deposit push (R$11.1B, +27%); credit relaunch; core equity-ratio hurdle cut 20%→17%. No material geographic expansion beyond Brazil (FACT).

APPENDIX B — Source Appendix

*All URLs accessed 2026-06-08 unless noted. Primary sources prioritized. StoneCo is a foreign private issuer (CIK 0001745431) filing Form 20-F (annual) and 6-K (interim) under IFRS in BRL. *

Primary — SEC Filings & Company Disclosure (CIK 0001745431)

• StoneCo Form 20-F FY2024 — https://www.sec.gov/Archives/edgar/data/0001745431/000162828025019653/stne-20241231.htm
• StoneCo Form 20-F FY2025 (filed Apr 23, 2026) — https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001745431&type=20-F
• StoneCo Q4-2025 / FY2025 earnings release (6-K) — https://www.sec.gov/Archives/edgar/data/0001745431/000207097926000026/earningsrelease4q25.htm
• StoneCo Q1-2026 earnings release (6-K, May 14 2026) — https://www.sec.gov/Archives/edgar/data/0001745431/000207097926000265/earningsrelease1q26.htm
• StoneCo FY2024 6-K — Linx goodwill impairment R$3,558.0M — https://www.sec.gov/Archives/edgar/data/0001745431/000129281425000956/ex99-1.htm
• StoneCo special dividend ($2.53/share) 6-K filings — https://www.stocktitan.net/news/STNE/
• StoneCo investor relations — https://investors.stone.co/
• StoneCo Q4-2025 earnings call transcript — https://www.fool.com/earnings/call-transcripts/2026/03/02/stoneco-stne-q4-2025-earnings-call-transcript/

Primary — Brazilian Macro & Regulatory

• Banco Central do Brasil — SELIC / Copom decisions — https://www.bcb.gov.br/en/monetarypolicy
• Brazil interest rate (SELIC path) — Trading Economics — https://tradingeconomics.com/brazil/interest-rate
• Brazil starts easing cycle (SELIC to 14.75%) — MercoPress — https://en.mercopress.com/2026/03/18/brazil-s-central-bank-starts-easing-cycle-cuts-selic-rate-to-14.75
• Pix Parcelado (installments) rules — Payment Expert — https://paymentexpert.com/2025/10/07/brazils-race-to-standardise-pix-parcelado-for-further-instant-payment-growth/
• PIX statistics (Banco Central) — paymentscmi.com — https://paymentscmi.com/insights/pix-in-brazil-latest-statistics-central-bank/
• USTR Section 301 investigation (PIX) — https://ustr.gov/ ; Federal Register determination (Jun 4, 2026) — https://www.federalregister.gov/
• USTR PIX “unreasonable” / Section 301 background — Rest of World — https://restofworld.org/2025/pix-brazil-us-investigation-digital-payments/
• 2026 Brazilian general election — https://en.wikipedia.org/wiki/2026_Brazilian_general_election

Secondary — Trade Press & Analysis

• StoneCo to sell majority of software (Linx) to TOTVS for ~R$3.19B — Investing.com — https://www.investing.com/news/company-news/stoneco-to-sell-majority-of-software-segment-for-r319-billion-93CH-4145574
• Linx/TOTVS deal (R$3.05B EV, ~79% of software revenue) — IndexBox — https://www.indexbox.io/
• StoneCo Q4 Linx impairment R$3,558M — Seeking Alpha — https://seekingalpha.com/news/4559744-stoneco-stock-dips-on-2025-adjusted-gross-profit-falling-below-expectations-q4-linx-impairment
• StoneCo slashes 2027 guidance ~29% — Rio Times — https://www.riotimesonline.com/stoneco-misses-estimates-slashes-2027-guidance-by-29/
• StoneCo Q1-2026 margin pressure / buyback EPS accretion — TipRanks — https://www.tipranks.com/news/company-announcements/stoneco-q1-2026-results-show-margin-pressure-as-it-returns-capital-after-linx-sale
• Mercado Pago overtakes Stone in acquiring profit (Itaú BBA) — NeoFeed — https://neofeed.com.br/negocios/mercado-pago-avanca-e-ja-supera-lucro-de-stone-e-pagbank-em-adquirencia-diz-itau-bba/
• Brazil banks customer ranking (Nubank/Mercado Pago lead; Stone declines) — Brazil Stock Guide — https://brazilstockguide.com/insights/brazil-banks-customer-ranking-2026/
• Cielo take-private by BB/Bradesco — Reuters via Yahoo — https://uk.finance.yahoo.com/news/controlling-shareholders-bid-brazil-payments-004816948.html
• Brazilian payments competitive landscape (Cielo/Stone/Linx/PagSeguro) — Brazil Journal — https://braziljournal.com/cielo-stone-linx-e-pagseguro-quem-vai-liderar/
• UBS Evidence Lab — Brazilian Payments (MDR repricing) — https://www.ubs.com/global/en/investment-bank/insights-and-data/2024/brazilian-payments.html
• Berkshire Hathaway sheds StoneCo (Q4 2023 exit) — Seeking Alpha — https://seekingalpha.com/news/4067238

Valuation / Comps / Market Data

• StoneCo statistics & forecast — stockanalysis.com — https://stockanalysis.com/stocks/stne/statistics/ ; https://stockanalysis.com/stocks/stne/forecast/
• PagSeguro (PAGS) statistics — https://stockanalysis.com/stocks/pags/statistics/
• Nu Holdings (NU) — https://stockanalysis.com/stocks/nu/
• MercadoLibre (MELI) — https://stockanalysis.com/stocks/meli/statistics/
• dLocal (DLO) — https://stockanalysis.com/stocks/dlo/statistics/
• StoneCo consensus / target — MarketScreener — https://www.marketscreener.com/quote/stock/STONECO-LTD-46772609/consensus/

Quantitative Data (reconciled to filings)

• yfinance — price, market cap, share count, multiples (unofficial; reconciled to SEC filings)
• FX: USD/BRL ~5.16 (June 2026), via yfinance