MercadoLibre, Inc. (NASDAQ: MELI) — The Continent’s Best Franchise, De-Rated for a Margin Dip It Chose and a Loan Book It Hasn’t Cycled

Independent Equity Research — MercadoLibre, Inc. Report date: 2026-06-11 · Price (2026-06-10): $1,588.29 · Market cap: ~$80.5B · Enterprise value: ~$87B · Shares out: ~50.7M Fiscal year: December · Filer status: US-domestic (10-K/10-Q, US GAAP, reports in USD) despite Argentine origin / pan-LatAm operations · CIK: 0001099590 · HQ: Montevideo, Uruguay (incorporated Delaware, 1999)

⚡ Claude’s Take

This block is the author’s own subjective opinion. It is general information and not investment advice. The analysis that follows takes no position, carries no price target, and remains recommendation-free throughout.

Verdict: HOLD / accumulate-on-weakness. Medium conviction. Not-a-short. Best risk/reward in the ~$1,300–1,550 zone (roughly 28–33× a compressed forward EPS, i.e. near where the stock trades today and just below); back up the truck only on a credit-driven break toward the high-$1,000s–$1,200s. Above ~$2,000 the compound bet is fully priced and I’d trim, not add.

MercadoLibre is the best operating franchise in Latin America and one of the better growth franchises on any exchange: a dominant, vertically-integrated marketplace (1P + 3P + owned logistics + a fast-scaling retail-media business) welded to the region’s leading commerce-native fintech (Mercado Pago — payments, a $14.6B credit book, ~$19B of AUM, a credible “largest digital bank in LatAm” ambition). It has now strung together 28 consecutive quarters of >30% revenue growth — a duration record at this scale — and the most recent quarter (Q1-2026, +49%) was its fastest growth in four years. The stock has nonetheless fallen ~40% from its November-2025 high to within a few percent of its 52-week low. The selloff is almost entirely a margin story, not a demand story: the company deliberately gave up ~5–6 points of operating margin in 2025 to slash Brazil’s free-shipping threshold (to BRL 19), scale 1P, accelerate credit cards, and invest in Mexico — and that choice became frightening in Q1-2026 when operating income fell −19.9% year-over-year even as revenue grew 49% and loan-loss provisions doubled. The cleanest tell of the dislocation: price-to-sales sits at the 22nd percentile of MELI’s own ten-year history (cheap) while price-to-earnings sits at the 74th (rich) — a contradiction that can only resolve if the margin compression is temporary.

My read is that this is a quality compounder in a self-inflicted air-pocket, which is a buyable setup but not a no-brainer, because two things are genuinely true at once. First, the bull’s core fact is unimpeachable: revenue never broke — unit metrics (items sold, buyers) are accelerating, the free-shipping flywheel is producing real unit-cost deflation (Brazil cost-per-shipment −17% while absorbing +56% volume), and ~80% of the scary FY2025 net-income stall is a transitory Argentine-disinflation tax artifact that will roll off. Second, the bear’s facts are quantitative and recent: the credit book doubled in a year, is now 45% credit cards (the lowest-margin, longest-duration, highest-loss product), its risk-adjusted spread (NIMAL) has halved to ~17.8%, off-balance-sheet card commitments have exploded to ~$12B against a ~$36M reserve, and the whole thing is un-seasoned through a real LatAm credit downturn — while Shopee, now #1 by orders in Brazil and EBITDA-positive, removes the old “competitors will run out of money” comfort. At ~$87B EV the market prices the base case (margins recover toward the mid-teens, credit seasons benignly, the premium multiple holds) at its midpoint, leaving no margin of safety if any leg breaks. The framing is quality-at-a-fair-(no-longer-stretched)-price after a sentiment de-rate — a contrarian lean, not a deep-value layup. The single fact that pushes me to “accumulate” rather than “avoid” is the director open-market buy — Alejandro Aguzin bought ~$1.0M of stock at ~$1,655 in a discretionary (non-10b5-1) trade right into the drawdown, the first such conviction signal in years from a company whose cash-settled comp means insiders rarely transact at all.

Conviction: medium. Flip-bullish trigger: two-to-three quarters of operating-margin inflection back toward the mid-teens with growth intact and credit NIMAL stabilizing — that vindicates the deliberate-investment thesis and both the “E” and the multiple re-rate up together. Flip-bearish trigger: NIMAL keeps compressing below ~15% while provisions outrun book growth and Brazil commerce take-rate/ad-load declines sequentially — that would mean the margin dip is structural (Shopee/Amazon forcing price) and the credit book is deteriorating, not seasoning, in which case ~42× trailing on peak ROE still has real downside. Catchy version: the continent’s best house, briefly on sale because management is re-roofing it and lending to the neighbors.

1. Executive Summary

MercadoLibre operates two deeply-integrated engines across 18 Latin American countries: Commerce (the Mercado Libre marketplace — predominantly third-party, with selective 1P; Mercado Envíos owned logistics; Mercado Ads retail media; classifieds), at ~56% of revenue, and Fintech (Mercado Pago — payments/acquiring, the Mercado Crédito lending book, Mercado Fondo asset management, a digital account/neobank, insurtech), at ~44%. FY2025 total net revenue (including financial income) was $28,893M, +39.1%; GMV was $65.0B (+26%) and total payment volume $277.8B (+41%). The business is a genuine compounder — 121M annual unique buyers, 78M fintech monthly actives, no customer above 5% of revenue — and it reinvests virtually all of its cash flow back into the moat rather than returning it (no dividend, no buyback; capex up 56% to $1.34B; ~$6.7B of net new lending in the year).

The investment debate is entirely about margins and credit, not demand. Management consciously compressed operating margin (11.1% in FY2025, down from 12.7%, and ~6.9% in Q1-2026) to fund a land-grab — cutting Brazil’s free-shipping threshold, scaling lower-margin 1P retail, accelerating credit cards across Brazil/Mexico/Argentina, and stepping up Mexico (a $4.6B 2026 commitment). The reported FY2025 net-income “stall” (+4.5% on +39% revenue) is ~80% a transitory artifact of Argentine disinflation lifting the effective tax rate from 21.4% to 29.7%; pretax income actually grew +16.9%. But Q1-2026’s −19.9% operating-income decline (provisions +106%) is real operating compression independent of tax, and it is what triggered the ~40% drawdown and a JPMorgan price-target cut.

The moat is wide but contested. The durable edge is Greenwald-style economies of scale in owned logistics (>50% of shipments from fulfillment centers, >75% of items delivered within two days) plus customer captivity (the Meli+ loyalty lock, deposit/credit “principality,” and high seller switching costs). Two-sided network effects are real but the weakest, most replicable pillar — Shopee built the #1 order-volume position in Brazil from a standing start in five years and is now e-commerce-EBITDA positive, and Nubank out-scales MELI in consumer banking. The fintech flywheel (commerce data → underwriting) is a genuine weapon but imports LatAm consumer-credit-cycle risk into what was once an asset-light fee business.

Quality of earnings is amber, not red. Coverage on the loan book is stable (~25–26%), the accounting is conservative (cash-settled comp, no dilution, reserve-building ahead of write-offs, no one-time gains games), and gross liquidity is strong (~$8.1B ex-restricted). But headline operating cash flow ($12.1B) overstates owner earnings ~3× — normalized free cash flow to equity is ~$4.1B once you strip deposit float and loan-book funding — and the single largest forward risk is a credit downturn hitting a doubled, card-heavy book at a halved NIMAL.

Valuation prices the optimistic-but-not-heroic outcome. Sum-of-the-parts (~$73–102B) brackets the ~$87B EV — no hidden discount. A reverse-DCF implies the market is underwriting ~20–25% revenue CAGR and a 300–500bp margin recovery and benign credit seasoning and a sustained premium multiple — a compound bet with little cushion. This memo takes no position; it lays out what must be true for each side and the falsification tests that will settle it.

2. Business Overview

MercadoLibre is the leading e-commerce and fintech ecosystem in Latin America — the regional commerce leader by GMV and the fintech leader by monthly active users in Argentina, Chile and Mexico (and #2 in Brazil, behind Nubank). The commerce platform operates in 18 countries; the Mercado Pago fintech platform in eight (Argentina, Brazil, Mexico, Chile, Colombia, Peru, Uruguay, Ecuador). The company reports two revenue streams (Commerce, Fintech) across four geographic segments (Brazil, Mexico, Argentina, Other). (FACT — FY2025 10-K, Item 1.) The strategic logic is an integrated flywheel: commerce generates payment volume and proprietary behavioral/cash-flow data; that data underwrites credit and deepens fintech engagement (“principality”); fintech and logistics in turn lower friction and raise commerce frequency.

Commerce segment — FY2025 revenue $16,294M (+34%), 56.4% of total

• Mercado Libre Marketplace — predominantly 3P (third-party sellers are most of GMV); 1P is <10% of GMV, used selectively to plug price/assortment gaps (notably heavy/bulky goods, including a Casas Bahia white-goods partnership). Monetized via final-value fees (a percentage of sale value, with the payment embedded) plus flat fees on low-value items. (FACT — 10-K Item 1.)
• Mercado Envíos (logistics) — fulfillment centers handling >50% of shipments, cross-docking, and a “MELI Places” network of thousands of partner drop-off/pickup points, served by dedicated aircraft, trucks and last-mile vans (mostly third-party-operated). >75% of items are delivered within two days region-wide. The network subsidizes free/discounted shipping to drive frequency — the Brazil free-shipping threshold was cut from BRL 79 to BRL 19 in mid-2025, the fourth such cut since 2017. (FACT — 10-K; Goldman Communacopia 2025-09-10.)
• Mercado Ads — product, brand, display and video advertising, on- and off-platform (Google Ad Manager, Disney, Roku, HBO Max integrations). A ~$1B+ run-rate growing ~67% (Q4-2025), but still only ~2% of GMV versus high-single-digit % at mature retail-media peers — the single largest margin-accretive runway in the company. (FACT — Q4-2025 call; Communacopia 2025.)
• Classifieds (vehicles, real estate, services — placement fees only), Meli+ loyalty (tiered bundles of Disney+/Netflix/HBO Max/Apple TV+ with shipping and fintech perks), and Mercado Play (ad-supported streaming). Mercado Shops (the storefront tool) was discontinued Dec 31, 2025, folded into “Mi Página” within the Marketplace. (FACT — 10-K.)

Fintech segment — FY2025 revenue $12,599M (+46%), 43.6% of total

• Mercado Pago payments/acquiring — online checkout, QR codes, and mPOS devices for offline merchants; acquiring TPV $188,105M in FY2025. On-marketplace payment processing is bundled into the commerce fee (no separate charge); off-platform volume earns MDR/commissions plus installment-financing fees. (FACT — 10-K.)
• Mercado Crédito (lending) — merchant loans (repaid directly from sellers’ own marketplace cash flows, structurally low-loss), consumer loans, credit cards (Brazil 2021, Mexico 2023, Argentina 2025), and Brazilian asset-backed auto loans. Gross portfolio $12.5B at FY2025 → $14.6B by Q1-2026, with proprietary ML underwriting on ecosystem data as the stated edge. (FACT — 10-K Note 5; Q1-2026 call.)
• Mercado Pago digital account / neobank — remunerated balances (paying more than bank checking), debit/prepaid/credit cards, transfers, bill-pay; the stated ambition is “Latin America’s largest digital bank.” A Mexican banking license is in process. (FACT — Q1-2026 call.)
• Mercado Fondo (asset management) — a remunerated money-market product with AUM ~$19B (+78%); a deposit-gathering and principality engine that funds the credit book. Plus insurtech, crypto (buy/hold/sell in BR/MX/CL), and the “Meli Dólar” USD stablecoin. (FACT — 10-K; Q4-2025 call.)

Revenue quality, geography, and KPIs

Revenue is transactional rather than contractual-recurring (no SaaS-style contracts) but quasi-recurring through habit and frequency — 121M annual unique buyers, 78M fintech monthly actives, rising purchase frequency, and a loyalty program explicitly engineered to convert users into “ecosystemic” members who both shop and bank. Customer concentration is negligible (no customer >5% of revenue). Geographic revenue mix (FY2025): Brazil $15,201M (53%), Mexico $6,475M (22%), Argentina $5,962M (21%), Other $1,255M (4%) — and Argentina is fintech-heavy, which makes consolidated fintech results sensitive to Argentine rates and FX. (FACT — 10-K MD&A.) Of the FY2025 total, ~30% ($8,558M) is financial/interest income from the credit book and float — lower-quality, rate- and credit-sensitive revenue that does not deserve a marketplace multiple.

** Verdict.** A genuinely two-engine compounder — a dominant 3P marketplace wrapped in owned logistics, retail media, and a fast-scaling neobank, with deep designed-in synergies. The model is heavier and more capital-intensive than a pure marketplace (owned fulfillment plus a $14.6B on-balance-sheet credit book funded partly by deposits), which is the source of both the moat and the margin volatility. High-quality and structurally advantaged — but now partly a lender, which imports credit-cycle risk into the story.

3. Industry Dynamics

A long runway at low penetration. Latin American online retail is roughly 12–15% of total retail versus ~16% in the US and ~47% in China; the region’s 650M+ people remain structurally under-penetrated, and management consistently frames its own GMV as “very small versus total retail.” (FACT — eMarketer/Statista 2025; Q2-2025 call.) This is the single most important industry fact: the pie is still small and compounding double-digits, which lets the leader gain share without the zero-sum knife-fight of a mature market — and which justifies MELI’s “invest, don’t optimize margin” posture. Brazil is ~45% of regional e-commerce, Mexico ~26%.

An ecosystem oligopoly forming over a contested field. The market is consolidating around a handful of integrated ecosystems — MercadoLibre, Amazon, Shopee (Sea Limited), Magazine Luiza/Magalu, Casas Bahia/Via, Americanas — with cross-border entrants (Shein, AliExpress, Temu) adding a cheap, slow-logistics tier. The most important competitive fact of the last two years is that Shopee is now #1 by order volume in Brazil and turned its e-commerce segment EBITDA-positive (Sea group e-commerce adjusted EBITDA ~$228M in Q2-2025), removing the “they’ll run out of money” comfort that long underpinned the MELI bull case. (FACT — Sea Q2-2025; Fortune 2025.) Shopee competes on price, gamified engagement and a long-tail/cross-border catalog, and is structurally lighter on owned logistics than MELI — its cost advantage and its weakness. Amazon is entrenched (strong in Mexico, #3 in Brazil) but lacks MELI’s owned last-mile density and embedded fintech in most LatAm geographies. The weakened local incumbents are increasingly distribution on MELI rather than against it — telling signal: Casas Bahia became a 3P seller on MELI in 2025.

Fintech is the harder battle. In consumer fintech, Nubank is the dominant rival — ~105M+ Brazil customers (131M+ region-wide), the world’s largest digital bank ex-China — and Mercado Pago is #2 by MAU in Brazil. In merchant acquiring, StoneCo and PagSeguro compete hard (both repricing-pressured), and Mercado Pago is gaining acquiring share (Brazil TPV +25%, Mexico +50% in Q4-2025). PicPay, the incumbent banks, and the cross-border PSP dLocal round out a crowded field. (FACT — Nubank/STNE/PAGS disclosures 2025-26; Q4-2025 call.)

The Pix question (central). Brazil’s central-bank instant-payment rail — free, real-time, ubiquitous (~42% of Brazilian e-commerce payment value) — is both threat and tailwind. It commoditizes the low-value payments/transfer layer that was Mercado Pago’s original wedge (removing float and basic MDR, and letting Nubank/PicPay/banks offer the same rails), but it has also accelerated digital-account adoption across the unbanked, expanding Mercado Pago’s user base and the deposit pool (AUM +78%) that funds high-margin credit. MELI’s monetization has migrated up-stack — to lending spread, asset management, credit-card interchange, and acquiring MDR on merchant volume — areas Pix does not commoditize, though interchange itself is now under regulatory scrutiny (Mexico interchange caps were raised as an open item, later deferred). Net: Pix commoditizes the rail but, on balance, enables the banking franchise. (FACT/INTERPRETATION — Q4-2025 call; BCB data.)

Other regulation and macro. Fintech licensing/AML across seven countries; Brazil’s LGPD privacy regime; and — most consequentially — Argentina’s FX controls and hyperinflation (USD functional currency, repatriation friction, rate-driven funding-cost swings that whipsaw credit margins). Argentina is simultaneously MELI’s highest-margin market and its largest macro tail.

Marathon capital-cycle lens. The classic warning — high returns attract capital, supply floods in, returns mean-revert — is visibly underway: Sea is pouring capital into Brazil logistics, Amazon keeps investing, Nubank is scaling lending, cross-border players are dumping cheap supply. The mitigant is that the TAM is expanding faster than capital can compete it away (12–15% vs 47% China), so this is a growth capital cycle, not yet a mature/replacement one. The fingerprints to watch are on MELI’s own supply side: capex up from $509M (2023) to $1,343M (2025), heavy fulfillment build, and a NIMAL compressing from 36% to ~18% — early evidence that competition is eroding marginal returns even as absolute growth stays spectacular.

** Verdict — structurally attractive but intensely contested.** The decisive feature is under-penetration: a still-small, double-digit-growth pie lets the scaled, vertically-integrated leader gain share without a mature-market brawl — genuinely good, and punishing for sub-scale players (witness Americanas’ fraud/bankruptcy and Via’s restructuring). But this is not an oligopoly at rest — Shopee is now #1 by orders and profitable, Amazon is entrenched, and fintech is a true fight with Nubank. The profit pool is real but is being competed for with aggressive capital, and returns at the margin are under pressure even as the top line compounds.

4. Competitive Position

Moat type (Greenwald taxonomy): primarily economies of scale + customer captivity, reinforced by — but not reliant on — network effects. MELI is a textbook case where the dominant advantage is localized scale economies (owned logistics density and a fixed-cost tech/credit-modeling base spread over the largest GMV/TPV in each market) combined with customer captivity (habit, the Meli+ loyalty lock, the principality of the digital account, and seller switching costs). Pure two-sided network effects are present but the weakest, most replicable pillar; the durable edge is the capital-intensive, locally-dense infrastructure a new entrant cannot cheaply replicate.

Pressure-testing each pillar

1. Logistics scale (strongest). This is the real moat. Density is self-reinforcing: every incremental package lowers cost-per-delivery and enables a lower free-shipping threshold, which drives frequency and more packages. The proof is in the unit economics — Brazil cost-per-shipment fell 8% → 11% → 17% YoY across Q3-2025 to Q1-2026 while absorbing 56% volume growth, powered by volume density plus a new low-cost “slow-shipping” layer that monetizes idle fulfillment/line-haul capacity. (FACT — Q1-2026 call.) This is the pillar Shopee most conspicuously lacks (it leads on orders but is racing to build Brazil fulfillment). It ties directly to a financial outcome (unit shipping cost), so it qualifies as a genuine Greenwald cost-advantage moat — one that the $1.34B capex is the price of widening.

2. The fintech flywheel / data advantage (high, but cyclical). Commerce generates proprietary behavioral and cash-flow data; merchant loans are repaid directly from sellers’ marketplace receivables (“meaningfully reducing uncollectability”); credit cards are originated inside the buying flow at near-zero acquisition cost (2.7M issued in Q1-2026 alone, most from the marketplace). This is a real underwriting edge banks lack — but it imports credit-cycle risk and is visibly straining: NIMAL fell 36.2% → 28.2% → 22.4% → 17.8% over three years and a quarter, allowances are $3.8B (~26% of gross book), and provisions are rising fast. A weapon, not a free lunch.

3. Network effects (moderate — the replicable pillar). Two-sided liquidity (121M buyers ↔ millions of sellers) is real but contestable: Shopee built #1 order volume in ~5 years from zero, proving the matching function alone is not impregnable. What makes MELI’s network sticky is what’s bolted onto it (logistics + payments + credit + loyalty), not the matching itself.

4. Brand / trust (moderate-high). #1 brand preference and #1 NPS in commerce and fintech across Brazil/Mexico/Argentina/Chile, plus a buyer-protection program. In a region with low institutional trust and high fraud, brand is more durable than in developed markets.

5. Switching costs (high for sellers, moderate for buyers). Sellers integrated into MELI fulfillment, ads, payments and working-capital credit face real friction (inventory in MELI warehouses, ad campaigns, reputation/reviews). Buyers stickier once “ecosystemic” (Meli+, remunerated balance, credit card) — the loyalty program is explicitly designed to be “very hard to replicate because [rivals] don’t have the two sides of it.”

Greenwald tests and head-to-head

• Share stability: MELI’s commerce share is rising, not merely stable (management says it doubled in five years, +4pp last year) — but it is gaining by spending (margin down 5–6 points), which under Greenwald is the signature of a contested moat being actively defended, not a placid dominant one. Fintech share is genuinely contested (#2 to Nubank in Brazil). (INTERPRETATION; share figures are management-sourced, OPEN QUESTION on independent reconciliation.)
• ROIC: returns on the commerce franchise (asset-light fees + ads) remain high, but the consolidated business now carries a $14.6B credit book plus $2.3B PP&E and $2.2B leases, diluting headline returns and adding cyclicality. The live debate is incremental ROIC — is the capex/credit/shipping investment earning its cost of capital? NIMAL compression and rising capex say “watch this closely.”
• vs Amazon (BR/MX): MELI wins on owned last-mile density and embedded fintech; Amazon wins on global selection and balance-sheet firepower. Local specialization is the durable edge.
• vs Shopee/Sea: the genuinely two-way fight. Shopee wins on price, gamified engagement, long-tail/cross-border supply, and a capital-lighter model now generating e-commerce EBITDA — and took #1 order volume in Brazil. MELI wins on logistics control, delivery speed, higher-AOV categories, and fintech.
• vs Nubank (fintech): Nubank wins on consumer-banking scale (105M+ Brazil), brand, and a deposit-funded lending machine; MELI wins on the commerce-to-credit data loop and merchant acquiring. In pure consumer neobanking MELI is the challenger.

** Verdict — a durable, multi-pillar moat, real and widening in commerce, strong but contested and cyclical in fintech.** This is not “a crowded market with weak differentiation”: MELI’s vertical integration (marketplace + owned logistics + payments + credit + ads + loyalty) is a combination no single rival matches, and that integration is the moat. But it is being actively defended with capital (margin down 5–6pts, capex up 56%, NIMAL down ~18pts over three years), Shopee has proven the marketplace pillar is contestable and is now profitable, and Nubank out-scales the consumer-banking ambition. The honest characterization: a wide but not impregnable moat around a leader that must keep spending to stay ahead — strongest in logistics, thinnest where payments are commoditized by Pix and where the credit book exposes it to the LatAm consumer-credit cycle.

5. Growth History and Forward Opportunities

The multi-year record — and how much is real

Net revenue compounded $15,107M (2023) → $20,777M (2024, +37.5%) → $28,893M (2025, +39.1%), with the quarterly cadence accelerating through 2025 rather than fading: Q3 +39%, Q4 +45%, then Q1-2026 +49% YoY — the fastest since Q2-2022 and the 28th consecutive quarter above 30%. (FACT — 10-K; Q1-2026 call.) The physical KPIs corroborate that this is volume, not price: GMV $65.0B (+26%), items sold 2,429M (+36%), unique active buyers 121M (+21%), TPV $277.8B (+41%), fintech MAUs 78M (+28%).

Organic vs FX/inflation. The composition is high-quality: items sold and unique buyers are FX-immune unit metrics, and they are accelerating, the cleanest evidence that growth is organic demand. The distortion is concentrated in Argentina (~21% of revenue) — Q3-2025 Argentine revenue grew 39% in USD but 97% in local currency, the gap being peso devaluation. As Argentina disinflates (211% → 117% → ~31%), that inflation tailwind mechanically fades, so the FX-clean acceleration is being carried by Brazil (53%) and Mexico (22%) on unit metrics — the strong signal. (FACT — Q3-2025 call.)

Commerce growth drivers

• Marketplace share gains — management states Brazil share doubled over five years (+~4pp last year), against still-low ~15% e-commerce penetration. (Mgmt claim; OPEN QUESTION on independent verification.)
• The free-shipping flywheel + the BRL 19 cut — the June-2025 threshold cut (from BRL 79) is the most consequential commercial decision of the period. Brazil items-sold growth went 26% → 42% → 45% → 56% across the four quarters; conversion +1pt; record retention for both new and existing cohorts; new post-cut cohorts buy more items across more categories with higher retention. Several lower-price brackets are “already breaking even,” tracking the original 2016-17 free-shipping rollout “even faster.” (FACT — Q4-2025/Q1-2026 calls.)
• Targeted seller take-rate cuts (Brazil, 2024 onward, conditional on the seller keeping competitive pricing) — these hit the P&L from Q2-2026, so further near-term margin drag is queued.
• 1P expansion — scaling into heavy/bulky selection gaps (Casas Bahia white goods); profitable on a variable basis but a net margin drag while scaling.
• Advertising — the highest-margin recovery lever: +67% (Q4-2025) but only ~2% of GMV vs high-single-digit peers, >80% search, with display/video doubling off a low base and off-platform reach via Google/Disney/Roku/HBO Max.
• Meli+ loyalty and Mercado Play — engagement and ad-inventory drivers that also tie the two sides together.

Fintech growth drivers

TPV +41%, with acquiring share gains (Brazil +25%, Mexico +50%). The credit book grew ~87–91% YoY to $14.6B, led by credit cards (TPV +90%, MAU +68%, 2.7M issued in Q1-2026; Brazil cohorts older than two years are NIMAL-positive, ~half the Brazil book is NIMAL-positive, Mexico ~2 years behind, Argentina just launched). AUM ~$19B (+78%) via remunerated “super-yielding” accounts that drive deposits and principality (≥50% of a user’s income flowing through Pago — up 11 points in Brazil). A measurable cross-sell flywheel (many new cardholders were marketplace-only users; cardholders show higher spend, retention and NPS), plus insurtech, the Meli Dólar stablecoin, and an AI assistant resolving 87% of Pago interactions. (FACT — Q4-2025/Q1-2026 calls.)

Forward opportunities

• Mexico — the #2 market, now backed by a $4.6B 2026 investment (announced 2026-06-08); already #2 by financial MAU and #1 by app downloads, with a banking license pending. The largest single forward bet. (FACT — Benzinga 2026-06-08; Q4-2025 call.)
• New categories — groceries, white goods/heavy, pharmacy, and cross-border (CBT, China–US–LatAm corridors), the last called a “multi-billion-dollar” B2B opportunity.
• Advertising penetration (~2% → high-single-digit % of GMV), credit penetration (Argentine private credit is single-digit % of GDP; Brazil card share ~2%), and geographic expansion (Chile fintech MAU +75%, Colombia GMV accelerating, Peru).

** Verdict — high-quality growth, increasingly subsidized at the margin.** Growth is predominantly organic, unit-driven, and accelerating, with durable engagement and cohort evidence — but a non-trivial slice is bought with margin and subsidy (5–6pts in 2025), and the lowest-quality increment sits in the fast-growing, down-the-curve Brazilian credit book (duration extended 5→8 months into thinner-spread segments). The thesis depends on the subsidy producing a structurally lower-cost, higher-frequency network (the unit-cost data supports this) and on ads/credit maturation eventually converting the bought growth back into margin.

6. Financial Quality

The net-income-vs-operating-income gap — the crux, decomposed

(FACT — FY2025 10-K consolidated statements; income-tax note.)

The headline “+4.5% net income on +39% revenue” frightens at a glance, but the dominant driver (~$324M of the gap) is tax, not operations. The effective rate jumped 8.3 points because Argentine disinflation eroded the tax-inflation-adjustment deduction (the Argentine inflation-adjustment benefit collapsed from −$118M to −$29M as inflation fell 117.8% → 31.5%, while the Argentine statutory-rate-difference cost rose from +$91M to +$256M). This is a transitory, disinflation-shaped headwind tied to Argentine macro normalization, not a permanent step-up — and notably the same disinflation is a tailwind to real Argentine purchasing power and credit quality. Pretax income grew +16.9%. (FACT — income-tax reconciliation note.)

But the operating-margin compression is real and not a tax illusion. Gross margin fell 46.1% → 44.5% on (a) lower-margin 1P retail mix (+70% growth), (b) logistics/fulfillment investment, and © Mexico expansion; the provision for doubtful accounts — now the single largest operating expense at $3,091M, +66%, 10.7% of revenue — and S&M (+47%) outgrew revenue. Q1-2026 confirms the operating squeeze is independent of tax: operating income fell −19.9% YoY to $611M, with provisions +106.5% (14.1% of revenue) and S&M +64% (a $250M marketing step-up plus 43% S&M-headcount growth). And it is geographically broad — Brazil’s direct-contribution margin fell 20.0% → 13.7%, with Brazil direct contribution declining in absolute dollars (−9.2%) despite +33% revenue; consolidated direct-contribution margin fell 23.7% → 20.4%. (FACT — segment Note 8; Q1-2026 MD&A.)

The fintech credit book — the key quality-of-earnings question

(FACT — 10-K Note 5; Q1-2026.)

The book nearly doubled in FY2025, and credit cards (+114%) are now the largest slice (~45% of gross loans) — the lowest-NIMAL, longest-duration, highest-loss product, and the engine of the NIMAL decline (36.2% → 28.2% → 22.4% → 17.8% in Q1-2026). On asset quality: allowance/gross loans is stable-to-slightly-rising (25.5% → 25.1% → 26.2%) and total past-due/gross is roughly flat (26.8% → 26.3%), so there is no evidence of a hidden asset-quality blowup. But the absolute delinquency level is very high by bank standards (the structural cost of lending to a financially-underserved population — write-offs ran ~18.8% of average gross loans, $1,796M in 2025), and off-balance-sheet undrawn credit-card commitments exploded from $2,872M (YE2024) to $9,001M (YE2025) to $11,892M (Q1-2026) — contingent credit risk now roughly equal to the entire on-book portfolio, against only a ~$36M CECL reserve. (FACT — Note 5; Q1-2026 Note.) The Q1-2026 provision spike was explicitly driven by originations growing 81% (mostly cards/consumer) and an increase in average consumer-portfolio duration. OPEN QUESTION: is the duration extension a deliberate product decision or early evidence of slower repayment? Versus PagSeguro/StoneCo (thinner credit books) and Nubank (a pure card lender), MELI’s disclosure is adequate but its blended charge-off and provision intensity are higher than a payments processor’s, and the rapid card mix-shift is the single biggest QoE watch-item.

Cash-flow quality — the $12.1B that isn’t $12.1B

OCF of $12,116M is ~6× net income — for a deposit-taking lender this is structurally distorted, not a sign of pristine earnings. The two biggest swing factors — +$5,341M customer-deposit float and the $3,091M non-cash provision add-back — are not distributable cash, and the real cost of growing the book sits in investing (changes in loans receivable −$6,659M). A defensible normalized FCFE ≈ OCF − capex − loan-book funding ≈ $12,116 − $1,343 − $6,659 ≈ $4.1B — still ~2× net income (the deposit float genuinely funds part of the book at near-zero cost, a real advantage), but far below the ~$10.8B “OCF − capex” an unwary analyst would quote. (FACT — cash-flow statement; INTERPRETATION on normalization.) Management’s own “adjusted free cash flow” strips the same fintech working-capital items — reasonable, but it was expanded in Q2-2025 to also add back management-restricted cash and digital assets, a definitional loosening worth watching. Capex inflected +56% to $1,343M (logistics + data centers); SBC is the cash-settled LTRP (~$303M), a P&L cash cost rather than dilution — share count is flat at ~50.7M, a conservative, non-dilutive treatment.

Balance sheet, Argentina, and returns

Total assets reached $42.7B (up 69%). The liability stack is a bank’s: funds payable to customers $13,029M (cheap, sticky deposit float), loans payable & other financial liabilities $9,193M (bank loans, deposit certificates/financial bills, $2,852M collateralized securitization debt, and 2031/2033 notes), and card-transaction payables $3,771M, against $6,748M equity. Liquidity is strong once you sum cash + short-term + long-term investments (ex-restricted) = $8,063M (the bare cash-equivalents line of $3,670M understates it); restricted cash of $9,867M is regulatory customer-fund segregation, not free liquidity. Leverage is appropriate for a profitable lender, not stretched — but it is materially higher than the market’s “asset-light marketplace” mental model, and the securitization reliance introduces wholesale-funding sensitivity if LatAm credit spreads gap out. (FACT — balance sheet; Note 16.)

Argentina (ASC 830, highly inflationary since 2018) flows through the P&L as FX losses ($337M in 2025), the tax-line swing above, and inflation-flattered Argentine revenue (+56% USD) and margin (41.6% direct contribution, the highest of any geography) — all of which should be discounted; MELI uses the official (not parallel) exchange rate, which historically overstated USD-translated Argentine results, a distortion now shrinking post-liberalization.

ROE is elite (~36%, NI $1,997M / avg equity $5,550M) but falling (from ~43% in 2024) on the tax-and-credit drag, even as the balance sheet levers up to fund the loan book.

** Verdict — economics improve with scale on Commerce, dis-improve on Fintech, and the reported earnings are currently obscured by Argentine macro.** MELI is a genuinely high-return franchise (~36% ROE, ~90%-take-rate marketplace economics, deposit-funded lending, no dilution, $8B+ ex-restricted liquidity) whose cash generation is real even after stripping float (~$4.1B normalized FCFE). But the FY2025 “earnings stall” exposes three structural pressures the +39% revenue hides: (1) operating and direct-contribution margins are compressing (Brazil 20.0% → 13.7%); (2) the credit-card pivot is trading margin and balance-sheet risk for growth, with ~$12B off-balance-sheet card exposure against a ~$36M reserve; (3) the +4.5% net-income print is ~80% a transitory Argentine-tax artifact — but the Q1-2026 operating decline proves the margin pressure is independent of tax and is what the market is repricing. The QoE flags are amber, not red: coverage is stable, accounting is conservative (cash-settled SBC, no dilution, reserve-building ahead of write-offs, no one-time-gain games), but the 2023 presentation recast complicates trend analysis, the adjusted-FCF definition was loosened in 2025, and the biggest forward risk is a LatAm credit cycle hitting a doubled, card-heavy book at a 17.8% NIMAL.

7. Capital Allocation

Philosophy: reinvest nearly everything, return nothing. MELI pays no dividend and runs no equity buyback program (FY2025 “repurchases” in the filings are $13M of bond repurchases, not stock). It plows essentially all internally-generated cash into the moat — logistics/fulfillment, 1P inventory, the Mercado Pago credit book, Mexico, and technology/AI. FY2025 operating cash flow was $12,116M against capex of $1,343M (+56% YoY) and $6,659M of net new lending. This is defensible because the opportunity set is large and the returns are evidenced — but it puts the entire burden of the thesis on reinvestment quality. (FACT — 10-K.)

Build-not-buy discipline (a genuine strength). The trailing five-year SEC corpus shows no material acquisitions — investing cash flow carries essentially no business-combination spend, and there is no goodwill-driven empire-building. Through the Greenwald lens, this avoids acquirer’s-curse value destruction; through Marathon’s, refusing to chase capital into M&A at cycle-peak multiples is exactly the supply-side discipline that preserves returns. The cost is forgone optionality, but MELI’s organic TAM is deep enough to make this the right trade.

A sensibly bifurcated capital structure. The 2018 $880M 2.00% convertibles were fully converted/redeemed by November 2023 with capped calls settled to cap dilution (shareholder-friendly); the current senior stack is modest ($700M 3.125% 2031 notes + a newly-issued $750M 4.90% 2033 notes, Dec-2025; the $400M 2.375% 2026 sustainability notes matured/repaid Jan-2026). Separately, ~$2,852M of securitization debt plus local secured lines and the $13.0B deposit float fund the lending operation — asset-backed, self-liquidating funding against the loan book, with modest corporate notes for general purposes. The company is net-cash-positive overall (~$5.3B liquidity ex-regulatory) and comfortably solvent; “debt/equity” looks high only because float and securitizations are fintech working capital, not corporate leverage. (FACT — 10-K debt note; 8-Ks Dec-2025.)

Is the reinvestment value-creative? On the available evidence, yes — with one un-cycled caveat. The “deliberate 5–6pt margin compression” is better read as growth-stage operating-leverage timing than value destruction: Brazil GMV >30% with share gains and record NPS attributed to the shipping/1P spend; advertising +67%; acquiring TPV +25%/+50% (Brazil/Mexico); credit cohorts turning NIMAL-positive at two-year maturity; and unit-cost deflation proving the free-shipping spend builds a structurally lower-cost network. It is not empire-building (no M&A, no diversification away from the core, no vanity capex), and it is funded from internal cash, not dilutive issuance. The genuine risk to the return thesis is concentrated in the credit book — it has grown ~90% into a benign environment and the newest, fastest-growing cohorts (Mexico, Argentina cards) are un-seasoned through a downturn; NIMAL on the card book is not yet positive on average. (INTERPRETATION.)

Compensation and alignment. Three elements: modest base salary, an annual cash bonus, and the dominant Long-Term Retention Program (LTRP) — a cash incentive paid over six years, where 50% of the variable portion moves with the stock price and the whole award is forfeitable on departure. The annual-bonus metrics are Net revenues & financial income 40%, Income from operations 35%, adjusted TPV 10%, Competitive NPS 15% — growth-and-profit weighted, with no per-share, ROIC, ROE, or return-on-capital metric anywhere in the plan. For a business deploying billions into a credit book, the absence of a risk-adjusted-return metric is a genuine alignment gap (mitigated by a flat share count, which makes growth largely per-share growth, and by the six-year stock-linked LTRP tail). CEO compensation is dominated by that multi-year stock-linked cash plan ($13.1M for Galperin in 2025, of which $12.3M was LTRP) — conservative for shareholders (cash, not dilution; forfeitable; long tail). (FACT — 2026 DEF 14A.)

Founder alignment and governance. Single share class — no dual-class, no super-voting. Founder Marcos Galperin’s economic stake sits in the Galperin Trust at 7.00% (3.55M shares); institutions own the float (Baillie Gifford ~7.5% the largest single holder; ~83% institutional overall). Effective Jan 1, 2026, Ariel Szarfsztejn (former Commerce President) became CEO and Galperin moved to Executive Chairman — an orderly, internally-promoted succession that signals bench depth while keeping the founder economically anchored. Clawback, anti-hedging/pledging policies are in place (one legacy pledge exception). (FACT — proxy; 13D/A.)

Insider behavior — a rare, clean positive. The headline is director Alejandro Aguzin’s discretionary open-market BUY — 600 shares at ~$1,655 (~$993k) on 2026-05-22, transaction code P with the 10b5-1 flag = 0 (discretionary, not a plan) — corroborated by CAO Marcelo Melamud’s smaller code-P buy at ~$1,756 in February. On the sell side, the five-year Form 4 corpus shows no pattern of discretionary insider selling (the only 2024 Galperin/Giménez Form 4s were tiny code-J fund-distribution receipts, not sales) — because the cash-settled LTRP means insiders simply don’t accumulate-and-dump equity. The Aguzin buy is a clean, if modest, signal that an insider viewed ~$1,650 as attractive right into the drawdown. (FACT — SEC Form 4, accession 000162828026038297, 2026-05-26.)

** Verdict — yes, with one un-cycled caveat.** MELI runs a disciplined, internally-funded, build-not-buy reinvestment model with no dilutive issuance, no value-destroying M&A, no buyback-at-any-price, and a sensibly bifurcated balance sheet; the deliberate margin compression is a defensible timing decision backed by real return evidence; founder alignment is genuine; and the rare insider buy reinforces it. Two honest deductions keep this from an unqualified A: the incentive plan lacks any per-share or return-on-capital metric, and the credit book’s attractive returns are not yet seasoned through a LatAm downturn. Net: high-quality, shareholder-aligned capital allocation — graded strong, watch the credit cycle and the missing return metric.

8. Changes and Headwinds — Last Two Years

The defining change is a conscious re-rating of the margin profile. Management explicitly traded ~5–6 points of 2025 operating margin for growth across four vectors: the Brazil free-shipping cut to BRL 19 (June 2025), credit-card acceleration (Brazil/Mexico/Argentina), 1P expansion, and cross-border. It frames margin as a “dial” and told investors in Q1-2026 not to expect the compressed level to change materially near-term — unusually specific from a company that refuses to guide. Other shifts: targeted Brazil take-rate cuts (P&L impact from Q2-2026), the Meli+ relaunch tying fintech perks to shipping/content, the Mercado Play launch, the slow-shipping network build, and the discontinuation of Mercado Shops (folded into the Marketplace, Dec-2025). (FACT — Q4-2025 letter; 10-K.)

The Q1-2026 result and the narrative shift. Q1-2026 is what spooked the market: revenue +49% but operating income −19.9% (op margin ~6.9%), provisions +106%, pretax −18%; the stock fell ~36% peak-to-trough and JPMorgan cut its price target on 2026-05-28 citing margin pressure ahead. The weight of evidence says investment, not execution failure: ~2/3 of the credit-driven compression is the mechanical IFRS-9 front-loading of lifetime losses on a book growing 87% versus revenue 49% (a reversible timing effect), with the remaining ~1/3 a genuine, deliberate move down the credit curve at thinner spreads (Brazil personal-loan duration 5→8 months) — still double-digit-margin profitable, just less so. Asset quality is described as stable (credit-card NPL at an all-time-low 4.4% in Q4-2025). The bear read: management is masking competitive pressure with an “investment” story, and the down-the-curve credit push raises cycle risk just as it leans in. (FACT/INTERPRETATION — Q1-2026 10-Q & call; Benzinga 2026-05-28.)

Competitive changes. Brazil intensity rose materially — Shopee became #1 by orders and EBITDA-positive, Amazon is making aggressive seller-side moves, Temu/Shein/AliExpress drive low-ASP cross-border pressure, and Nubank is scaling lending against MELI’s credit book. MELI’s response (threshold cut, conditional take-rate cuts, AI-search deployment lifting conversion) is front-footed and its own KPIs are at records — but the margin sacrifice is at least partly a defensive response to low-ASP encroachment.

Regulatory / macro. Argentina under Milei — FX liberalization and disinflation (211% → 117% → ~31%): structurally positive long-term (a real credit market can finally form; private credit is single-digit % of GDP) but a near-term drag (it collapses the inflation-inflated NIMAL/FX-neutral growth that flattered 2023-24, and drove the FY2025 tax-rate jump). Brazil — selective pullback from thin-margin Pix acquiring toward credit-card acquiring; a new 10% withholding law; rising rates/oil as 2026 watch items. Mexico — the interchange cap was postponed/put on hold (a removed headwind), banking license in process. Cross-border — tariffs/de-minimis dynamics on Asian imports are an evolving swing factor for the CBT opportunity. (FACT — 10-K; calls.)

Leadership/insider. The CEO transition (Szarfsztejn) was orderly; the Aguzin open-market buy (~$994k at ~$1,655) is a bullish post-selloff signal; offsetting, NWI Management sold ~$82.4M (a fund-level disposal, not an insider). No material litigation surfaced in the window.

** Verdict — net thesis-strengthening on a 3–5 year view, clearly thesis-pressuring near-term.** The strategic shifts (threshold cut, credit-card scale, Mexico, ads infrastructure) are building durable competitive assets — the unit-cost deflation, cohort maturation, record engagement/NPS and share gains are real and corroborated. But they were bought with a deliberate 5–6 points of margin, the Q1-2026 print exposes how reliant the bull case now is on an unproven future margin recovery, and competition (Shopee profitable and #1 by orders) plus a credit book leaning down-the-curve into a higher-rate Brazil are genuine risks. The changes strengthen the business; they have weakened the near-term financial narrative and raised the bar for what must go right.

9. Risk Analysis

The risk profile is dominated by two correlated mid-likelihood/high-impact items — a LatAm credit cycle and structural margin compression — both of which would bite hardest in Brazil and both of which the current price does not cushion. The catastrophic-loss risk is genuinely low: this is a profitable, net-cash, diversified franchise, not a balance-sheet-fragile story.

10. Valuation Discussion (Embedded Expectations)

No price target and no recommendation appear in this section — only embedded-expectations and scenario framing. The single positional view in this article is the labeled author’s-take block at the top.

The P/S-vs-P/E split is the entire story

(FACT — third-party valuation data and public market data, 2026-06-10; EV/EBIT and FCFE computed from the financials.)

A stock cannot sit at the 22nd P/S percentile and the 74th P/E percentile simultaneously unless its margin has compressed — which is exactly what happened. Sales kept compounding; the profit conversion deteriorated, and the ~40% drawdown re-priced the multiple. The market is implicitly betting the compression is cyclical/deliberate (so P/S is the right lens and the stock is cheap) rather than structural (so P/E is the right lens and the stock is fair-to-rich). One caveat against over-weighting the “cheap on sales” signal: ~30% of revenue is rate/credit-sensitive interest income that does not deserve a marketplace multiple, so the blended 2.53x P/S understates the richness of the high-margin commerce/ads revenue and overstates the value of the interest income.

Sum-of-the-parts — a sanity check that passes, not a value unlock

• Commerce (~$16.2B revenue) — dominant marketplace + high-margin retail media + logistics; a growth-marketplace ~3.0–3.8× revenue ≈ $48–62B.
• Fintech (~$12.7B revenue) — payments/acquiring on a fintech multiple, plus the $14.6B credit book valued as a bank (P/B-of-book adjusted for ROE and loss content; cross-read PAGS ~1.0× book/~15% ROE, STNE ~1.3×/~20%, Nubank ~5–6×/~29–33%). At ~2.5–4× the credit equity ≈ $25–40B.
• SOTP total ≈ $73–102B vs the ~$87B EV — the midpoint sits right on top of the current EV. Conclusion: no hidden discount. SOTP only screams cheap if you apply a generous multiple to all fintech revenue (ignoring that much is interest income) and assume the credit book seasons benignly — i.e., assume away the two biggest risks. (INTERPRETATION/ASSUMPTION.)

Embedded expectations / reverse-DCF

Working backward from ~$87B EV with a ~12–13% LatAm (USD, EM-premium) cost of capital and capitalizing normalized FCFE, the price embeds a compound bet: ~20–25% revenue CAGR for five years (decelerating from +39% to ~$70–85B by 2030) AND operating margin recovering from ~11% toward ~14–16% by 2028-2030 AND benign credit seasoning (provisions prove front-loaded, NIMAL stabilizes near ~18%) AND a sustained ~18–22× EV/EBIT terminal multiple. None is heroic in isolation; all four together with no cushion if one breaks is the embedded-expectations risk. Critically, the market is not capitalizing the Q1-2026 −19.9% print as the run-rate — it is treating the margin dip as deliberate investment. (ASSUMPTION — illustrative reverse-DCF.)

Scenario analysis (embedded-expectations outputs — NOT price targets)

(ASSUMPTION throughout — illustrative ranges, not forecasts.) The current price sits at the base-case midpoint — the market is paying for the optimistic-but-not-heroic outcome, with the bear a ~30–45% de-rating and the bull a ~+50–90% re-rating. The distribution is roughly symmetric-to-slightly-favorable only if you assign the base case the highest probability, which itself requires faith in the margin-recovery story.

Peer comparison

(FACT — public market data 2026-06-10; P/B/ROE/growth cross-read from prior peer research, Jun-2026.)

MELI is the highest-quality and highest-priced name in the LatAm cohort and roughly in line with the global e-commerce comps (AMZN ~32×, SE ~34× trailing) despite faster growth. The striking gap is versus Nubank — comparable growth and arguably superior fintech ROE at ~20× P/E and ~5–6× book, versus MELI’s ~42× and ~11–13×. A pure fintech investor would find NU the cheaper way to own LatAm digital finance; MELI’s premium is justified only by the commerce/ads franchise welded on top — the cleanest evidence that MELI’s fintech leg, stand-alone, is not the cheap part of the SOTP. Against PAGS/STNE (~6–8×), MELI is a different quality tier (those are melting-“E” deep-value plays, not comps for a 30%+ compounder).

** Verdict — what the market is underwriting correctly vs incorrectly.** Correctly: that MELI is the dominant, highest-quality LatAm digital franchise with a real moat, ~36% ROE, and a long runway — and that revenue has not deteriorated (the 22nd-percentile P/S is a multiple reset, not a demand problem); the premium is earned on quality. Potentially incorrectly: the market is capitalizing a margin recovery as near-certain and the credit book’s losses as benignly seasoning — two assumptions the data does not yet confirm (Q1-2026 op income −19.9%; provisions +66% FY/+106% Q1; NIMAL halved). At ~$87B EV the stock is priced at the base-case midpoint with no margin of safety if the margin trough proves structural or the credit cycle turns. The “cheap on sales” headline is real but partly an artifact of blending high-margin commerce with rate-sensitive interest income. The market is making a compound bet and pricing it as if the components were independent and each highly likely.

11. Variant Perception

Consensus (embedded in the ~$2,230 sell-side targets and ~83% institutional ownership): MELI is a best-in-class, founder-led compounder whose margin dip is a deliberate, temporary investment cycle, not a structural break; the ~40% drawdown is a sentiment overreaction to a transitory Argentine-tax artifact and front-loaded provisioning; ads and logistics scale will re-expand margin toward the mid-teens; the credit book is seasoning normally; and at ~33× forward / ~1.05 PEG you are paying a fair price for a 30%+ grower with a 36% ROE.

The strongest bull case: revenue never broke. Unit metrics are accelerating, the free-shipping flywheel is producing real unit-cost deflation, cohort/frequency data is deepening, ~80% of the net-income stall is a transitory tax artifact, the director bought stock into the drop, and the highest-margin lever (advertising, ~2% of GMV vs high-single-digit peers) is barely tapped. P/S at the 22nd own-history percentile is the cleanest dislocation signal.

The strongest bear case: the margin compression is structural. Q1-2026 op income −19.9% is the first hard evidence that “deliberate investment” masks competitive and credit pressure; Shopee (now #1 by orders and profitable) can compress take-rates in a structurally price-competitive market; the $14.6B credit book is a time bomb (cards 45%, provisions +66%, NIMAL halved, ~$12B off-B/S exposure, losses lagging origination); ~30% of revenue is low-multiple interest income; and at ~42× trailing / ~13× book on peak ROE, any ROE normalization compresses both the “E” and the multiple at once.

The 3–5 assumptions that matter most: (1) margin recovery (does op margin recover toward 14–16% or stay ~8–11%? — the fulcrum); (2) credit asset quality (do seasoned loss rates stabilize or keep climbing?); (3) competitive position on take-rate (can MELI hold commerce take-rate and ad load against Shopee/Amazon?); (4) the advertising ramp (the most credible bull lever); (5) Argentina/FX and Brazil/Mexico rates (does the tax/FX distortion roll off and does funding cost cooperate?).

Falsification tests. Falsify the bull: operating margin stays ≤11% across 2–3 quarters with growth intact; provisions keep outrunning book growth and NIMAL falls below ~15%; commerce take-rate or ad load declines sequentially. Falsify the bear: margin inflects toward the mid-teens within 2–4 quarters as the Argentine tax artifact rolls off and ads scale; credit loss rates plateau and NIMAL stabilizes/re-expands; commerce take-rate and ad/GMV keep rising despite Shopee.

** Verdict — the variant-perception edge.** The genuine question is binary and unusually clean: is the market over-reacting to a deliberate-investment margin dip (a contrarian buy of a quality compounder at the 22nd-percentile P/S), or correctly de-rating a business facing structural margin compression + intensifying competition + an un-seasoned credit book (in which case ~42× on peak ROE still has downside)? The evidence is genuinely mixed and not yet resolved — the bull’s strongest fact is that revenue never broke; the bear’s strongest facts are quantitative and recent. Near-zero short interest (~2% of float) confirms this is a crowded long that got hurt, not a crowded short to fade — so the asymmetry favors the patient analyst who waits for the margin-and-credit data to break one way before the multiple re-rates with it. The edge belongs to whoever correctly calls the margin-recovery + credit-seasoning pair over the next 2–4 quarters.

12. Fact vs. Interpretation Table

13. Open Questions

1. NIMAL trajectory: Is the 36% → 17.8% slide a deliberate mix-shift toward credit cards (which re-price up as cohorts mature) or early evidence of deteriorating underwriting? The consumer-duration extension (5→8 months) — product decision or slower repayment?
2. Off-balance-sheet card exposure: ~$12B of undrawn commitments against a ~$36M reserve — what is the realistic drawdown/loss behavior in a downturn?
3. Margin recovery timing: Management guides no near-term recovery; how many quarters until ads + logistics scale + credit maturation visibly inflect operating margin?
4. Commerce take-rate under Shopee pressure: Are the Brazil take-rate cuts a temporary competitive response or a permanent reset of marketplace economics?
5. Advertising ceiling: Can ad load close the gap to high-single-digit % of GMV, and how much margin does that recover?
6. Argentina normalization: How much of consolidated fintech revenue/margin is Argentine rate/FX-driven (and therefore low-quality/volatile), and how does it look fully normalized?
7. Independent share data: Management’s “Brazil share doubled in 5 years” is un-reconciled to a third-party source.

14. What Must Be True

For the bull (the deliberate-investment / quality-compounder thesis):

• Operating margin must inflect back toward the mid-teens within ~2–4 quarters as the Argentine tax artifact rolls off and advertising/logistics scale — falsified if margin stays ≤11% with growth intact across 2–3 quarters.
• Credit losses must prove front-loaded seasoning, not deterioration — NIMAL stabilizes/re-expands and loss rates plateau — falsified if provisions keep outrunning book growth and NIMAL falls below ~15%.
• The marketplace must hold take-rate and ad load against Shopee/Amazon — falsified if commerce take-rate or ad/GMV declines sequentially.

For the bear (the structural-compression / credit-risk thesis):

• Margin compression must prove structural (competition forcing permanent take-rate/price resets) — falsified if margin recovers toward the mid-teens on schedule.
• The credit book must deteriorate through the cycle (rising seasoned loss rates, NIMAL below ~15%, off-B/S commitments converting to losses) — falsified if loss rates plateau and the book stays double-digit-margin profitable.
• The premium multiple must fade as ROE normalizes — falsified if ROE holds in the mid-30s and the franchise sustains a premium re-rating.

The two cases share the same scoreboard — operating-margin trajectory and credit NIMAL/loss rates over the next 2–4 quarters — which is why this is a cleanly testable, evidence-driven setup rather than a faith-based one.

15. Source Appendix

Primary filings (SEC EDGAR):

• MercadoLibre FY2025 Form 10-K, filed 2026-02-25 (10-K/2026-02-25_meli-20251231.htm) — consolidated statements, segment Note 8, loans-receivable Note 5, debt Note 16, income-tax note, highly-inflationary (ASC 830) note, Item 1 Business.
• Q1-2026 Form 10-Q, filed 2026-05-08 (10-Q/2026-05-08_meli-20260331.htm).
• FY2024 10-K (2025-02-21) and FY2023 10-K (2024-02-23) for trend.
• DEF 14A proxy, filed 2026-04-23 (DEF_14A/2026-04-23_meli-20260423.htm) — executive comp, LTRP, beneficial ownership, board.
• SEC Form 4 — director Alejandro Aguzin open-market purchase, accession 000162828026038297, filed 2026-05-26.
• EDGAR XBRL (scripts/edgar.sh concept) — revenue (Revenues total = $28,893M FY2025; RevenueFromContractWithCustomerExcludingAssessedTax $20,335M + RevenueNotFromContractWithCustomer $8,558M), net income, operating income, OCF, equity, cash.

Transcripts (company investor relations / public transcripts): Q1-2026 earnings call (2026-05-07), Q4-2025 (2026-02-24), Q3-2025 (2025-10-29), Q2-2025 (2025-08-04), Q1-2025 (2025-05-07); Goldman Sachs Communacopia presentations (2025-09-10, 2024-09-10).

Market/quant data: third-party valuation data (own-history percentiles) (own-history percentiles), 2026-06-10; public market data (price, market cap, EV, debt, cash), 2026-06-10; financial news aggregators (JPMorgan PT cut, Benzinga 2026-05-28; Mexico $4.6B investment, Benzinga 2026-06-08; Aguzin Form 4 row).

Peer cross-reads (prior peer research): AMZN (2026-06-09), SHOP (2026-06-10), BABA (2026-06-07), PAGS (2026-06-08), STNE (2026-06-08); peer multiples for SE/NU from public market data.

Industry/external (qualitative): eMarketer/Statista LatAm e-commerce penetration (2025); Sea Limited Q2-2025 results and trade press on Shopee Brazil (Fortune, 2025); Nubank customer-base disclosures (2025-26); EBANX/PCMI LatAm payments data (2025); Banco Central do Brasil Pix statistics.

Cited figures are reconciled to the primary filings above; The analyst target (~$2,230) is third-party color and is explicitly not adopted as a target.

APPENDIX A — Standard Diligence Questionnaire

MercadoLibre, Inc. (NASDAQ: MELI) — Standard Diligence Questionnaire Appendix

Supplemental to the analysis. Fact / Interpretation / Assumption labels applied where it matters. Greenwald (Competition Demystified) and Marathon (Capital Returns) frameworks applied where they add insight.

General

What thoughtful questions have other investors asked about this company? The dominant investor questions cluster on the recent margin shock: (1) Is the Q1-2026 operating-income decline (−19.9% YoY) deliberate investment or the start of structural margin compression? (2) Is the rapidly-doubling, credit-card-heavy loan book ($14.6B, NIMAL halved to ~17.8%) seasoning normally, or is it pre-storing a credit problem the reported delinquency ratios don’t yet show? (3) Can MELI hold marketplace take-rate and ad load against a now-profitable Shopee? (4) How much of reported results is Argentine FX/inflation noise? (5) Why no per-share/ROIC metric in management comp? (6) Is the “cheap on sales” (22nd-percentile P/S) signal real, or an artifact of margin compression and a 30%-interest-income revenue mix?

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Interpretation: operating earnings are at a deliberately-depressed level (management gave up 5–6 margin points to invest), not a classic cyclical high or low. Net income is artificially depressed by a transitory Argentine-disinflation tax-rate jump (21.4%→29.7%). So reported earnings understate normalized earning power — but the credit component of earnings may be at a benign-cycle high (low loss rates not yet stressed).

Driven by the external environment or internal actions? Predominantly internal (the margin-investment decision, credit-book expansion, free-shipping cut) overlaid on external macro (Argentine disinflation/FX, Brazil rates, competitive intensity).

How stable are revenues? Very stable and growing — 28 consecutive quarters >30% revenue growth; Q1-2026 +49%. Revenue is transactional but quasi-recurring through frequency/habit. No customer >5% of revenue.

Outlook for products/services? How big will this market be? LatAm e-commerce is ~12–15% penetrated (vs ~47% China) across 650M+ people — a large, growing, under-penetrated, international (pan-LatAm) market. Fintech TAM is similarly under-penetrated (Argentine private credit single-digit % of GDP; Brazil card share ~2%). The market is growing, not shrinking.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More — Shopee is now #1 by orders in Brazil and EBITDA-positive; Amazon is investing; Temu/Shein/AliExpress add cross-border price pressure; Nubank scales lending. Mitigated by an expanding pie.

How profitable is the business (ROIC, ROE)? ROE ~36% (down from ~43%), elite but falling on tax/credit drag. Consolidated direct-contribution margin 20.4% (Brazil 13.7%). Commerce economics are high-return/asset-light; the credit book consumes capital and lowers/cyclicalizes incremental ROIC.

How profitable is the industry — competitors, barriers to entry? Bifurcated: scaled, integrated leaders (MELI, Shopee, Nubank) earn well; sub-scale players bleed (Americanas bankruptcy, Via restructuring). Barriers: owned logistics density (high capital), fintech licensing, two-sided liquidity, brand/trust. Greenwald: genuine economies-of-scale + customer-captivity barriers in logistics/fintech; weaker (contestable) in pure marketplace matching.

Can the business be easily understood? Mostly — two clear engines (Commerce, Fintech). The complication is the fintech credit book (a bank embedded in an e-commerce company), Argentine hyperinflation accounting, and the float-distorted cash flow.

Can it be undermined by foreign low-cost labor? Not directly (services/logistics are local). The relevant analog is low-cost cross-border supply (Temu/Shein/AliExpress) pressuring marketplace pricing.

Do brands matter? Yes — #1 brand preference and NPS in commerce and fintech across core markets; trust is a more durable asset in a low-institutional-trust, high-fraud region.

Nature of competition / switching costs? Competition on price, logistics speed, selection, and fintech breadth. Switching costs high for sellers (fulfillment, ads, working-capital credit, reputation) and moderate for buyers (rising with Meli+/principality/credit card).

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The brand, the logistics network density, the proprietary underwriting data, and the deposit-funding franchise are under-recognized economic assets.

Off-balance-sheet liabilities? Yes and material: ~$11.9B of undrawn credit-card commitments (Q1-2026) against a ~$36M CECL reserve — a genuine contingent credit exposure. Also operating leases and securitization structures.

How conservative is the accounting? Conservative on the equity side (cash-settled comp/no dilution, reserve-building ahead of write-offs, no one-time-gain games, official-FX Argentina treatment). Watch items: the 2023 presentation recast complicates trend analysis, and the “adjusted FCF” definition was loosened in Q2-2025.

How CapEx-hungry is the business? Increasingly — capex +56% to $1.34B (logistics/fulfillment + data centers), the “asset-light marketplace” getting heavier. Still modest vs revenue (~4.6%), but rising.

Capital Allocation & Management

How much FCF, and how is it used? Headline OCF $12.1B but normalized FCFE ~$4.1B (strip deposit float + loan-book funding). Used almost entirely for reinvestment — logistics, credit book, Mexico, tech. No dividend, no buyback.

Significant acquisitions recently? No — build-not-buy is a consistent discipline (no material M&A in 5 years; no goodwill empire-building). A capital-allocation strength.

Buying back / issuing shares? Neither materially — share count flat at ~50.7M. Comp is cash-settled (LTRP), so no dilution and no insider sell-flow.

Compensation policy / incentive alignment? Modest base + annual bonus (revenue 40% / op income 35% / adj TPV 10% / NPS 15%) + a six-year stock-linked cash LTRP (CEO 2025 ~$13.1M, ~$12.3M LTRP). Gap: no ROIC/per-share/return-on-capital metric — a real flaw for a capital-deploying lender, partly mitigated by flat shares + the stock-linked LTRP.

Motivations of management? Founder-led (Galperin → Executive Chairman; Szarfsztejn CEO from Jan-2026), single share class, Galperin Trust 7% economic stake; long-horizon, growth-and-moat-oriented. The director (Aguzin) open-market buy at ~$1,655 signals insider conviction at current prices.

Valuation & Market Data

ADR, MLP, or K-1? No — MELI is a US-domestic 10-K/10-Q filer (US GAAP, USD), Delaware-incorporated, common stock (not an ADR, not a K-1 issuer).

Dividend policy? None (no dividend; reinvests all cash).

How profitable? ~36% ROE; ~11% operating margin (deliberately compressed); ~$2.0B net income; ~$4.1B normalized FCFE.

Is net income diverging from cash from operations? Yes, massively — OCF $12.1B vs NI $2.0B (~6×), but the divergence is structural (deposit float + non-cash provision add-back), not an earnings-quality red flag. Normalized FCFE (~$4.1B) is the honest cash measure.

Risks & Downside

What would cause the stock to decline? A credit-cycle hit to the un-seasoned loan book; confirmation that margin compression is structural (Shopee/Amazon take-rate pressure); failure of the margin recovery to materialize; an Argentine/Brazilian macro shock; adverse fintech regulation (Pix/interchange).

Risk of a catastrophic loss? Low — profitable, net-cash-positive, diversified across countries and two engines; no solvency, fraud, or covenant issues.

Chance of a total loss? Negligible. The realistic downside is a multiple/earnings de-rating (~30–45% in the bear case), not impairment of the franchise.

Recent News & Events

Has the business environment changed recently? Yes — (1) Q1-2026 margin shock + JPMorgan PT cut (2026-05-28); (2) Shopee turning EBITDA-positive and #1 by orders in Brazil; (3) Argentine disinflation/FX liberalization under Milei; (4) the deliberate 5–6pt margin investment; (5) a $4.6B Mexico investment announced 2026-06-08.

Significant acquisitions? None.

Change in accounting policies? A 2023 income-statement presentation recast (financial expenses into cost of revenue) and a Q2-2025 broadening of the “adjusted FCF” definition — both worth noting for trend comparability.

Recent changes — new markets, facilities, management? CEO transition (Szarfsztejn, Jan-2026); Mercado Shops discontinued (Dec-2025); BRL 19 free-shipping threshold (Jun-2025); credit cards launched in Argentina (2025); Mexico banking license in process; heavy fulfillment/data-center build.

APPENDIX B — Source Appendix

MercadoLibre, Inc. (NASDAQ: MELI) — Source Appendix

Primary sources prioritized over secondary; every material number reconciled to a filing. Accessed 2026-06-11 unless noted.

A. SEC Filings (SEC EDGAR)

B. EDGAR XBRL (SEC EDGAR XBRL)

• Total net revenue: Revenues tag — FY2023 $14,473M / FY2024 $20,777M / FY2025 $28,893M. Reconciliation: RevenueFromContractWithCustomerExcludingAssessedTax (contract revenue, FY2025 $20,335M) + RevenueNotFromContractWithCustomer (financial/interest income, FY2025 $8,558M) = $28,893M.
• OperatingIncomeLoss: FY2023 $2,207M / FY2024 $2,631M / FY2025 $3,201M.
• NetIncomeLoss: FY2023 $987M / FY2024 $1,911M / FY2025 $1,997M.
• NetCashProvidedByUsedInOperatingActivities: FY2023 $5,140M / FY2024 $7,918M / FY2025 $12,116M.
• StockholdersEquity: YE2024 $4,351M / YE2025 $6,748M. CashAndCashEquivalentsAtCarryingValue: YE2025 $3,670M (sum with ST+LT investments for true liquidity ~$8.06B).

C. Earnings Call & Event Transcripts (company IR / public transcripts)

• Q1-2026 earnings call, 2026-05-07 (margin compression, credit duration, cost-per-shipment −17%, items sold +56% Brazil).
• Q4-2025 earnings call, 2026-02-24 (the 5–6pt deliberate margin investment; 28 straight >30% quarters; NPL 4.4%).
• Q3-2025 (2025-10-29), Q2-2025 (2025-08-04), Q1-2025 (2025-05-07) earnings calls.
• Goldman Sachs Communacopia + Technology Conference presentations, 2025-09-10 and 2024-09-10 (share gains, Meli+ two-sided lock, ads runway).

D. Market / Quantitative Data (third-party, reconciled)

• public market data (yfinance), 2026-06-10 — price $1,588.29; market cap ~$80.5B; EV ~$87.2B; total debt ~$12.4B; cash ~$5.7B; 52wk $1,495–$2,645.
• third-party valuation data (own-history percentiles), 2026-06-10 — P/E 41.9x (74th own-history pctile), P/B 11.1x (67th), P/S 2.53x (22nd), composite 54th; forward P/E ~32.9x; ROE 31% (snapshot); short interest ~2.05% float; institutions ~83%, insiders ~7.3%.
• financial news aggregators — JPMorgan PT cut after Q1-2026 (Benzinga, 2026-05-28); Mexico $4.6B investment (Benzinga, 2026-06-08); NWI Management ~$82.4M disposal (2026-05-30); Aguzin Form 4 buy row (2026-05-26).
• Peer multiples (public market data, 2026-06-10): AMZN, SE (Sea), SHOP, NU (Nubank), PAGS, STNE.

E. Peer Comparison (public filings)

• Amazon (AMZN), Sea Limited (SE/Shopee), Shopify (SHOP), Alibaba (BABA), Nu Holdings (NU), PagSeguro (PAGS), StoneCo (STNE) — public filings and market data used for the peer-multiple comparison.

F. Industry / External (secondary, qualitative)

• eMarketer / Statista — LatAm e-commerce penetration ~12–15% vs ~16% US / ~47% China; Mexico ~17.7% (2025).
• Sea Limited Q2-2025 results; Fortune and trade press on Shopee Brazil #1-by-orders and e-commerce-EBITDA positive (2025).
• Nu Holdings customer-base disclosures (~105M+ Brazil / 131M+ region, 2025-26).
• EBANX / PCMI — LatAm payments and digital-wallet share data (2025).
• Banco Central do Brasil — Pix transaction statistics (~42% of Brazilian e-commerce payment value).

Note: management commentary (transcripts, investor materials) is treated as hypothesis and validated against filings, financials, and external evidence per a primary-source evidence standard. The ~$2,230 analyst consensus target is recorded as third-party color and explicitly not adopted.