MercadoLibre MELI
Consumer Cyclical · Specialty Retail · Synthos Deep Dive · 2026-08-04
The Overview
MercadoLibre is Latin America's Amazon, and also its PayPal, and increasingly its bank. It runs the biggest online marketplace in Brazil, Mexico and Argentina; it processes the payments; it lends money to buyers and sellers; and it has been building its own network of warehouses and delivery vehicles.
The business is growing very fast. Revenue rose 39% last year and 49% in the first three months of this year. Very few companies of this size grow like that.
But profits are not growing at all. In that same quarter, operating profit fell 20% and net profit fell 16%. Out of every dollar of sales, the company kept about 7 cents in operating profit, against 13 cents a year earlier.
There are two explanations and they lead to opposite conclusions.
The optimistic one — which several commentators in our research library make — is that this is deliberate. MercadoLibre is spending heavily on warehouses and delivery, exactly as Amazon did for years, and on building a lending business. Those costs are largely fixed once built, so as sales keep growing the profits eventually arrive in a rush. Analysts believe this: they expect profits to be flat this year and then jump 43% next year.
The pessimistic one is that the spending is competitive — Amazon and Sea Limited are attacking the same markets — and it does not stop.
The market has been voting for the pessimistic reading. The shares are down 20% over the past year while the American market rose 24%, and they sit 25% below their high. The company has also come in below analysts' profit forecasts in each of the last four quarters.
At $1,888.50 the shares cost about 50 times last year's profits and 34 times what analysts expect for 2027. Our estimate of fair value is $2,140, about 13% above the price. Analysts on average say $2,167.
Two things to know before deciding. First, the company reports results tomorrow, and analysts themselves expect profits to fall 16% while sales rise 44%. Second, MercadoLibre returns nothing to shareholders — no dividend, and last year it bought back one million dollars of stock, which on a $96 billion company is nothing. If the shares do not rise, you earn zero.
We are saying Watch rather than Buy: the business is excellent, the price is not obviously wrong, and the single fact that would settle it — whether profit margins have turned — arrives tomorrow. We would buy this at $1,600.
- Downside Risk 7/10. Emerging-market, credit, competitive and multiple risk, with no yield to cushion it.
- Growth Quality 8/10. The best in this batch on revenue; the worst on the translation into earnings.
- Exponential Potential 7/10. The best in this batch — a genuine multi-sided flywheel in a large unpenetrated market.
Putting a number on it: our fair-value estimate is $2,140 against a current price of $1,966.25 — real upside if our numbers are right.
Our summary metrics
"Rated 7 — a genuinely dominant franchise carrying emerging-market, credit, competitive and multiple risk at once, with no yield to cushion any of it. The supports: 29.6% return on equity; 20.4% return on capital employed; interest coverage of 17.5x; $6,299 million of cash and short-term investments; and a share count that has been essentially static at 50,697,000 diluted for two years — no dilution, and the FY2025 buyback was literally $1 million. Against that, five risks. FIRST, the earnings line has stopped growing: March-quarter operating income FELL 19.9% on revenue up 49.0%, and consensus FY2026 EPS of $39.34 is flat against the $39.40 delivered in 2025. SECOND, the forecasting record is broken: FOUR CONSECUTIVE MISSES, by 13.6%, 8.8%, 3.7% and 5.9%. A company the street cannot model is a company whose forward multiple carries less information than the number suggests. THIRD, this is a lender as well as a marketplace — days of sales outstanding are 212.4 and average receivables are $12,886 million, which is Mercado Crédito's loan book, not a trade balance; total debt is $11,392 million against $6,748 million of equity, a 1.70x debt-to-equity and 6.45x financial leverage, and much of that funds the credit portfolio. FOURTH, currency and country: 20.6% of fiscal-2025 revenue is Argentina, a jurisdiction whose exchange rate and inflation have repeatedly re-based reported results. FIFTH, valuation and yield: 49.8x trailing earnings, no dividend, and a buyback that rounds to zero, so the entire return must come from earnings and the multiple. Add a founder-to-professional-CEO transition — Ariel Szarfsztejn is now Chief Executive Officer with Marcos Galperin as Executive Chairman — and 7 is the honest number."
"Rated 8 — the highest growth score in this batch, and it is real, organic and accelerating on the top line while stalling on the bottom. Revenue: $10,780M (FY2022), $15,107M (FY2023, +40.1%), $20,777M (FY2024, +37.5%), $28,893M (FY2025, +39.1%), and the March 2026 quarter at $8,845M was +49.0% year on year — the fastest quarterly growth in the nine-quarter series. Consensus wants $40,737M in FY2026 (16 analysts), $52,035M in FY2027 (17) and $64,065M in FY2028 (18) — 41.0%, 27.7%, 23.1%. Geographic detail from `seg_geo`, which reconciles exactly: Brazil $15,201M (52.6% of revenue, +33.3%), Mexico $6,475M (22.4%, +38.8%), Argentina $5,962M (20.6%, +56.2%), Other $1,255M (4.3%). All three principal markets are compounding above 30%. The problem is below the revenue line. March-quarter operating income was $611M against $763M — DOWN 19.9% — and net income $417M against $494M, down 15.6%; the operating margin went from 12.9% to 6.9%. Consensus FY2026 adjusted EPS of $39.343 is essentially identical to the $39.40 the four 2025 quarterly actuals sum to. What holds this at 8 rather than 9: a year of 41% revenue growth producing zero earnings growth is either an investment decision that reverses or a competitive cost that does not, and the file does not settle which."
"Rated 7 — the highest exponential score in this batch, and the argument is the flywheel rather than a single product. MercadoLibre operates the dominant e-commerce marketplace across Latin America, a payments network in Mercado Pago that processes on and off its own platform, a credit book in Mercado Crédito, an asset-management product in Mercado Fondo, and a logistics network it has built itself. Each reinforces the others: marketplace volume generates payment volume, payment data underwrites credit, credit raises conversion, and owned logistics raises delivery reliability which raises marketplace share. The knowledge base's clearest statement of the addressable opportunity is `we_study_billionaires` (2026-04-18, conviction 80, speaker Daniel): 'LatAm e-commerce penetration only 14-15% vs 25% US will close, MELI holds 30-35% Brazil share and is uniquely South-America-first.' If penetration closes toward US levels on a base MercadoLibre already leads, the revenue opportunity is a multiple of today's $28.9 billion, and the company would capture it without needing a new product. The 123,670 employees and the $1,343 million of fiscal-2025 capital expenditure are what building that looks like. Against it: e-commerce is not a natural monopoly, Amazon and Sea Limited compete directly in the same markets, and the credit book converts a capital-light platform into a leveraged lender — 212 days of receivables and $11,392 million of debt are not exponential characteristics. A 7: a genuine multi-sided flywheel with a large unpenetrated market, attached to a lending balance sheet."
What does “fair value” mean?
Fair value is Synthos’s estimate of what one share is worth today, based on our model of the company’s future cash generation and the risks to it. It is not a price target or a prediction of where the stock trades next quarter — it’s the price where we think risk and reward are balanced. Above it, you’re paying for outcomes better than our base case; below it, the market is offering a margin of safety.
The model’s inputs and weightings are proprietary — they’re the product. What we publish is the output, We don’t publish a reverse-DCF cross-check for pre-profit companies — negative or missing earnings break that math — so take this number on our modeling alone.
What do the 5 tiers mean? (Core · Tactical · Watch · Hold · Avoid)
The Road Ahead
What we expect to matter in each window, and the evidence that would prove us wrong.
Short term 0-6 months
No differentiated view- Driver
- "The company reports tomorrow and the consensus itself forecasts an earnings decline. Second-quarter results are due 2026-08-05 against adjusted EPS of $8.67 — down 15.9% from the $10.31 delivered a year earlier — on revenue of $9,742M, up 43.5%. That is the shape of the whole problem in one line. MercadoLibre has MISSED consensus in four consecutive quarters, by 13.6%, 8.8%, 3.7% and 5.9%, and the one beat in the file is fifteen months old. Technically the position is neutral-to-mildly-constructive: at $1,888.50 the stock is 24.8% below its 52-week high of $2,510.97, 22.1% above its low of $1,546.81, 8.8% above a 50-day average of $1,736.43 and 0.1% BELOW a 200-day average of $1,890.37 — sitting almost exactly on the longer average. RSI is 59.6 and MACD is +39.96, both constructive. But three-month return is +4.1% against SPY's +7.6%, six-month MINUS 12.0% against +11.1%, and twelve-month MINUS 20.5% against PLUS 24.3% — a 45-point deficit. The maximum drawdown from peak over the trailing year is 27.7%. No position is taken into tomorrow's print."
- What we’re watching
- "The 2026-08-05 print, and within it one number above all others: the operating margin. It was 6.9% in the March quarter against 12.9% a year earlier, and the entire investment case turns on whether that compression is a deliberate, reversible investment in logistics and credit — which is what the knowledge base argues — or a permanent competitive cost. Second: whether the four-quarter miss streak extends to five against the $8.67 estimate. Third: any guidance or commentary on the credit book — loans receivable are classified as Merchant, Consumer, Credit cards and Asset-backed in the 10-Q, average receivables are $12,886M, and the knowledge base's specific bullish claim is that net interest margin after losses is rising while bad-loan losses fall; that is a checkable assertion. Fourth: Argentina, 20.6% of fiscal-2025 revenue and growing 56.2%, where currency movements can re-base the reported figures in either direction. Fifth: any commentary from Ariel Szarfsztejn, in his first full year as chief executive with founder Marcos Galperin moved to Executive Chairman, on how long the investment phase runs."
- Confidence
- Low
Medium term 6-24 months
Tailwind- Driver
- "The medium-term case is that the investment phase ends and the operating leverage that 41% revenue growth should produce finally arrives. Consensus has adjusted EPS at $39.343 in FY2026 — flat — then $56.318 in FY2027 (+43.1%) and $78.075 in FY2028 (+38.6%), on revenue rising from $28,893M actual to $40,737M, $52,035M and $64,065M. The FY2027 step is the entire valuation case, and it is a bet that margins recover rather than that revenue accelerates: revenue growth is forecast to DECELERATE from 41.0% to 27.7% while earnings growth jumps from 0% to 43.1%. That can only happen through margin expansion. The mechanism the bulls describe is legible: logistics capital expenditure ($1,343M in fiscal 2025 against $509M in fiscal 2023) builds fixed infrastructure whose cost stops growing once built, while volume keeps growing over it; and the credit book, once seasoned, earns a spread rather than absorbing provisions. `invest_like_the_best` (2026-03-08) calls it 'Amazon-style investment mode that will unlock profitability later.' Against that: nothing in this file demonstrates the margin has turned, and a company that has missed four consecutive quarters has not earned the benefit of the doubt on a forecast that requires a 43% earnings step."
- What we’re watching
- "Whether the operating margin recovers from 6.9%, and in which quarter. That is the single variable on which the FY2027 consensus rests. Whether logistics capital expenditure moderates from the fiscal-2025 run rate of $1,343M — a capital programme that keeps growing is an investment phase that has not ended. Whether the credit book's economics behave as the knowledge base claims: rising net interest margin after losses, falling bad-loan losses, on a receivable base averaging $12,886M with a 212-day cycle. Whether competitive intensity from Amazon and Sea Limited in Brazil and Mexico is visible in marketing spend or take rates. Whether Argentina's 56.2% growth rate is real volume or currency re-basing. Whether MercadoLibre begins returning capital — the fiscal-2025 buyback was $1 million and there is no dividend, so a shareholder currently owns growth and nothing else. And whether the miss streak breaks: four consecutive quarters below consensus is a pattern, and its continuation would suggest the company itself does not know where its margin is going."
- Confidence
- Low
Long term 2+ years
Tailwind- Driver
- "Long-run this is the strongest structural franchise in this batch. MercadoLibre is the dominant e-commerce marketplace in Latin America, a region where the knowledge base puts online penetration at 14-15% against 25% in the United States, and it holds an estimated 30-35% share in Brazil, its largest market. It owns the payments rail (Mercado Pago), the credit underwriting (Mercado Crédito), the savings product (Mercado Fondo) and, increasingly, the logistics network — a vertically integrated stack that no regional competitor has assembled and that global competitors have repeatedly failed to replicate in these markets. `we_study_billionaires` (2025-09-18, conviction 95) lists it among seven long-held '100-baggers' alongside Amazon, Netflix, Nvidia and Tesla; `we_study_billionaires` (2022-11-24, David Gardner) called it 'a smaller younger version of Amazon at ~$40B market cap' — it is now $95.7B. The long-run bet is simply that penetration closes on a base the company already leads, and that the fixed logistics and payments infrastructure being built now carries far more volume at far higher margin later. The risks that could break it are equally structural: Latin American currency and political instability, a credit cycle in an unsecured consumer loan book, and the possibility that e-commerce in these markets settles into a competitive equilibrium with normal retail margins rather than a platform monopoly."
- What we’re watching
- "Whether Latin American e-commerce penetration actually closes toward US levels, and whether MercadoLibre keeps its share while it does — the two are separable, and only the second is within the company's control. Whether Mercado Pago becomes a genuine financial institution with regulatory capital requirements, which would change the return profile of the whole business. Whether the credit book survives a real Latin American credit cycle; it has grown through a benign period. Whether Amazon or Sea Limited commit to sustained loss-making share purchase in Brazil or Mexico. Whether the founder transition holds — Marcos Galperin built this company over twenty-seven years and is now Executive Chairman, with Ariel Szarfsztejn as chief executive and a six-year cash Long Term Retention Program adopted 2026-03-31 designed to hold senior management in place. Whether Argentina, 20.6% of revenue and the fastest-growing market, remains an asset rather than a liability. And whether the company ever converts scale into cash returned to shareholders, or reinvests indefinitely."
- Confidence
- Medium
Exponential Potential
"Rated 7 — the highest exponential score in this batch, and the argument is the flywheel rather than a single product. MercadoLibre operates the dominant e-commerce marketplace across Latin America, a payments network in Mercado Pago that processes on and off its own platform, a credit book in Mercado Crédito, an asset-management product in Mercado Fondo, and a logistics network it has built itself. Each reinforces the others: marketplace volume generates payment volume, payment data underwrites credit, credit raises conversion, and owned logistics raises delivery reliability which raises marketplace share. The knowledge base's clearest statement of the addressable opportunity is `we_study_billionaires` (2026-04-18, conviction 80, speaker Daniel): 'LatAm e-commerce penetration only 14-15% vs 25% US will close, MELI holds 30-35% Brazil share and is uniquely South-America-first.' If penetration closes toward US levels on a base MercadoLibre already leads, the revenue opportunity is a multiple of today's $28.9 billion, and the company would capture it without needing a new product. The 123,670 employees and the $1,343 million of fiscal-2025 capital expenditure are what building that looks like. Against it: e-commerce is not a natural monopoly, Amazon and Sea Limited compete directly in the same markets, and the credit book converts a capital-light platform into a leveraged lender — 212 days of receivables and $11,392 million of debt are not exponential characteristics. A 7: a genuine multi-sided flywheel with a large unpenetrated market, attached to a lending balance sheet."
What could take this further than the base case — and where the market may be underpricing it. Full forward-growth and acceleration math in Deeper analysis, §4 below.
Deeper analysis
Technicals, fundamentals, valuation, and the full expert-claim evidence panel — the detail behind the numbers above.
⚠ EARNINGS BANNER — RESULTS ARE RELEASED TOMORROW
MercadoLibre reports second-quarter 2026 results on 2026-08-05, ONE DAY after this dive is struck. Vendor consensus is adjusted EPS $8.67 and revenue $9,742M.
Read those two numbers together, because they are the whole thesis in one line. Against the year-ago quarter's $10.31 and $6,790M, consensus implies revenue growth of 43.5% and an EPS DECLINE of 15.9%.
And the forecasting record is broken. MercadoLibre has MISSED consensus in four consecutive quarters:
| Report date | Actual | Estimate | Variance |
|---|---|---|---|
| 2025-05-07 | $9.74 | $8.27 | +17.8% |
| 2025-08-04 | $10.31 | $11.93 | −13.6% MISS |
| 2025-10-29 | $8.32 | $9.12 | −8.8% MISS |
| 2026-02-24 | $11.03 | $11.45 | −3.7% MISS |
| 2026-05-07 | $8.23 | $8.75 | −5.9% MISS |
Four in a row, and the only beat in the file is fifteen months old. The Watch verdict below is set with that timing and that record as explicit inputs, and no position is taken into tomorrow's print.
Reference table
| Street consensus | $2,166.67 (+14.7%) · median $2,150 · high $2,600 (+37.7%) · low $1,750 (−7.3%) · 1 strong buy / 23 buy / 9 hold / 0 sell across 33 analysts |
| Valuation | 49.8x trailing adjusted EPS ($37.89) · 48.0x FY2026E $39.343 · 33.5x FY2027E $56.318 · 24.2x FY2028E $78.075 · 30.8x EV/EBITDA · 13.1x book |
| The problem, in one row | Q1 2026 revenue $8,845M, +49.0% · operating income $611M, −19.9% · net income $417M, −15.6% · operating margin 6.9% vs 12.9% |
Geography — seg_geo reconciles EXACTLY | Brazil $15,201M (52.6%, +33.3%) · Mexico $6,475M (22.4%, +38.8%) · Argentina $5,962M (20.6%, +56.2%) · Other $1,255M (4.3%) · sum = $28,893M = reported revenue to the dollar |
| Capital return | No dividend (dividendPerShareTTM 0) · fiscal-2025 share repurchase $1 million · total shareholder yield is ZERO. The entire return must come from earnings and the multiple |
| Governance | Ariel Szarfsztejn is Chief Executive Officer; founder Marcos Galperin is Executive Chairman (8-K, 2026-04-03). A six-year cash 2026 Long Term Retention Program was adopted 2026-03-31 |
| Conviction | Moderate, uniformly bullish, and concentrated — 7 raw KB hits, 7 used, 0 discarded (the cleanest sweep in this batch), but we_study_billionaires supplies 5 of 7 (71%) and the concentration test fails |
| Technicals | −24.8% from the 52-week high of $2,510.97; +22.1% above the low of $1,546.81; +8.8% above the 50-DMA ($1,736.43) and −0.1% BELOW the 200-DMA ($1,890.37) — sitting on it; RSI 59.6; MACD +39.96; max drawdown −27.7%; 12-month −20.5% vs SPY +24.3% |
What the experts actually said 6 traceable claims on MELI · showing the highest-conviction voices
“MercadoLibre runs Amazon's scale-economics-shared playbook; LatAm e-commerce penetration only 14-15% vs 25% US will close, MELI holds 30-35% Brazil share and is uniquely South-America-first.”
“Flat 5-year stock hides a healthy business; heavy logistics capex is Amazon-style investment mode that will unlock profitability later.”
Every claim reconciles to a real claim_id in the Synthos knowledge base — this is the evidence the verdict is built on, not vibes. Management (the company itself) is shown but half-weighted; one cautionary voice is included on purpose.
Price & moving averages 12 months · 50 & 200-day averages · 52-week range
Solid line = price · dashed line = 50-day average · dotted line = 200-day average · the two thin horizontal lines mark the 52-week high and low. Price above both averages is an uptrend.
Data summary: last close $1,966.25, 8% above the 50-day average ($1,822), 6% above the 200-day average ($1,862) — an uptrend. 22% below the 52-week high of $2,511, 27% above the 52-week low of $1,547.
Bollinger Bands 20-day average ± 2 standard deviations
The shaded band widens when the stock gets more volatile. Riding the upper edge = strong momentum (sometimes stretched); the lower edge = weak / potentially oversold.
Data summary: price $1,966.25 is currently inside the band (band $1,763–$2,011).
RSI (14) momentum gauge · 0–100
Above 70 (overbought zone, shaded) = overbought, below 30 (oversold zone, shaded) = oversold. Currently 60.
MACD 12 / 26 / 9 · trend & momentum
The MACD line crossing above the signal line (bars flip to the up color) = momentum turning up; crossing below (bars flip to the down color) = turning down. Bar height = the size of that gap.
Data summary: MACD is currently above its signal line by 7.36, positive momentum.
Relative performance vs S&P 500 & its sector (XLY (sector)), set to 100 a year ago
Solid = MELI · dashed = S&P 500 · dotted = XLY (sector). A rising line means it is beating that benchmark — the sector line shows whether it is a leader or laggard within its own group.
Forward revenue & earnings actual → estimate · "FY" = fiscal year, "E" = estimate
Darker bars = actual results, brighter = analyst estimates. Taller bars to the right = expected growth.
Key stats an RIA wants
1. The business, and both segment blocks reconcile
MercadoLibre is a Delaware-incorporated company with operations across Latin America, headquartered for reporting purposes in Uruguay, employing approximately 123,670 people. Ariel Szarfsztejn is Chief Executive Officer and founder Marcos Galperin is Executive Chairman, per the 8-K filed 2026-04-03; Osvaldo Giménez is Fintech President and Daniel Rabinovich runs Technology and Operations.
The 10-Q describes the segment structure and the metric management runs on:
> "The Company manages the business country-by-country... The Company's segments include Brazil, Mexico, Argentina and other countries (which includes Bermuda, Chile, China, Colombia, Costa Rica, Ecuador, Peru, Uruguay and the U.S.). The CODM evaluates the performance of the Company's operating segments based on their direct contribution... consists of net revenues and financial income from external customers less segment costs, which include expenses, such as shipping operation costs (including warehousing costs), carrier and other operating costs, provision for doubtful accounts, cost of goods sold, collection fees, funding cost, salaries and wages, marketing expenses and hosting expenses."
Revenue by geography, from seg_geo — and this block reconciles to the dollar:
| FY2025 | Revenue | share | FY2024 | growth |
|---|---|---|---|---|
| Brazil | $15,201M | 52.6% | $11,406M | +33.3% |
| Mexico | $6,475M | 22.4% | $4,664M | +38.8% |
| Argentina | $5,962M | 20.6% | $3,818M | +56.2% |
| Other countries | $1,255M | 4.3% | $889M | +41.2% |
| Sum | $28,893M | 100.0% | $20,777M | +39.1% |
inc_a FY2025 revenue | $28,893M | $20,777M | EXACT |
Both years reconcile exactly. After MDT's 147%-of-revenue geography, SPGI's omitted segment, DUK's fiscal-2025 block losing both reportable segments and KKR's 122% double-count, a geographic block that ties to the dollar in two consecutive years is a genuinely clean result and is recorded as one.
seg_prod also reconciles exactly — Product $3,607M plus Service $25,286M equals $28,893M — but it is only a two-line product/service split, not the Commerce/Fintech division MercadoLibre and its analysts actually discuss. Note that the FY2021 and FY2020 entries DO carry the Commerce/Fintech split (FY2021: Commerce $4,635M, Fintech $2,434M), and it disappears from FY2022 onward. So the payload has the useful segmentation for the years it does not matter and the useless one for the years it does. No Commerce/Fintech figure in this dive comes from the payload.
Argentina is the fastest-growing market at +56.2% and 20.6% of revenue. That is a real business and it is also a currency exposure: reported growth in a high-inflation, managed-exchange-rate economy reflects price and translation as well as volume, and the payload provides no constant-currency breakdown.
2. The margin — the entire investment case in one comparison
| Q1 2026 | Q1 2025 | change | |
|---|---|---|---|
| Revenue | $8,845M | $5,935M | +49.0% |
| Gross profit | $3,862M | $2,771M | +39.4% |
| Gross margin | 43.7% | 46.7% | −300 bps |
| Operating income | $611M | $763M | −19.9% |
| Operating margin | 6.9% | 12.9% | −597 bps |
| Interest expense | $65M | $39M | +66.7% |
| Net income | $417M | $494M | −15.6% |
| Adjusted EPS | $8.23 | $9.74 | −15.5% |
| EBITDA | $890M | $917M | −2.9% |
Revenue up 49% and operating income down 20% is the most important fact in this file, and it is not an aberration. The nine-quarter operating-income series runs $528M, $726M, $557M, $820M, $763M, $825M, $724M, $889M, $611M — the March 2026 figure is the second-lowest in the series, on the highest revenue in the series.
The full-year picture confirms it:
| Fiscal year | Revenue | growth | Operating income | Operating margin | Net income | Adjusted EPS |
|---|---|---|---|---|---|---|
| FY2022 | $10,780M | — | $1,069M | 9.9% | $482M | — |
| FY2023 | $15,107M | +40.1% | $2,207M | 14.6% | $987M | — |
| FY2024 | $20,777M | +37.5% | $2,631M | 12.7% | $1,911M | — |
| FY2025 | $28,893M | +39.1% | $3,201M | 11.1% | $1,997M | ~$39.40 |
| FY2026E | $40,737M | +41.0% | — | — | — | $39.343 |
The operating margin peaked at 14.6% in fiscal 2023 and has fallen every year since. Consensus FY2026 adjusted EPS of $39.343 is flat against the $39.40 the four 2025 quarterly actuals sum to — a year of 41% revenue growth producing no earnings growth at all.
The bull explanation, from the knowledge base, is investment. invest_like_the_best (2026-03-08, conviction 75): "Flat 5-year stock hides a healthy business; heavy logistics capex is Amazon-style investment mode that will unlock profitability later." The capital expenditure series supports the description: $455M (FY2022), $509M (FY2023), $860M (FY2024), $1,343M (FY2025) — nearly tripled in three years, and research and development is $2,269M, 7.9% of revenue.
The bear explanation is competition, and this file cannot distinguish the two. What it can say is that the FY2027 consensus of $56.318 requires 43.1% earnings growth on 27.7% revenue growth — that is not an operating-leverage forecast, it is a margin-recovery forecast, and it must happen while revenue growth decelerates by thirteen points. That is the load-bearing assumption in the price and it is the one with the least evidence behind it.
3. Valuation — priced in or room?
At $1,888.50 (market capitalisation $95.74B, 50,697,000 diluted shares):
| Trailing | FY2026E | FY2027E | FY2028E | |
|---|---|---|---|---|
| Adjusted EPS | $37.89 | $39.343 (12) | $56.318 (14) | $78.075 (15) |
| Estimate range | — | $35.524–$49.438 (39%) | $47.572–$66.053 (39%) | $61.918–$94.438 (53%) |
| Growth | — | ~0% | +43.1% | +38.6% |
| P/E | 49.8x | 48.0x | 33.5x | 24.2x |
| Revenue | $28,893M (FY2025) | $40,737M (16) | $52,035M (17) | $64,065M (18) |
| Revenue growth | +39.1% | +41.0% | +27.7% | +23.1% |
| EV / EBITDA | 30.8x | — | — | — |
| Price / book | 13.1x | — | — | — |
| Return on equity | 29.6% | — | — | — |
| Dividend yield | 0% | — | — | — |
Estimate coverage is good and dispersion is wide. Twelve analysts on FY2026 adjusted EPS, fourteen on FY2027, fifteen on FY2028. The FY2027 range of $47.572 to $66.053 is 39% wide — narrower than FCX's 107% but far wider than Duke Energy's 1.4%, and it reflects genuine disagreement about when the margin turns. The FY2029 row shows $125.965 on 10 analysts, implying 61.3% growth — an acceleration off an already-elevated base that we treat as an extrapolation rather than a forecast — and FY2030's $175 rests on 6 analysts. Both are excluded from every conclusion.
est.ebitdaAvg and est.ebitAvg carry the fixed-ratio fabrication signature — REJECTED. ebitdaAvg is exactly 16.613% of revenueAvg and ebitAvg exactly 15.810% in FY2026, FY2027, FY2028, FY2029 and FY2030 alike. Five consecutive years to three decimal places. Note also that the implied EBIT margin of 15.8% is far above the 6.9% the company actually earned in the March quarter, so the block is not merely fabricated but optimistic. All forward valuation uses epsAvg, validated against the earn_cal actuals (the FY2025 estimate of $40.410 against a quarterly-actual sum of $39.40, a 2.5% overstatement — itself a small illustration of the miss pattern).
Enterprise value is usable. $104,410M, being market capitalisation plus net debt of $7,722M and there are no noncontrolling interests in this structure. EV/EBITDA of 30.8x on trailing EBITDA near $3.39 billion. Finance lease obligations of $2,329M ARE included in total debt of $11,392M — no lease double-count of the T or TJX class.
Peer context is poor. The vendor set is Coupang, Carvana, eBay, JD.com, Lowe's, Nike, PDD, Starbucks, Sea Limited and TJX. Only Coupang ($30.1B), Sea Limited ($66.9B) and arguably PDD ($32.4B) are comparable — emerging-market e-commerce platforms with adjacent fintech or logistics. Sea Limited is the single most apposite: a South-East Asian analogue running the same marketplace-plus-payments-plus-logistics model, and a direct competitor in Brazil through Shopee. Lowe's, Nike, Starbucks, TJX and Carvana are unrelated. No cross-multiple is drawn.
3a. What today's price assumes (the inversion)
At $1,888.50 — 48.0x FY2026 consensus, 33.5x FY2027, 30.8x EV/EBITDA — the price embeds:
- Adjusted EPS of $39.343 in 2026 and $56.318 in 2027. (Consensus; 12 and 14 analysts.) The March quarter delivered $8.23 and the June guide is $8.67, so the second half needs roughly $22.44 — a very heavy back-half weighting.
- The operating margin recovers from 6.9%. (Filing-derived; our inference.) This is the most fragile assumption in the price. FY2027's 43.1% earnings growth on 27.7% revenue growth is arithmetically impossible without margin expansion of several hundred basis points. Nothing in the nine quarters of data in this file shows the margin turning.
- Logistics capital expenditure stops compounding. (Our number, from
cf_a.) Capex has gone $455M → $509M → $860M → $1,343M in four years. The bull case requires it to plateau while volume keeps growing over it. A capital programme still tripling is an investment phase that has not ended. - The credit book behaves. (Filing-disclosed structure; our extrapolation.) Days of sales outstanding are 212.4 and average receivables $12,886M — this is Mercado Crédito, classified in the 10-Q as Merchant, Consumer, Credit cards and Asset-backed. The knowledge base claims net interest margin after losses is rising and bad-loan losses falling. That is checkable and unverified in this archive.
- The market keeps paying more than 30x forward. (Our number.) At 28x FY2027E the stock is $1,577; at 40x it is $2,253. Twelve months ago the market paid materially more; the de-rating has already been 25% from the high, and the multiple is the largest single swing factor in the outcome.
3b. The return bridge (why the multiple moves)
Expected return over the next twelve months decomposes as: adjusted EPS growth (+43.1%, from FY2026E $39.343 to FY2027E $56.318) + multiple COMPRESSION (48.0x on the forward year moving to roughly 38x on the then-forward year, −21%) + shareholder yield (ZERO) ≈ +13%.
A caution the addendum requires: the rolldown from 48.0x FY2026 to 33.5x FY2027 at a constant price is mechanical, not compression, and is not counted as such. What our base DOES assume is that the market pays 38x the FY2027 number a year from now, against the 48.0x it currently pays on FY2026 — genuine de-rating, on the ground that a company whose earnings compound at 40% deserves a lower multiple than one whose earnings are flat only if you believe the 40% is durable, and four consecutive misses argue for discounting it.
Note what this makes the return: all of it is earnings growth, roughly half of which is handed back in de-rating, and there is no yield whatsoever. MercadoLibre pays no dividend and repurchased $1 million of stock in fiscal 2025 — on a $95.7 billion market capitalisation, that is 0.001%. A shareholder here owns growth and nothing else, and if the growth is deferred the wait is uncompensated.
If the multiple fell to 28x FY2027E the price is $1,577 (−16.5%). If it held at 48x, $2,703 (+43.1%).
3c. Variant perception (where we differ, what would surprise)
- We are 1.2% below consensus and that is not the disagreement. Our $2,140 against a mean of $2,166.67 and a median of $2,150 — effectively identical. Where we differ from the street is not the number but the confidence attached to it: the FY2027 estimate that produces it has a 39% range and comes from a company that has missed four consecutive quarters. A fair value derived from an estimate with that track record deserves a Watch rather than a Buy, and that is the whole of our variant perception.
- We think the margin question is being answered too generously in both directions. The bulls treat 6.9% as temporary because Amazon's was; the bears treat it as permanent because the competition is real. Watchable number: the operating margin in tomorrow's print. A reading above 9% would be the first evidence of a turn; a reading below 7% for a second consecutive quarter would make the FY2027 consensus untenable.
- We think the vendor's cash-flow presentation is actively misleading and the company agrees. The payload reports a free-cash-flow yield of 11.18% and price-to-free-cash-flow of 8.94x — implying MercadoLibre is one of the cheapest cash generators in the market at 49.8x earnings. The 10-Q defines the company's own adjusted free cash flow as operating cash flow LESS customer funds held under regulatory restriction, LESS changes in loans receivable, LESS net proceeds from loans payable related to Fintech, "since we consider those liabilities as the working capital of the Fintech activities." The vendor's $10,773 million of fiscal-2025 free cash flow includes all of it.
incomeQualityTTMof 6.22 — operating cash flow at 6.2x net income — is the visible symptom. Anyone screening this name on free-cash-flow yield is looking at customer float. - We think Argentina is under-examined. It is 20.6% of revenue and grew 56.2%, the fastest of the three principal markets, in an economy with a history of currency re-basing. The payload contains no constant-currency disclosure and the archive's tables were not preserved.
- Positive surprise that would force a re-rate: an operating margin above 9% tomorrow with management framing the investment phase as ending; a break in the four-quarter miss streak; disclosed credit-book metrics showing rising net interest margin after losses, corroborating the knowledge base's specific claim; or the initiation of a buyback of any consequence.
- Negative surprise that would break the thesis: a fifth consecutive miss; an operating margin below 7% again; capital expenditure guided materially above the $1,343M fiscal-2025 run rate; any deterioration in the loan book, on a 212-day receivable cycle and $12,886M of average receivables; or evidence that Sea Limited's Shopee or Amazon is taking share in Brazil.
Synthos fair values
All three anchors are multiples of the FY2027 consensus adjusted EPS distribution (mean $56.318, low $47.572, high $66.053, 14 analysts).
- Bear ~$1,427 — 30.0x the FY2027 consensus LOW of $47.572, cross-checked at 36.3x FY2026E. 7.7% BELOW the 52-week low of $1,546.81. The scenario: the margin compression proves competitive rather than discretionary, the miss streak extends, FY2027 lands near the bottom of its range, and a de-rating growth stock trades at 30x a disappointing number. −24.4%.
- Base ~$2,140 — 38.0x the FY2027 consensus MEAN of $56.318, cross-checked at 54.4x FY2026E and 27.4x FY2028E. Sensitivity, stated openly: 33.5x — the current multiple held flat — gives $1,887, which is spot; 45x gives $2,534. The scenario: the margin turns during 2026, FY2027 delivers the consensus 43.1% earnings step, revenue keeps compounding above 25%, and the multiple de-rates modestly as growth is proven rather than assumed. +13.3%.
- Bull ~$2,972 — 45.0x the FY2027 consensus HIGH of $66.053, cross-checked at 38.1x FY2028E. 18.4% above the 52-week high of $2,510.97. The scenario: the logistics investment stops compounding, operating margin returns toward the 14.6% of fiscal 2023 on a revenue base 3.4x larger, the credit book seasons profitably, and the market pays a full growth multiple for the region's dominant platform. +57.4%.
Base is 13.3% above spot; asymmetry roughly 2.35:1 (24.4% down, 57.4% up), with zero shareholder yield. That asymmetry clears the arithmetic bar. What does not clear is the confidence: the base rests on an estimate this company has missed four times running, from a report that lands tomorrow, in a stock that pays nothing while you wait. Watch, with the trigger named in Section 6.
4. Knowledge base — seven claims, all bullish, from two channels
Raw hits: 7. Entity matches: 7. Used: 7. Discarded: 0. Distinct channels: 2.
The sweep ran case-sensitive entity tokens MELI, MercadoLibre, Mercado Libre and Mercado Pago, plus case-sensitive free-text patterns for the same, across all 51,928 distilled claims. It returned seven claims and ZERO free-text noise — the cleanest sweep in this batch, and worth recording because "Mercado" is a common Spanish noun and a looser pattern would have generated substantial noise.
Every one of the seven is bullish. The four most substantive, verbatim:
> 2026-03-08 · bullish · conviction 75 · horizon: thesis · entities: MercadoLibre, MELI · channel: invest_like_the_best · speaker: null
> "Flat 5-year stock hides a healthy business; heavy logistics capex is Amazon-style investment mode that will unlock profitability later."
> 2026-03-08 · bullish · conviction 70 · horizon: fact · entities: MercadoLibre, MELI · channel: invest_like_the_best · speaker: null
> "Revenue up ~400% over 5 years, GMV accelerating, and the credit book is improving (NIM after losses rising, bad-loan losses falling)."
> 2026-04-18 · bullish · conviction 80 · horizon: thesis · entities: MELI · channel: we_study_billionaires · speaker: Daniel
> "MercadoLibre runs Amazon's scale-economics-shared playbook; LatAm e-commerce penetration only 14-15% vs 25% US will close, MELI holds 30-35% Brazil share and is uniquely South-America-first."
> 2026-07-24 · bullish · conviction 62 · horizon: thesis · entities: MELI · channel: we_study_billionaires · speaker: Luke Hallard · speaker_role: independent
> "Mercado Libre is in great shape despite ~6% price over 5yr: GMV and revenue growing at accelerating rates (revenue +400% over 5yr), credit book healthy (NIM after losses rising, bad-loan losses falling), capex builds logistics moat — Amazon-like."
The remaining three: we_study_billionaires 2025-09-18 (conviction 95) naming MELI among seven long-held "100-baggers" alongside Amazon, Netflix, Nvidia, Tesla, Booking and Intuitive Surgical; we_study_billionaires 2026-04-18 (conviction 65) on next-generation warehouse robotics making e-commerce structurally more profitable, naming AMZN and MELI; and we_study_billionaires 2022-11-24 (conviction 65, speaker David Gardner) describing MELI as "a smaller younger version of Amazon at ~$40B market cap."
Concentration test — RUN and FAILED badly. we_study_billionaires supplies five of seven claims — 71%, more than twice the one-third threshold. Removing that channel leaves two claims, both from invest_like_the_best, both dated the same day. The lane therefore rests on effectively two independent voices, and conviction is scored moderate rather than high with breadth 2.
A second, less common problem: the lane contains substantive duplication. The 2026-07-24 we_study_billionaires claim and the 2026-03-08 invest_like_the_best claim make the same argument in nearly the same words — revenue up ~400% over five years, credit book healthy with net interest margin after losses rising and bad-loan losses falling, capex building a logistics moat, Amazon-like. Two channels four months apart restating one thesis is weaker evidence than two independent analyses reaching the same conclusion, and counting them as two claims overstates breadth.
Attribution note. Three of the seven carry named speakers — Daniel and Luke Hallard on we_study_billionaires, David Gardner on the same — and four carry speaker: null. No management voice appears anywhere in this lane, and no commercially interested party is identifiable. The one speaker_role present reads independent and is not relied upon.
GRADING THE LANE. The thesis is coherent, specific and unproven on this horizon. The claims span 2022 to 2026-07-24 — eleven days before this dive — and the specific assertion that the flat share price "hides a healthy business" has not yet been vindicated where it matters: the stock is down 20.5% over the trailing twelve months against SPY's +24.3%, and March-quarter net income FELL 15.6%. The most checkable element — the credit book with rising net interest margin after losses and falling bad-loan losses — cannot be verified from this archive, because the MELI filings contain no preserved statement tables. We report the lane as bullish, concentrated, partly duplicative, and awaiting evidence.
Conclusion. Breadth 2, claim count 7, net conviction positive-moderate.
5. Data integrity — what we rejected and why
Six findings. One would materially mislead a screen; two blocks passed tests that have failed elsewhere in this programme.
1. freeCashFlow and every derived measure include Mercado Pago's customer float and credit-book funding — REJECTED, and the company's own filing is the authority. The vendor reports fiscal-2025 operating cash flow of $12,116M against net income of $1,997M — incomeQualityTTM of 6.22 — and free cash flow of $10,773M, giving freeCashFlowYieldTTM of 11.18% and priceToFreeCashFlowRatioTTM of 8.94x on a stock trading at 49.8x earnings. The 10-Q defines the company's own measure and explicitly excludes exactly what the vendor includes:
> "Adjusted free cash flow represents cash from operating activities less the increase (decrease) in cash and cash equivalents and investments related to customer funds due to regulatory requirements and other restrictions... investments in property and equipment and intangible assets, changes in loans receivable, net and net proceeds from/payments on loans payable and other financial liabilities related to our Fintech solutions, since we consider those liabilities as the working capital of the Fintech activities."
A payments business holding customer balances shows those balances as operating cash inflows; they are not the company's money. freeCashFlowYieldTTM, priceToFreeCashFlowRatioTTM, evToFreeCashFlowTTM (9.75x), priceToOperatingCashFlowRatioTTM (8.01x), freeCashFlowToEquityTTM ($14,557M — 15.2% of market capitalisation), freeCashFlowToFirmTTM and incomeQualityTTM are all rejected. This is the same conceptual class as the bank and lender rejections elsewhere in this batch, and it is more dangerous here because the numbers look attractive rather than absurd.
2. est.ebitdaAvg and est.ebitAvg carry the fixed-ratio fabrication signature and are optimistic in level — REJECTED. ebitdaAvg is exactly 16.613% of revenueAvg and ebitAvg exactly 15.810% in FY2026, FY2027, FY2028, FY2029 and FY2030 alike. The implied EBIT margin of 15.8% compares with the 6.9% operating margin the company actually earned in the March quarter and the 11.1% it earned in fiscal 2025 — so the block is fabricated AND materially above the achieved level. The historical rows use different ratios (FY2023: EBITDA 6.48% and EBIT 3.68% of revenue), confirming the forward rows are constructed rather than reported. All forward valuation uses epsAvg.
3. daysOfSalesOutstandingTTM of 212.4 days and averageReceivablesTTM of $12,886M are a LOAN BOOK, not a trade balance — do not read as working-capital deterioration. The 10-Q states: "The Company classifies loans receivable as 'Merchant,' 'Consumer,' 'Credit cards' and 'Asset-backed.'" A 212-day receivable cycle on a marketplace that collects at the point of sale is not a collection problem; it is Mercado Crédito's portfolio sitting in the receivables line. operatingCycleTTM of 226.3 days and cashConversionCycleTTM of 123.3 days inherit the same issue. They are reported here for completeness and are not used as efficiency measures. The same applies to debtToEquityRatioTTM of 1.70x and financialLeverageRatioTTM of 6.45x: a material share of the $11,392M of debt funds the credit portfolio rather than the marketplace.
4. dividendPerShareTTM reads 0 while profile.lastDividend reads 0.6 — an internal contradiction. dividendYieldTTM and dividendPayoutRatioTTM are both 0. We treat MercadoLibre as paying no dividend, which is consistent with the two zero fields and with dividendPaidAndCapexCoverageRatioTTM being identical to capitalExpenditureCoverageRatioTTM (9.643) — i.e. the dividend term in that ratio is zero. The lastDividend field is not used.
5. priceToEarningsGrowthRatioTTM is NEGATIVE at −7.32 and is meaningless. A PEG ratio requires positive earnings growth; trailing earnings fell. priceToEarningsDilutedGrowthRatioTTM of −7.34 is the same. Neither is used. forwardPriceToEarningsGrowthRatioTTM of 1.156 is derived from the rejected est block and is likewise not used.
6. seg_prod reconciles exactly but carries the wrong segmentation for the years that matter. FY2025's Product $3,607M plus Service $25,286M equals $28,893M — exact. But a two-line product/service split tells a reader nothing about a company whose central question is the balance between Commerce and Fintech. The FY2021 and FY2020 entries DO carry the Commerce/Fintech split (FY2021: Commerce $4,635.4M, Fintech $2,434.0M) and it disappears from FY2022 onward, and the FY2019 and earlier entries use yet a third structure (Marketplace/Nonmarketplace). Three incompatible bases across the series, with the most informative one available only for the least relevant years. No Commerce/Fintech figure in this dive comes from the payload.
Tests that PASSED, and they are worth as much as the failures. seg_geo reconciles to the reported revenue line EXACTLY in both fiscal 2025 ($28,893M) and fiscal 2024 ($20,777M) — Brazil, Mexico, Argentina and Other Countries, with no missing bucket and no double-count. seg_prod also reconciles exactly in both years. inc_q passed the MO/APP two-basis test — all nine quarters are on one basis, the four fiscal-2025 quarters sum to $28,893M exactly matching inc_a, and the four most recent match earn_cal.revenueActual exactly. earn_cal passed the SPGI mixed-basis test — all six actual-versus-estimate pairs are on the adjusted-EPS basis. The share count is exact and static: inc_a.weightedAverageShsOutDil is 50,697,000 in both FY2025 and FY2024, against 50,697,000 implied by market capitalisation ÷ price — an exact match, and evidence of genuinely zero dilution over two years despite $303M of annual stock-based compensation. There are no noncontrolling interests, so no NCI omission in enterprise value is possible. Finance lease obligations of $2,329M are correctly included in total debt of $11,392M — no lease double-count.
Also checked and recorded. quote.yearHigh/yearLow ($2,548.50 / $1,495.00) against tech.hi52/lo52 ($2,510.97 / $1,546.81) — a 1.5% and 3.5% discrepancy; tech is used throughout. Note that tech.max_dd_from_peak of −27.74% is LARGER than tech.pct_from_hi of −24.79%, which is internally consistent (the stock fell further at some point than it currently sits below the high) and is the only name in this batch where the two differ materially — evidence of a genuine intra-year trough. effectiveTaxRateTTM of 15.7% is well below the FY2025 statutory-equivalent rate implied by incomeTaxExpense of $845M on incomeBeforeTax of $2,842M (29.7%), and no explanation is available in the payload; the trailing figure is not used.
returnOnEquityTTM — source stated. Present in km_ttm at 29.58% and absent from ratios_ttm. Used as reported; note the denominator is a small equity base ($6,748M) on a $42,667M balance sheet, so the figure flatters relative to a less levered peer.
Vendor composite rating — noted, low weight. B / 3 overall, with 5 out of 5 on discounted cash flow and 5 out of 5 on return on equity, against 1 out of 5 on price-to-earnings, price-to-book and debt-to-equity. The discounted-cash-flow score rests on the free-cash-flow figure rejected in finding 1 and carries no weight. The debt-to-equity score of 1 penalises a company for funding a credit book.
Non-equity tripwire — checked and passed. MELI is common stock, Nasdaq-listed, 50,697,000 diluted shares. Price of $1,888.50 is not par-like; beta is 1.312; there is no dividend; volume was 432,433 shares (~$817M of turnover); the 52-week band of $1,546.81 to $2,510.97 is a 62% range. This is common equity.
6. Verdict, kill-criteria and flip conditions
Watch.
The case for interest, and it is substantial. MercadoLibre is the dominant e-commerce and payments platform across a region where online penetration is roughly 14-15% against 25% in the United States, holds an estimated 30-35% share in Brazil, and grew revenue 49.0% in the March quarter — the fastest in this batch by a wide margin. It earns 29.6% on equity, has not diluted its shareholders in two years (50,697,000 diluted shares in both FY2025 and FY2024), and reconciles both its segment blocks to the dollar. The stock is 24.8% below its 52-week high and sits on its 200-day average. Our own knowledge base is uniformly bullish — seven claims, none discarded, zero homograph noise. Base fair value $2,140 is 13.3% above spot with 2.35:1 asymmetry.
The case for waiting, which is the operative one. The earnings line has stopped growing. March-quarter operating income fell 19.9% on revenue up 49.0%; the operating margin went from 12.9% to 6.9%; and consensus FY2026 adjusted EPS of $39.34 is flat against the $39.40 delivered in 2025. The FY2027 consensus of $56.318 requires 43.1% earnings growth on 27.7% revenue growth — a pure margin-recovery forecast — from a company that has MISSED consensus in four consecutive quarters, by 13.6%, 8.8%, 3.7% and 5.9%. It reports tomorrow, and consensus itself expects a 15.9% earnings decline in that quarter. There is no yield whatsoever: no dividend, and a fiscal-2025 buyback of $1 million on a $95.7 billion company. And the knowledge-base lane, though uniformly positive, collapses to two claims from one channel once its dominant voice is removed, and contains a near-verbatim duplication that inflates apparent breadth.
Watch rather than Hold, because there is a genuine catalyst path and a specific price at which the arithmetic becomes compelling. Watch rather than Buy, because a 13.3% base case resting on a four-times-missed estimate, the day before a print, in a stock that pays nothing, is not a position.
The named trigger: $1,600. At $1,600 the same FY2027 consensus of $56.318 is 28.4x, the base case becomes +33.8%, asymmetry improves past 3:1, and the entry sits 3.4% above the 52-week low of $1,546.81. That is where this becomes Buy — Tactical on arithmetic alone, absent an adverse development below.
Pre-registered conditions that would move MELI to Buy — Tactical:
- A price at or below $1,600, on no adverse fundamental news.
- An operating margin above 9% in any quarter, breaking the four-year decline from the 14.6% of fiscal 2023 — this is the single highest-value disclosure available.
- A break in the four-quarter miss streak, with adjusted EPS at or above consensus.
- Capital expenditure guided flat or down against the fiscal-2025 run rate of $1,343M, which would evidence that the investment phase is ending rather than compounding.
- Disclosed credit-book metrics corroborating the knowledge base's specific claim — rising net interest margin after losses, falling bad-loan losses.
- The initiation of a share repurchase of any consequence, which would put a floor under the return for the first time.
Pre-registered conditions that would move MELI to Avoid:
- A fifth consecutive miss, or an operating margin below 7% for a second straight quarter.
- Revenue growth decelerating below 25% while the margin is still compressing — that combination would indicate competitive rather than investment spending.
- Any deterioration in the loan book: rising provisions, a growing allowance against the $12,886M average receivable base, or an extension of the 212-day cycle.
- A material Argentine currency event, on 20.6% of revenue.
- Evidence of share loss to Sea Limited's Shopee or to Amazon in Brazil or Mexico.
- A break below the 52-week low of $1,546.81 on company-specific news.
Where MELI fits in the Synthos Framework Portfolio. The international-growth sleeve — on the watchlist at zero weight today, with a $1,600 limit and a review scheduled immediately after the 2026-08-05 print. Sizing note: this is the highest-growth and highest-exponential name in this batch and it pays nothing while you wait, so the entry price is the entire risk control. A name with no yield and a broken forecasting record must be bought at a price, not at a moment. Logged as a tracked Synthos call (Watch) as of 2026-08-04 at $1,888.50, with the fair-value anchors, the $1,600 trigger and both condition sets gradeable.
Single biggest risk: that the margin compression is competitive rather than discretionary. The entire valuation rests on FY2027 adjusted EPS of $56.318, which requires 43.1% earnings growth on 27.7% revenue growth — arithmetically impossible without several hundred basis points of margin expansion. The operating margin has fallen every year since fiscal 2023, from 14.6% to 12.7% to 11.1% to 6.9% in the most recent quarter, while capital expenditure has nearly tripled to $1,343 million and research and development runs at 7.9% of revenue. The knowledge base says this is Amazon-style investment that unlocks profitability later, and that is a coherent and historically-supported argument. The alternative reading — that Amazon and Sea Limited are forcing permanent spending in the same markets — is equally consistent with every number in this file. Nothing in nine quarters of data yet distinguishes them, and the first evidence either way arrives tomorrow. That is precisely why this is a Watch with a dated review rather than a position.
Provenance & disclosures
- Traceability: 7 knowledge-base claims name MercadoLibre out of 51,928 distilled claims (raw hits 7, entity matches 7, used 7, discarded 0; breadth 2, net conviction positive-moderate). ALL SEVEN ARE BULLISH. The principal claims:
invest_like_the_best2026-03-08 (conviction 75, "Flat 5-year stock hides a healthy business; heavy logistics capex is Amazon-style investment mode that will unlock profitability later") and its companion at conviction 70 ("Revenue up ~400% over 5 years, GMV accelerating, and the credit book is improving (NIM after losses rising, bad-loan losses falling)");we_study_billionaires2026-04-18 (conviction 80, speaker Daniel, "LatAm e-commerce penetration only 14-15% vs 25% US will close, MELI holds 30-35% Brazil share");we_study_billionaires2026-07-24 (conviction 62, speaker Luke Hallard, a near-verbatim restatement of the invest_like_the_best thesis);we_study_billionaires2025-09-18 (conviction 95, MELI among seven long-held "100-baggers");we_study_billionaires2026-04-18 (conviction 65, warehouse robotics, naming AMZN and MELI); andwe_study_billionaires2022-11-24 (conviction 65, speaker David Gardner, "a smaller younger version of Amazon at ~$40B market cap"). CONCENTRATION SENSITIVITY TEST — RUN and FAILED BADLY:we_study_billionairessupplies FIVE of SEVEN claims (71%), more than twice the one-third threshold; removing it leaves two claims from a single channel dated the same day. A second weakness is recorded: the 2026-07-24 and 2026-03-08 claims make the same argument in nearly the same words — revenue +400% over five years, rising NIM after losses, falling bad-loan losses, capex building a logistics moat — so counting them as two independent claims overstates breadth. Three of seven carry named speakers; four carryspeaker: null; no management voice appears and no commercially interested party is identifiable. HOMOGRAPH CHECK — the cleanest sweep in this batch: ZERO free-text noise despite "Mercado" being a common Spanish noun, and nothing was discarded. GRADING: the thesis is coherent and unproven on this horizon — the newest claim is eleven days old, and over the trailing twelve months the stock is DOWN 20.5% against SPY's PLUS 24.3% while March-quarter net income fell 15.6%. The most checkable element — the credit book's net interest margin after losses — cannot be verified from this archive, because the MELI filings contain no preserved statement tables. All quotes are verbatim from the stored claim text. - Data as-of: fundamentals — the four-country segment structure, the direct-contribution measure, the loans-receivable classification and the adjusted-free-cash-flow definition through 2026-03-31, from the 10-Q filed 2026-05-08; the chief-executive and executive-chairman roles and the 2026 Long Term Retention Program from the 8-K filed 2026-04-03 · estimates 2026-08-04 · prices 2026-08-04, quote timestamp 1785873601 = 2026-08-04T20:00:01Z ($1,888.50, −0.54%; 50-DMA $1,736.43; 200-DMA $1,890.37; RSI 59.6; MACD +39.96; max drawdown −27.7%) · knowledge-base claims 2026-08-04. MercadoLibre's fiscal year is the calendar year. All figures come from the Synthos vendor data file for MELI or from the SEC filings in the MELI archive; no figure comes from memory, recall or external retrieval.
- Filing archive contents and its limits: 10-K filed 2026-02-25 (fiscal 2025); 10-Q filed 2025-10-30 (September 2025 quarter); 10-Q filed 2026-05-08 (March 2026 quarter — the source of the segment definition, the loans-receivable classification and the adjusted-free-cash-flow definition); 8-K filed 2026-04-03 (2026 Bonus Program and 2026 Long Term Retention Program — the source confirming Ariel Szarfsztejn as Chief Executive Officer and Marcos Galperin as Executive Chairman, alongside Osvaldo Giménez as Fintech President and Daniel Rabinovich as Technology and Operations); 8-K filed 2026-05-07 (first-quarter results, Item 2.02 — the press release is Exhibit 99.1 and is not in the extracted text); 8-K filed 2026-06-12 (annual meeting voting results). NO file in this archive contains a
[TABLE]marker — statement tables were not preserved for this name, verified bygrep -creturning zero on all six files. Every filing-sourced figure in this dive is therefore a prose statement, and no primary statement line item or credit-book metric was independently verified. There is no June-quarter 10-Q — those results are released tomorrow. - Where the filings contradicted or corrected the vendor: free cash flow — the vendor reports fiscal-2025 free cash flow of $10,773M and a
freeCashFlowYieldTTMof 11.18% for a stock at 49.8x earnings, while the 10-Q defines the company's own adjusted free cash flow as operating cash flow "less the increase (decrease) in cash and cash equivalents and investments related to customer funds due to regulatory requirements and other restrictions... changes in loans receivable, net and net proceeds from/payments on loans payable and other financial liabilities related to our Fintech solutions, since we consider those liabilities as the working capital of the Fintech activities" — the vendor includes all of it, andincomeQualityTTMof 6.22 is the visible symptom; the receivable cycle —daysOfSalesOutstandingTTMof 212.4 andaverageReceivablesTTMof $12,886M are Mercado Crédito's loan book, which the 10-Q classifies as "Merchant, Consumer, Credit cards and Asset-backed", not a trade balance; the segment basis — the 10-Q's four country segments evaluated on "direct contribution" againstseg_prod's two-line Product/Service split; and the governance structure — the 8-K names Ariel Szarfsztejn as Chief Executive Officer with founder Marcos Galperin as Executive Chairman, a transition absent from every vendor field except theprofile.ceostring. Where vendor and filing AGREED — recorded, including three tests that have failed on other names:seg_georeconciles to the reported revenue line EXACTLY in both fiscal 2025 ($28,893M) and fiscal 2024 ($20,777M);seg_prodreconciles exactly in both years;inc_qpassed the two-basis test (nine quarters on one basis, the four fiscal-2025 quarters summing to $28,893M exactly matchinginc_a, and the four most recent matchingearn_cal.revenueActualexactly);earn_calpassed the mixed-basis test (all six pairs on the adjusted-EPS basis); and the share count is exact and static at 50,697,000 diluted in both FY2025 and FY2024, matching the market-capitalisation-implied count — genuinely zero dilution over two years despite $303M of annual stock-based compensation. - Rejected as a class — the fintech-float cascade:
freeCashFlowYieldTTM(11.18%),priceToFreeCashFlowRatioTTM(8.94x),evToFreeCashFlowTTM(9.75x),priceToOperatingCashFlowRatioTTM(8.01x),evToOperatingCashFlowTTM(8.74x),freeCashFlowToEquityTTM($14,557M, 15.2% of market capitalisation),freeCashFlowToFirmTTM($11,395M),incomeQualityTTM(6.22) andoperatingCashFlowSalesRatioTTM(37.6%), because operating cash flow includes customer funds held under regulatory restriction and changes in the credit book that the company itself excludes. The vendor composite rating's discounted-cash-flow sub-score of 5 out of 5 rests on the same figure and carries no weight. Separately rejected:est.ebitdaAvgandest.ebitAvgfor the fixed-ratio signature (exactly 16.613% and 15.810% ofrevenueAvgin five consecutive forward years, against an actual 6.9% operating margin in the March quarter);priceToEarningsGrowthRatioTTM(−7.32, negative and therefore meaningless) andpriceToEarningsDilutedGrowthRatioTTM(−7.34);profile.lastDividendof 0.6, which contradictsdividendPerShareTTM,dividendYieldTTManddividendPayoutRatioTTMall reading zero; andeffectiveTaxRateTTMof 15.7%, unreconciled against the 29.7% implied by the fiscal-2025 income statement.daysOfSalesOutstandingTTM,operatingCycleTTMandcashConversionCycleTTMare reported for completeness and are not used as efficiency measures, because the receivable is a loan book. returnOnEquityTTM— source stated: present inkm_ttmat 29.58% and absent fromratios_ttm. Used as reported, with the note that the denominator is a $6,748M equity base on a $42,667M balance sheet.- Estimate coverage: 12 analysts on FY2026 adjusted EPS, 14 on FY2027 (the anchor for all three fair values) and 15 on FY2028. The FY2027 range is $47.572 to $66.053, a 39% spread. FY2029 rests on 10 analysts and implies 61.3% growth — an acceleration off an already-elevated base that we treat as extrapolation — and FY2030 rests on 6; both are excluded from every conclusion. Note that the FY2025
epsAvgof $40.410 overstated the actual quarterly-actual sum of $39.40 by 2.5%, which is itself a small illustration of the four-quarter miss pattern. - Peer note: the vendor peer set is Coupang, Carvana, eBay, JD.com, Lowe's, Nike, PDD, Starbucks, Sea Limited and TJX. Only Coupang, Sea Limited and arguably PDD are comparable emerging-market e-commerce platforms with adjacent fintech or logistics; Sea Limited is the single most apposite, being both a structural analogue and a direct competitor in Brazil through Shopee. Lowe's, Nike, Starbucks, TJX and Carvana are unrelated. No cross-multiple is drawn.
- Fair-value caveat: the $1,427 / $2,140 / $2,972 anchors are multiples of the FY2027 consensus adjusted EPS distribution — 30.0x the low of $47.572, 38.0x the mean of $56.318, and 45.0x the high of $66.053 — cross-checked against FY2026 and FY2028. Stated arithmetic, not a discounted cash flow. No cash-flow cross-check is applied, because the vendor's free-cash-flow figures include customer float that the company itself excludes. The base is sensitivity-disclosed: 33.5x — the current FY2027 multiple held flat — gives $1,887, which is spot; 45x gives $2,534. The base assumes genuine multiple DE-RATING from the 48.0x currently paid on FY2026 to roughly 38x on FY2027 a year hence; the mechanical rolldown from 48.0x to 33.5x at a constant price is NOT counted as compression. All three anchors rest on an estimate the company has missed in four consecutive quarters, and that fact — not the level of the number — is why the verdict is Watch.
- Timing — and it governs the verdict. Second-quarter 2026 results are released 2026-08-05, ONE DAY after this dive, against consensus of $8.67 adjusted EPS and $9,742M of revenue — implying 43.5% revenue growth and a 15.9% EPS DECLINE against the year-ago quarter. MercadoLibre has MISSED consensus in FOUR CONSECUTIVE QUARTERS — by 13.6% (2025-08-04), 8.8% (2025-10-29), 3.7% (2026-02-24) and 5.9% (2026-05-07) — and the only beat in the file is from 2025-05-07, fifteen months ago. No position is taken into tomorrow's print, and the Watch verdict carries a review scheduled immediately after it. The most recent insider transactions are a 2026-06-11 open-market purchase of 124.64 shares at $1,604.62 by Marcelo Melamud, SVP and Chief Accounting Officer — approximately $200,000, small but the only directional transaction in the file — alongside a Form 3, a director gift and five director restricted-stock-unit awards dated 2026-06-12. 2026-08-04 carried no company-specific filing.
- Accessibility note: no information in this dive is conveyed by colour. All emphasis is carried by bold text, table structure and explicit labelling.
- Not investment advice. Independent research, educational and informational only, never personalised. No recommendation to buy, sell or hold any security is made to any person.
- Version: 2026-08-04-full.