Phnom Penh Condo Market: Cross-Source Evidence

Phnom Penh market reports often place very different numbers side by side as if they were competing answers to one question. They are not. An official index tracks price movement in its statistical sample; a consultancy estimates completed stock and new completions; another consultancy series may track launch pricing; a portal or agency sees the buyers and renters moving through its own business; and listings show what individual sellers are asking at a particular moment. The problem is not that these readings differ. The problem is treating any one of them as a universal citywide condo price.

Coverage is especially important for the official series. NBC's June 2026 economic and monetary statistics bulletin carries the RPPI through May. The Phnom Penh sub-index was about 99.92 on a 2020=100 base, roughly 1.88% above April but about 1.93% below May 2025. That is useful evidence of movement within the official series, but it is not automatically a condominium index. The IMF's May 2024 technical assistance report said the published RPPI then covered houses because there were too few condominium and apartment observations in the lending data to estimate their price change reliably. NBC has continued to revise and publish the series in 2026, but the current statistical bulletin itself does not establish that a representative condo sample has since been added. On this page, the RPPI is therefore treated as broad residential context rather than a direct measure of Phnom Penh condo values.

Consultancy stock estimates answer a different question. Knight Frank counted 63,334 existing condominium units in Phnom Penh in H2 2025, while APS's Q1 2026 presentation shows more than 76,000 units across more than 150 projects. It would be a mistake to subtract one from the other and call the difference quarterly completions. The dates differ, but so can the monitored project universe, the point at which a unit enters completed supply and the segment definitions. For change over time, a continuous series from one provider is more defensible; the two headline totals are better read as separate views of the market.

Realestate.com.kh adds another layer. Its 2026 Condo Investment Guide says the company handled more than 2,000 condominium transactions over the preceding 24 months, explicitly combining sales and rentals, and it publishes buyer-preference data and a Phnom Penh price reference of about US$1,800 per square metre on a gross-area basis. That is useful commercial evidence from a large platform and agency, not a census of every Phnom Penh transaction. Live listings can move even closer to the unit a buyer is considering, but they show asking expectations, not necessarily the price achieved at completion. Each source becomes more useful once its role is kept narrow.

The same city can still mean a different sample. The useful comparison is the unit of observation and what the number can legitimately support.

The most striking gap is Knight Frank's US$676 per square metre versus roughly US$1,800 from Realestate.com.kh. Stripped of their labels, the numbers look like competing estimates of the same market. They are not. Knight Frank's figure is the average developer-advertised selling price of new launches in H2 2025 and uses net saleable area. Realestate.com.kh publishes a Phnom Penh reference from its own commercial dataset on a gross-area basis. The period, property mix, development stage, segment and area denominator differ, so averaging the two or choosing one as the city's 'correct' price would create a number with no clear meaning.

Area definitions matter even before market mix enters the picture. Gross floor area can include space that is outside a net saleable-area definition, and projects do not always use identical measurement conventions. A price per square metre cannot be converted responsibly without the underlying area figures for the same property. Even a perfect area conversion would not remove the larger difference: Knight Frank's measure is tied to a particular set of new launches during a period when affordable and mid-market product was prominent, while the Realestate.com.kh figure is drawn from a different pool of properties and activity.

The NBC RPPI belongs to yet another category. An index reading of 99.92 is not an absolute price per square metre and cannot be turned into a 'fair discount' on a listing. It expresses relative movement inside a statistical series against a base period. NBC's 2025 Financial Stability Review reported a 4.3% decline for Phnom Penh over 2025, and the 2026 monthly series gives a more current view of subsequent movement, but those figures only make sense within the index methodology. Because the IMF documented a house-only published RPPI in 2024, the index is especially unsuitable as a shortcut for saying every Phnom Penh condo moved by the same percentage.

A live resale listing answers a third pricing question: what the seller is asking today. It becomes more informative when several genuinely comparable units can be found in the same building or in close substitutes with similar size, floor, view, condition and ownership terms. Yet a carefully assembled set of listings remains asking-price evidence until an achieved transaction is known. The practical lesson from the US$676 and US$1,800 figures is therefore not to decide which source wins; it is to identify whether the buyer is trying to understand a new-project launch price, a current resale ask, an achieved transaction or the direction of an official index.

Within Knight Frank's own series, supply growth is straightforward: existing condominium stock rose from 57,772 units in H2 2024 to 63,334 in H2 2025, a 9.6% increase. Two projects, Vue Aston and La Vista One River View, added 1,770 completed units in the second half of 2025 alone. Those figures answer a supply question well: how much completed stock Knight Frank was tracking and how that series changed. They do not, by themselves, reveal how many units end-users bought, how much stock is vacant or how quickly a particular building can be rented.

APS provides a later but separate view. Its Q1 2026 presentation shows more than 76,000 units across more than 150 projects, with 61% of stock classified as mid-range, 26% affordable and 13% high-end. The same presentation records six launches totalling more than 3,300 units in the quarter, with only 26% of the new launches in the high-end category. That is useful evidence of how new supply was being positioned within APS's coverage. The gap between 63,334 and 76,000+ is also a reminder that a 'Phnom Penh stock' series cannot be assembled by jumping between consultancies: some of the difference may be genuine market change and some may reflect different monitoring universes.

Demand creates even more denominator problems. Knight Frank reports an average sales rate of roughly 3–4% for monitored available units toward the end of 2025, and its chart is explicitly framed around units available and monitored by quarter. That is not 3–4% of every condo in Phnom Penh and it is not a count of registered citywide transactions. The H2 2024 report contains a different measure: the launch sales rate in the first quarter after launch, which reached 29% then. Comparing 29% with 3–4% as if they were the same series would be misleading because the metric and denominator changed.

Realestate.com.kh observes demand through its own sales and rentals, buyer mix, unit preferences and location interest. Its 2026 guide says one-bedroom units account for 61% of Phnom Penh buyer preferences and Cambodian purchasers represent just over one-fifth of buyers; a September 2026 article repeats the latter figure against 18.2% in the previous 2025 analysis. This is useful evidence about the audience moving through a large portal and agency. It does not automatically become the buyer mix of the entire city without a disclosed citywide denominator. Supply can therefore be described relatively cleanly with stock and completions inside one series; demand always needs a second question: whose activity, which units, over what period and out of what available pool?

A defensible trend starts where the definition stays consistent. The clearest example here is Knight Frank's stock series: 57,772 existing condominium units in H2 2024 and 63,334 one year later. Knight Frank itself reports the increase as 9.6%. Provider, geography, property type and metric are aligned, so the change can reasonably be described as growth in Knight Frank's monitored existing stock. APS's 76,000+ figure is useful as a later cross-check from another researcher, but it is not the next point on the Knight Frank line.

The official RPPI should also be compared only with itself. NBC's Financial Stability Review 2025 reported that the residential property index fell 3.8% nationwide and 4.3% in Phnom Penh over 2025. The more recent series through May 2026 is no longer moving in one direction: Phnom Penh rose about 1.88% from April while remaining roughly 1.93% below its level a year earlier. Those time horizons can coexist; a monthly rebound does not erase a year-on-year decline. The same scope limit still applies, however. Until current methodology confirms representative condominium coverage, movement in the official series should not be relabelled as condo-price movement.

New-launch prices need another layer of caution even within one consultancy. Knight Frank maintains a half-yearly series of developer-advertised launch prices, but the mix of projects can change materially from one period to the next. A lower average can reflect more affordable launches entering the sample, not only price cuts on an unchanged product. The US$676 per square metre figure for H2 2025 is therefore a valid description of that launch cohort, but it does not prove that the same apartment type fell to that level year on year. The same logic applies to portal buyer shares: moving from 18.2% to just over 20% can describe a change inside Realestate.com.kh's analysis without becoming an official census of all buyers.

Series can also be revised. NBC publishes an updating statistical series, and methodological changes can alter historical observations; commercial researchers may expand their project coverage or adjust classifications. A current figure therefore needs a compatible previous observation, not merely an older number with a similar label. If the definition changed, the honest presentation is two releases with a note about the break rather than a precise growth rate. That is less dramatic than a single trend line, but it avoids creating momentum that comes from a changed measuring stick.

Consider a completed one-bedroom unit in BKK1. The 63,334-unit Knight Frank stock figure tells a buyer that the apartment sits within a large citywide condominium market, while Realestate.com.kh's location-interest data show that BKK1 is prominent among that platform's audience. Neither figure says what this particular unit should cost. Several one-bedroom comparables in the same building, or in genuinely similar buildings listed at roughly the same time, are much closer to the pricing question; verified sales and rents in the same project are closer still. Floor, view, fit-out, condition and the actual area definition can matter more to this unit than a Phnom Penh average.

A new affordable project on the edge of the city needs a different evidence set. Knight Frank's launch-price series and APS's breakdown of recent launches can help place the developer's pricing within the current new-build market. Even if a project is close to the US$676 per square metre H2 2025 launch average, that does not establish its future resale value or the price a completed unit will achieve several years later. The area basis, what is included in the quoted price, the segment used for comparison and the launch period all need to match before the benchmark becomes useful.

The official index sits further away from both unit-level questions. It can show that the wider residential price environment in Phnom Penh was weaker year on year, or that the series rebounded in a particular month. That context is useful when a seller's case rests on a broad statement such as 'the whole market is rising'. It does not determine the premium for a particular view, the quality of building management, the liquidity of a floor plan or the condition of a resale unit. The further a metric is from the property type and sample at hand, the less weight it should carry in a unit-level decision.

The most useful hierarchy is therefore based on proximity to the question, not prestige. NBC is strongest for official residential context; Knight Frank and APS for monitored stock, launches and market structure; Realestate.com.kh for the behaviour visible in its commercial dataset; and close comparables for the pricing of an individual apartment. When numbers disagree, period, property type, stage, area basis and denominator should be aligned before any conclusion is drawn. The more of those characteristics match the unit being considered, the more practical weight the evidence deserves.