San Antonio

US$275k – 573k

Asking range (no closed-sale data exists)

MXN 32k – 55k

Asking price per m²

San Antonio is flat, it is close to the historic centre, and it is the cheapest of the colonias that are both. The published asking band is MXN 4.8 million to MXN 10 million, which is roughly $275,000 to $573,000 at the ECB reference rate of 17.4601 MXN per USD on 23 July 2026. That is the whole of the reason most foreign buyers look here rather than in Centro Historico or Guadiana.

It is also an asking band scraped from listing portals by a content aggregator, not a record of what anything sold for. This page is written around that limitation rather than over it.

What San Antonio costs, and what that number is

At 17.4601, the San Antonio per-square-metre band converts to roughly $1,830 to $3,150. That conversion is our arithmetic on the published peso figures, not a separately sourced number.

Three things are wrong with using this table the way buyers usually use it.

First, these are asking prices scraped from Propiedades.com, Properstar, Vivanuncios and Inmuebles24. In a market carrying 29.6 months of inventory, asking prices sit systematically above clearing prices. No sample size is disclosed for any cell.

Second, the bands overlap so heavily that the colonia name carries little price information. The top of San Antonio’s per-square-metre band, MXN 55,000, is above the bottom of Centro Historico’s, MXN 45,000, and identical to the top of Atascadero’s. Its whole-property top, MXN 10 million, sits inside Centro’s range. A buyer who assumes San Antonio means cheap will meet listings that are not.

Third, the same publisher gives two irreconcilable numbers for this colonia. A page published 2 April 2026 states San Antonio “has around MXN 39,600/m2”. A page published ten weeks later gives MXN 32,000 to 55,000. A point estimate and a range are not the same claim, and neither page states a method that would let a reader reconcile them. To the April page’s credit, it says outright that it avoids fake precision where official neighbourhood-level data does not exist. That admission is the most reliable sentence on it.

There is a fourth calibration problem, and it is the one most likely to cost a buyer time. The aggregator’s citywide whole-property average, MXN 6.8 million or about $390,000, sits far below the closed-sale average of $568,829 that Realty San Miguel reported for 2025. These two datasets are measuring different markets: the portals capture the peso-denominated domestic segment, and the brokerage MLS captures the dollar-denominated foreign-buyer segment. They should never be blended on the same line. A buyer who arrives quoting the San Antonio portal band to a foreign-buyer-facing agent will find the conversation starts somewhere else.

Land, separately, is quoted at MXN 11,000 per square metre in San Antonio against a city average of MXN 10,650, with Centro Historico at 18,000, Guadiana at 16,000, Ojo de Agua at 13,500 and Los Frailes at 9,500. San Antonio land is priced at almost exactly the municipal average. That is the same aggregator, with the same untraceable methodology, and its own Centro land figure of MXN 18,000 sits oddly below its Centro built-property figure of MXN 45,000 to 90,000.

Supply: one of the few liquid submarkets outside Centro

This is the closest thing to genuine neighbourhood-level supply data that exists in public, and it is one brokerage’s listing book rather than a market census. It mixes houses, condos and land. The nineteen colonias in that directory sum to 595 by our arithmetic, against 860 active listings citywide in May 2026, so either the directory is incomplete or the two counts are drawn on different bases. We could not resolve which.

The shape still matters. San Antonio’s 48 listings mean a buyer can assemble something resembling a comparable set. Guadiana at 11 and Ojo de Agua at 13 cannot support that: in a submarket of a dozen listings, a single sale is an anecdote, not a comparable.

The citywide backdrop is a supply build, not a demand collapse. Active listings rose from 593 in May 2025 to 860 in May 2026, an increase of 45 percent. Average months of inventory over the first five months of 2026 climbed to 29.6, from 17.6 in the same period of 2025, against the 15 to 18 months the same brokerage calls normal for this market. The closing count for the same window is unresolved: the source reported 197 through May, then 186 through June, and a year-to-date count cannot fall, so we treat the 2026 volume as unpublished. Even on the source’s own May reading, roughly flat sales against 45 percent more listings is a market where sellers, not buyers, changed behaviour.

Every one of those figures traces to a single brokerage’s own MLS view. San Miguel de Allende has no public transaction registry, AMPI’s local chapter publishes no downloadable statistics, and the local MLS is not publicly accessible. There is no second reading to check the first against.

Appreciation: the number in circulation, and the number that is measured

One aggregator publishes 10 to 13 percent twelve-month appreciation for San Antonio, against 9 to 12 percent for Guadalupe and Centro Historico and 8 to 10 percent citywide. It is the highest neighbourhood figure on its own table, and it is the number that circulates when San Antonio is described as an up-and-coming area.

It carries no measurement method, no sample size, and no repeat-sales or hedonic basis. It is contradicted by the only measured series available: the citywide average resale price fell from $650,462 in 2024 to $568,829 in 2025, and inventory nearly doubled in months terms by mid-2026. The publisher lists Realty San Miguel among its own sources, so it had access to the contradicting data.

The 12.5 percent fall should not be read as negative appreciation either. Sales above $2 million fell from 19 in 2024 to 12 in 2025, and a change in mix of that size moves a mean on its own. It is a mix effect at least as much as a price effect.

The defensible position, for a publisher with no commission at stake, is that measured neighbourhood-level appreciation for San Miguel de Allende does not exist. Not for San Antonio, not for Centro, not for anywhere. Any percentage you are shown for a specific colonia was modelled or asserted.

Who buys here

The strongest evidence on where foreigners actually live is a peer-reviewed study of INEGI census tracts, not brokerage copy. The seven census tracts inside the UNESCO polygon held 1,803 foreign residents, 42 percent of the municipality’s entire foreign-born population of 4,312, and foreigners made up 8 percent of those tracts’ 22,405 residents. The single most concentrated tract inside the polygon was 15 percent foreign. One peripheral south-east tract reached 29 percent.

No source maps census tracts to colonia boundaries. So we cannot tell you what share of San Antonio is foreign-born, and neither can anyone else. What the study establishes is that foreign residency in San Miguel is concentrated but not total, and that the popular figures are unreliable: the census counts 4,312 foreign-born residents against local press claims of “more than 15,000” and a widely repeated estimate of 20,000 to 25,000. Those claims are unsourced and internally inconsistent with each other. Nobody has published a defensible current count.

The qualitative picture is agent copy, and we mark it as such. Three brokerages independently describe San Antonio as flat, mixing local families with long-term and international residents, with restaurant growth along Sterling Dickinson. They are mutually consistent, which is weak corroboration and nothing more. All three sell property.

The walk, and the part of it that is real

The load-bearing fact is terrain, and terrain is the one thing the brokerage descriptions agree on and can be checked against a map: San Antonio, Guadalupe and Guadiana are flat; Atascadero, Ojo de Agua and Balcones are on the hillside. Slope drives construction cost and, for anyone over sixty carrying groceries, actual walkability.

The walking time is not real data. One brokerage describes San Antonio and Guadalupe as a “flat few minutes walk to the center” on an undated page with an obvious incentive to compress the distance. No source gives a measured distance in metres or a verified walking time from any colonia to the Jardin. The unreliability is demonstrable elsewhere in the same body of material: Guadiana is stated as a 15-minute walk by one brokerage and a 10-minute walk by another, and Ojo de Agua’s “fifteen-minute downhill walk” quietly omits that the return leg is uphill.

Treat every stated walking time in San Miguel as a sales claim until you have walked it yourself, in the afternoon, carrying something.

Street to street

The brief a buyer usually arrives with is that San Antonio is cheaper than Centro but further out. The more accurate version is that San Antonio is a wide band with a low floor.

Within a single colonia the published asking range runs from MXN 32,000 to MXN 55,000 per square metre, a spread of roughly 72 percent between floor and ceiling, on 48 listings that mix houses, condos and raw land. There is no colonia-level days-on-market data to tell you which end of that band is transacting. The only DOM figure in circulation, approximately 150 days citywide, traces to an aggregator citing a brokerage update that contains no days-on-market figure at all. We checked the cited source. It is not there.

So the practical consequence is unglamorous: in San Antonio the neighbourhood average tells you very little and the specific street tells you most of it. That is true everywhere, but it is more true here than in Centro, where 220 listings and a tightly bounded UNESCO core give you something to average against.

What it costs to hold

Predial on urban built property runs on marginal bands from 0.234 percent, on fiscal value up to MXN 500,000, rising to 0.354 percent above MXN 1.5 million. Suburban built property is a flat 0.207 percent. These apply to the cadastral value, not market value, and Mexican cadastral values typically sit far below market, so the real carrying cost is very low by US or Canadian standards.

We cannot convert that into a San Antonio number. The municipality’s tabla de valores unitarios de suelo y construccion, which sets cadastral values by zone, could not be located in any public record. The 2026 Ley de Ingresos sets the rates and references the value tables without reproducing them. Until that document surfaces, per-colonia carrying cost is not knowable from published sources.

Acquisition tax is the largest closing line and it is not the flat 2 percent widely repeated. The 2026 municipal tariff is progressive, reaching a fixed MXN 39,125 plus 4 percent on value above MXN 1.5 million. At MXN 5 million that is MXN 179,125, an effective 3.58 percent; at MXN 10 million, the top of the published San Antonio band, it is MXN 379,125, an effective 3.79 percent.

The risks that apply here specifically

Water is the under-discussed one. CONAGUA’s 2024 determination for the Cuenca Alta del Rio Laja aquifer, which underlies San Miguel, records a deficit of 61,984,190 cubic metres a year and states plainly that there is no volume available for new concessions. Well-drilling vedas covering the Allende municipality date to 1958 and 1964. For a buyer of an existing serviced house on municipal supply this is a long-run question; for anyone buying land, or a property on a private well, it is an immediate due-diligence item, and water quality in outlying wells is rarely mentioned in brokerage material.

Security is the other. San Miguel de Allende’s profile differs sharply from the industrial south of Guanajuato: three homicides in the municipality in June 2026 against 45 in Leon, 32 in Salamanca and 25 in Irapuato. It is not, however, insulated from the state. Over the twelve months to August 2025 the municipality recorded 88 homicides, a rate of 49.3 per 100,000, making it Mexico’s 50th most violent municipality. In February 2026, cartel reprisal violence reached roughly 23 Guanajuato municipalities, including San Miguel, with an arson incident in the municipality on 22 February and a body found on 25 February. Any page telling you San Miguel is unaffected by Guanajuato’s violence is describing a place that does not exist.

What would change this page

Four documents would replace most of the guesswork above with measurement: the municipal tabla de valores unitarios, which would give real relative land value by zone; an AMPI or MLS release broken out by colonia; aggregated deed values from the Registro Publico de la Propiedad; and AirROI’s per-neighbourhood short-term rental metrics, which are purchasable rather than genuinely unavailable. We are pursuing the first by records request.

Until then, San Antonio has one honest headline. It is the lowest entry point among the flat, close-in colonias, it is one of the few non-Centro submarkets with enough listings to compare within, and every price figure attached to it is an asking price of unstated provenance. Buy the street, not the colonia.

All USD conversions at the ECB reference rate of 17.4601 MXN per USD, 23 July 2026. Annual average exchange rates are a different measure and are not used here.