Every few weeks somebody sends me a screenshot of an Airbnb listing in Roma Norte and asks whether they should buy the apartment two doors down. The screenshot always shows the nightly rate. It almost never shows the calendar, and the calendar is where the money actually lives.
I should say up front where I sit. I run StayWork CDMX, the guest-facing side of a furnished apartment operation in Roma Norte and Narvarte, and I also run Roma Norte Host, which manages units for owners who do not want to do any of this themselves. Same operator, two brands. That is the only reason I can talk about the cost side with any confidence: I pay the cleaners, argue about the cuota de mantenimiento, and watch what happens to a calendar in the second week of September.
This article is about the model, not about promising you a return. I am deliberately not going to invent an average nightly rate for Condesa or a yield figure for Narvarte, because anyone quoting those to a stranger on the internet is guessing. What I can give you is the arithmetic, the costs that eat it, and enough neighborhood texture to know which questions to ask about a specific building.
Quick answer
Is a Mexico City Airbnb still profitable in 2026?
Sometimes, and much less automatically than in 2019. Three things changed. The city introduced a registry for eventual tourist stays with a night cap that limits a registered property to roughly half the calendar year, so “365 nights” is no longer a defensible planning assumption. Purchase prices in the central colonias rose faster than nightly rates did. And the operating cost stack got heavier as the market professionalized.
The units that still work are the ones bought at a price that survives a conservative available-nights assumption, in buildings whose rules and cuota do not fight you, run by someone who treats it as an operation rather than a passive asset. The units that fail are usually bought on a spreadsheet that assumed a full year.
Not investment, legal, or tax advice. I am an operator, not your accountant. Anything below that touches money or law should be checked with a Mexican contador and, for the building rules, a Mexican lawyer.
The revenue equation, and the variable everyone gets wrong
Gross revenue on a short-term rental is three numbers multiplied together:
average daily rate × occupancy × available nights
Most first-time investors obsess over the first number. They screenshot the highest rate they can find in the neighborhood and build a year around it. The second number gets a round guess, usually something optimistic like 80%. The third number they do not think about at all, because in most of the world it is simply 365.
In Mexico City it is not 365, and that is the whole story.
The reform to the Ley de Turismo introduced a registry for what the law calls an estancia turística eventual, along with a cap of roughly 183 nights per year per registered property. Treat that number as approximate rather than exact, because the way nights are counted and the way platforms enforce it have both moved. The cap has been challenged through hundreds of amparos, and litigation is still working through the courts. Nothing about it is settled, and an amparo granted to some other owner is not a permit for your unit. If you are relying on someone else’s suspension, you are relying on a document with someone else’s name on it.
So model available nights as a range, not a constant. Run your numbers three ways: at a full year, at roughly half a year, and somewhere in between. If the deal only works at 365, it is not a deal, it is a bet on litigation.
That single change reorders everything downstream. When available nights are capped, a marginal peso of nightly rate matters more than it used to, because you cannot make up a soft rate with volume. It also makes vacancy more expensive in relative terms, and it makes the medium-term lane much more interesting, which I will get to.
For a fuller treatment of the registry itself and what compliance looks like in practice, I wrote a separate piece on whether your Mexico City Airbnb is legal in 2026 .
A worked example you are supposed to overwrite
The numbers below are invented. I picked round figures so the arithmetic is legible. They are not market data for any neighborhood, and you should replace every one of them with a real quote before you make a decision.
Hypothetical unit. Every input is a placeholder, not a market rate.
| Line | Placeholder input | Where your real number comes from |
|---|---|---|
| Nightly rate | 2,000 MXN | Comparable listings in the same building type, checked across a full year |
| Occupancy of available nights | 70% | Booked nights divided by nights you actually offered |
| Available nights | 183 | Your own regulatory assumption, run at three levels |
| Gross | 2,000 × 0.70 × 183 = 256,200 MXN | |
| Platform and payment fees | minus ~15% | Platform’s published host fee plus card and FX costs |
| Management | minus your manager’s published commission | Roma Norte Host publishes 20% of lodging revenue with no setup fee and no fixed monthly; get any quote’s exact calculation base in writing |
| Cleaning and linen | minus per-turnover cost × turnovers | Quote per clean, then multiply by realistic stay length |
| Utilities and internet | minus monthly fixed | Prior owner’s bills, not an estimate |
| Cuota de mantenimiento | minus monthly fixed | The administración, in writing, with derramas history |
| Furnishing amortization | minus annual replacement reserve | Your furnishing quote divided over three to five years |
| Taxes | minus whatever your contador says | ISR, IVA, the CDMX lodging tax, and platform retentions |
Two things usually surprise people when they fill this in honestly.
The first is how much the cost stack takes. Between platform fees, management, per-turnover cleaning, utilities, the cuota, furniture that wears out faster than you think, and tax, the distance between gross and what lands in your account is wide. Furnishing amortization is the line people skip most often. A sofa in a unit doing forty turnovers a year is not a ten-year sofa. Neither is the mattress, and guests notice the mattress before they notice anything else.
The second is what happens to the cuota de mantenimiento when it is spread over 183 nights instead of 365. A fixed monthly charge does not care whether you are allowed to rent. In a building with a heavy cuota, a fancy lobby, and a pool nobody uses, that fixed cost can quietly become the reason the unit does not clear. Ask for the last two years of ordinary cuota plus any derramas, which are the special assessments for elevator repairs, façade work, or a cistern that finally gave up.
There is a working calculator for the short-term versus traditional-rental comparison on the management side of the business: run your own numbers here . It exists precisely because I got tired of rebuilding the same spreadsheet in emails.
The 21-night lane changes the shape of the problem
Everything above assumes a nightly-churn model. There is another way to run the same apartment, and it is what we actually operate: stays of 21 nights and up, mostly 30 to 90 days, for remote workers, medical staff, corporate placements, and people relocating.
The economics differ in ways that matter to an investor.
Turnover cost collapses. One clean per guest instead of one clean every two or three nights takes an entire cost line and shrinks it. Wear on furniture drops for the same reason, since most damage happens on arrival and departure days.
Occupancy behaves differently. Instead of filling a calendar with many small bookings that each need marketing, you fill it with a few large blocks. Fewer blocks means each vacancy hurts more, so the risk profile is lumpier even when the average is better. A three-week gap between two two-month guests is a real event, not a rounding error.
Rate per night is lower. Nobody pays a weekend nightly rate for night forty-five. You trade rate for volume and stability, and whether that trade is good depends entirely on what you paid for the unit.
Regulatory exposure is different too. The night cap and the registry are built around eventual tourist stays. Longer furnished tenancies sit in a different part of the picture, which is a large part of why professional operators in this city drifted toward the monthly market as the rules tightened. I am not going to tell you the boundary is crisp, because it is the kind of thing a Mexican lawyer should draw for your specific setup. But the direction of travel is not subtle: the longer your average stay, the less of your business model depends on how the amparos resolve.
Practically, this also changes what you should buy. A one-bedroom with a genuine desk, a quiet interior-facing bedroom, a working kitchen and reliable fiber will outperform a photogenic studio with a great balcony and nowhere to sit for eight hours. Guests staying two months buy a life, not a weekend.
Neighborhood by neighborhood
What follows is character, not data. Relative price positioning is safe to state. Absolute figures are not, so I am not stating them.
Six central colonias, by demand profile and building stock
| Colonia | Who books there | Seasonality | Building stock | Entry price, relative |
|---|---|---|---|---|
| Roma Norte | Remote workers, design and food tourism, first-time visitors, long creative stays | Steadiest of the six; softens least in the slow months | Porfirian and art deco conversions, small boutique new-builds, plenty of walk-ups | Upper tier, below Polanco |
| Condesa | Couples, park-oriented long stays, returning visitors, dog owners | Similar to Roma Norte with a slightly heavier weekend and holiday skew | Deco apartment blocks, older buildings with charm and plumbing history, some fine renovations | Comparable to Roma Norte, often slightly above |
| Narvarte | Medical staff and hospital-adjacent stays, value-focused monthly guests, families visiting patients | Flatter and less tourist-driven; weekday demand carries it | 1950s to 1970s residential blocks, larger floorplans, more two-bedrooms, more parking | Clearly below Roma and Condesa |
| Polanco | Corporate placements, executive relocations, luxury leisure, embassy and consultancy traffic | Weekday-corporate rhythm; quiet in Mexican holiday periods | Modern towers with amenities, doormen, and the highest cuotas in this list | Top of the list |
| Del Valle | Long-stay professionals, families, guests who want quiet and space over nightlife | Flat and residential; low tourist seasonality | Solid mid-century blocks and newer mid-rises, generous square meters | Between Narvarte and Condesa |
| Escandón | Budget-aware remote workers, guests priced out of Condesa, longer stays | Follows Condesa at a lag, thinner tourist demand | Mixed: old walk-ups next to recent small developments, uneven quality block to block | Below Condesa, near or slightly above Narvarte |
A few notes that do not fit in a table.
Roma Norte has the deepest and most competitive market. That cuts both ways: demand is reliable, and so is the supply of people competing for it. You will not win there on price, and you will not win on photos either, because everyone has good photos now. You win on the unit being genuinely good and on operating it well. It is also the neighborhood carrying the most public tension about short-term rentals, which is worth understanding before you buy rather than after; I wrote about that in the Roma Norte gentrification guide . The owner-facing page for the colonia is here .
Condesa reads similar to Roma Norte from a spreadsheet and feels different from inside a building. The stock skews older, and older here can mean beautiful proportions and a plumbing riser installed when the building went up. Get a plumber and an electrician into any Condesa unit before signing.
Narvarte is the one investors underestimate, usually because it does not photograph like a magazine. Its demand is structurally different: hospital corridors, weekday professionals, families on long medical stays. That demand is less seasonal and less exposed to tourism sentiment. Floorplans are bigger for the money, which suits the medium-term lane well. If that market is unfamiliar, compare it against Condesa for a monthly stay .
Polanco is a corporate market wearing luxury clothes. Entry price is the highest here and the cuotas are heavy, so the gap between gross and net is wider than the headline rate suggests. It works best for owners already positioned for corporate demand with billing and invoicing to match. The Roma Norte, Condesa and Polanco comparison covers the guest-side differences in more depth.
Del Valle is quiet, residential, and genuinely underrated for longer stays. Nobody flies to Mexico City to be in Del Valle. Plenty of people spend three months there and extend.
Escandón is the highest-variance option of the six. Two buildings on the same street can be entirely different investments. If you buy here, buy the specific building, not the colonia; the Escandón guide covers the block-by-block texture. If you have not settled on an area at all, the neighborhood overview for monthly stays is the wider map.
When the math does not work
I would rather talk someone out of a bad unit than manage it, so here is the honest list.
The purchase price already assumes a full calendar. If the seller’s or agent’s projection quietly uses 365 available nights, the price you are being asked to pay has the regulatory risk baked into it and pointed at you.
The building forbids it. Read the reglamento interno and the escritura. Some condominios in these colonias have restricted or banned short-term letting outright, and a friendly administrador is not a substitute for the document. This is the single fastest way to kill a deal, so check it first.
The cuota is structural, not incidental. Amenity-heavy buildings carry costs that do not shrink when your allowed nights do.
You need financing at Mexican rates and the spread is thin. Debt service is unforgiving when available nights are uncertain.
You plan to self-manage from another country. Guest messaging at 2am in a timezone you are asleep in, a cleaner who cancels, a boiler that fails on a Sunday. Either you are hiring someone or you are becoming someone, and management is a real cost line either way. What that looks like in practice is on the management services page, and past performance is documented under results .
The unit is wrong for the lane. A large three-bedroom in a corporate corridor, or a windowless studio in a family neighborhood, will underperform no matter how well it is run.
And the quiet one: you wanted a passive asset. This is a hospitality business attached to real estate. If passive income is the actual goal, a traditional long-term tenancy in the same apartment is a legitimate answer, and often the better one. The calculator linked above compares exactly those two paths.
How to sanity-check a specific unit before you buy
Work in this order. Steps one through three kill more deals than everything else combined, so do them before you fall in love.
- Read the building’s reglamento interno and the escritura for restrictions on short-term or furnished letting. Have a Mexican lawyer read them too.
- Get the cuota de mantenimiento in writing, with two years of history and every derrama.
- Confirm the unit can actually be registered: proof of property, RFC, predial current, and whoever will hold the folio.
- Study comparable calendars rather than comparable rates. A listing showing a high nightly price with an empty September is telling you more than the price is.
- Get a real furnishing quote for the whole unit, then divide it over three to five years and carry that as an annual cost.
- Run available nights at three levels: full year, roughly half, and the middle. If it only clears at the top, walk.
- Model the same unit as a 30-day furnished rental and as a traditional unfurnished tenancy. Compare all three.
- Take the numbers to a Mexican contador before you offer. ISR, IVA, the CDMX lodging tax, and platform retentions all behave differently depending on how you hold the property.
- Visit the block at 8am on a Tuesday and 11pm on a Saturday. Test the internet from inside the unit, on the floor the bedroom is on.
Where this leaves you
The Mexico City short-term rental market has not stopped working. It has stopped working automatically. The version of this business where you buy a central apartment, put it on a platform, and let the calendar fill itself belonged to a period with looser rules and lower entry prices, and that period ended.
What replaced it rewards two things: buying at a price that survives a conservative view of available nights, and running the unit properly once you own it. Both are boring. Both are also the entire difference between a unit that clears and a unit that becomes an expensive lesson about cuotas.
If you want to pressure-test a specific apartment against real operating costs rather than a listing screenshot, start with the calculator and then send me the unit . I will tell you if I think it does not work, which happens more often than the other answer.
