MELI — MercadoLibre
Five businesses in a trench coat. Somehow the trench coat is the moat.
MercadoLibre is marketplace, logistics, payments, credit and ads all feeding each other across Latin America. It is messy to summarize because the mess is kind of the moat. I am fine with management spending aggressively to deepen that ecosystem. I am not fine with 'we are investing for the long term' becoming a permanent excuse for owner earnings never showing up.
Every time I try to summarize MercadoLibre I end up listing half the company: marketplace, logistics, payments, credit, ads, asset management.
Why I care about this one
Latin America still has a long runway in both e-commerce and digital finance, and MELI is attacking both with one customer relationship.
Commerce brings users in. Logistics improves the shopping experience. Mercado Pago increases frequency. Credit can deepen the relationship. Ads monetize the same traffic again.
That flywheel is why I am willing to tolerate a company that currently looks worse on margins than it does on almost every engagement metric.
I am not willing to tolerate that forever.
Q2 was ridiculous on growth and deliberately ugly on profitability.
Revenue plus financial income reached $10.17B, up 50% year over year. GMV grew 36% FX-neutral to $21.9B. TPV reached $101B, up 56%. The credit portfolio passed $16B and grew 75%. (Q2 shareholder letter)
Operating margin was 6.7%.
That last number is not an accident. Management has been lowering shipping thresholds, subsidizing parts of the marketplace value proposition, pushing cards and credit harder and spending to acquire users in payments.
So far the engagement data says the money is doing something useful: users are buying more frequently, using more categories and entrusting more money to Mercado Pago.
That is what I want to keep seeing. Growth alone is not the proof. Better user behavior is.
Q2 2026
Primary sources: Q2 2026 shareholder letter ↗ · Quarterly results ↗
Q2 showed strong commerce and fintech growth while the company continues reinvesting heavily. My central question is whether credit economics and operating leverage can turn that growth into earnings: the five-year Base is 28% annual EPS growth at 27x, roughly $3,680. The position remains meaningful within the 4%–7% sizing framework.
My position
What I actually bought and sold.
| Date | Type | Account | Dollars | Shares | Price | Return now* |
|---|---|---|---|---|---|---|
| Dec 12, 2025 | Buy | Taxable | $1,799 | 0.888 | $2,026 | -8.3% |
| Mar 9, 2026 | Buy | Roth IRA | $5,000 | 2.852 | $1,753 | +6.0% |
| Mar 13, 2026 | Buy | Roth IRA | $3,000 | 1.799 | $1,668 | +11.4% |
| Mar 16, 2026 | Buy | Taxable | $2,499 | 1.441 | $1,734 | +7.1% |
| May 15, 2026 | Buy | Taxable | $4,195 | 2.639 | $1,589 | +16.9% |
*Return now uses the latest available quote for buys and reinvestments. Sales and fees show “—”.
Valuation & scenario math
MELI is one of the few names where I still like a drawdown trigger—but only as a trigger.
The rolling-high framework gets my attention around 20%, 30% and 40% drawdowns. It does not tell me the stock is cheap.
For the actual model, I use roughly $40 of normalized current EPS power and ask what per-share earnings could look like in five years.
Bear assumes 15% EPS growth at 20x, Base assumes 28% at 27x, and Bull assumes 36% at 33x.
The Base still lands around $3,680 per share, or roughly 15% annualized from the Oct. 6 reference close.
MELI can plausibly compound earnings at an unusual rate, but that growth has to come with healthy credit performance and actual margin conversion. At 27x in five years, I am still paying a premium for that quality.
At 24x instead of 27x, the same Base earnings path still gives me roughly a 12% annualized return.
That is the bet I am more comfortable calling Base.
20% / 30% / 40% drawdown trigger from the Sep. 25 rolling high
Starter zone · 8.9% below $2,039 starter
| Band | Buy price | Base return to Sep 2031 |
|---|---|---|
| Starter | $2,039 | +12.7% |
| Normal | $1,784 | +15.7% |
| Aggressive | $1,529 | +19.4% |
A drawdown is an invitation to rerun the math. It is not proof the stock got cheap.
This is a five-year earnings-power model. The growth today is absurd. The margin today is not. So the return case only works if all of this ecosystem spending eventually becomes much more per-share earnings rather than just more scale.
| Normalized EPS today | $40 |
| Annual EPS growth | 28% |
| EPS by Sep 25, 2031 | ~$136 |
| Terminal P/E | 27x |
| Implied base price | ~$3,680/share |
The base case is basically ~$136 of Sep 25, 2031 EPS at 27x, or about $3,680/share. Same math, just without making the reader decode equation soup.
| Case | What has to be true | Target range | Annualized return from today |
|---|---|---|---|
| Bear | 15% annual EPS growth · 20x terminal P/E. Commerce and fintech keep growing, but credit costs and persistent reinvestment prevent much margin expansion and the market stops paying a premium for the story. | $1,473–$1,730 center ~$1,602 | -3.0% |
| Base | 28% annual EPS growth · 27x terminal P/E. Commerce and fintech keep compounding rapidly, credit stays disciplined and operating leverage finally turns ecosystem scale into much higher per-share earnings. The endpoint still deserves a premium, but not an automatic 28x after five years of extraordinary compounding. | $3,386–$3,975 center ~$3,680 | +14.8% |
| Bull | 36% annual EPS growth · 33x terminal P/E. MELI keeps taking share across commerce and financial services, credit economics hold and the ecosystem produces substantial margin expansion without losing growth. | $5,592–$6,565 center ~$6,078 | +27.0% |
Returns use the current quote and the time remaining to Sep 25, 2031. The scenario targets stay fixed until the underwriting changes.
Capital allocation
The reinvestment looks productive. Credit is the thing that stops me getting too cute with the score.
Commerce frequency, fintech usage and broader engagement are responding to the spending, which earns management a lot of credit. The deductions are mostly credit/balance-sheet complexity, limited proof around mature excess-cash deployment, and the fact management is funding a lot of capital-hungry fronts at once.
Stage-relative, not some objective rating-agency grade. The number is there to force the judgment into the open. See the rubric →
What management is actually doing with the cash
MELI is spending in the places I would expect: shipping, pricing, cards, credit and user acquisition, all inside the same ecosystem. That is the good version. The uncomfortable version is that logistics and credit can consume a lot of capital before the economics are obvious. I will happily accept ugly quarterly margins if user behavior and underwriting keep proving the spend is working. I will not accept 'long term' forever.
What I’m doing with the next dollar
MELI is already a meaningful position inside the 4%–7% target.
So “I love the business” is not enough.
The live valuation above tells me whether the return is attractive enough today. The next dollar still has to compete with the fact that I already own a real position, credit is expanding quickly and other compounders may offer similar returns with less execution risk.
At this size I can still add if the expected return gets better. I just do not have some missing MELI-shaped hole in the portfolio that needs filling.
What can break this
Credit is the easiest way for this story to become less fun very quickly.
The portfolio is growing much faster than the company, and even disciplined underwriting can look brilliant while the credit environment is friendly. I care about spreads after losses, not just loan growth.
There is also the broader problem of trying to compound across Brazil, Mexico, Argentina and the rest of Latin America while dealing with currency swings, regulation, logistics intensity and very different local competitive markets.
And eventually, yes, I want margins.
I do not care if management makes this quarter look pretty. I do care whether today’s spending actually buys a higher long-term earnings ceiling. “We are investing” cannot become the permanent answer to why the margin never shows up.
More engagement with controlled credit losses and visible operating leverage would make the current spending look increasingly smart.
The opposite is pretty simple: credit outruns underwriting, the marketplace spend stops changing behavior, or spectacular revenue growth keeps failing to produce a believable path to much better owner earnings.
Latin American volatility by itself is not the kill thesis.
The flywheel looking busy while the economics stop getting better is.
The evidence I’m watching
Commerce and fintech engagement keep deepening, credit losses stay controlled and the current reinvestment wave starts producing visible operating leverage without slowing the flywheel.
Credit growth outruns underwriting discipline, marketplace/logistics spend stops improving user behavior, or spectacular top-line growth keeps failing to become a credible path to much higher owner earnings.