Articles · 2 November 2026
We think Dubai's villa oversupply is real on the schedule and unlikely on the ground
Developers have scheduled 16,400 to 35,400 villas and townhouses a year for 2026 to 2028. If they delivered that, villas would be oversupplied. They are delivering about 6,000 to 8,400 a year, and we expect 9,000 to 10,000 a year to 2028. That is below the 15,000 to 16,600 villa households Dubai's population growth adds each year.
By BUY DUBAI · 2 November 2026
Summary
A widely shared Reddit post argues that Dubai's villas and townhouses are the most overbuilt and most overpriced part of the market, and will fall hardest. We tested its seven claims against the Dubai Land Department sales record, the Land Department project register and the DXB Interact pipeline. The post is right on two counts: villas fell as hard as flats in 2014 to 2020, and they have risen much further since. Ready villa prices are up 145% from 2020 to 2025, against 83% for flats. It overstates the pipeline. On a full-stock base the villa ratio is about 35%, not 48%, and nothing on the record separates townhouses from villas. On delivery, the schedule would oversupply villas, but the rate at which villas are actually being finished does not. Our base case is 9,000 villas and townhouses in 2026, 9,800 in 2027 and 9,400 in 2028. Most of the villa stock dated 2028 to 2030 has barely started: projects due in 2028 average 3% built, and the big Damac, Emaar and Danube villa phases are under 1%. Villas would be oversupplied only if handovers rose to about two to three times today's pace (two to six times if measured against the announced schedule), or if new villa households fell by roughly half. Neither is happening yet.
Claim by claim
1. Total pipeline is 62% of stock. Agree. DXB Interact schedules 542,481 homes for 2026 to 2032, which is 61% of the post's 890,000 homes. We could not verify the 890,000 stock figure; it is the post's.
2. Apartments 79%, townhouses 65%, villas 48% of their own stock. Disagree. The post's footnote says its 642,000 per-type base leaves out older stock. Rescaled to 890,000, the ratios fall to about 57%, 47% and 35%. Both stock figures are the post's and unverified.
3. Villas and townhouses are largely sold. Not supported. In 2022 to 2025 the record shows about 35,000 off-plan villa sales against 309,000 for flats, and apartments are more than 86% of the pipeline (Cushman & Wakefield via Khaleej Times). Both look similarly pre-sold.
4. A 300% launch wave, the highest ever. Partly. Launches in 2023 to 2025 were 433,948 homes, 3.4 times 2014 to 2016 and 2.1 times 2014 to 2018. The source has nothing before 2014, so "ever" cannot be tested.
5. Villas fell as much as apartments in past downturns. Agree for 2014 to 2020: ready villas fell 37% per sq ft, flats 32%.
6. Villas rose far more in the last five years. Agree: +145% against +83% (ready, 2020 to 2025).
7. The most overpriced segment will correct most. Not yet seen. From the first to the third quarter of 2026, the ready villa median rose 2.0% while flats fell 2.9%.
How an overhang reaches prices
A pipeline does no harm until it is finished. Once it is, new homes need tenants, vacancy rises and landlords cut rents. If investors want a set yield, price follows rent. Villas carry a thinner cushion: their gross yield is about 4.6%, against 5.4% to 6.1% for flats. Two other channels matter. Buyers who cannot fund the final payment sell at handover, as in Melbourne in 2019. And when villas get dear, families move back to cheap flats.
The fair way to compare falls is give-back, meaning the share of the previous rise that is lost in the fall. From the 2009 trough to the 2014 peak, ready flats rose 88% and villas 122%. By 2020 both had given back about 75% of that rise. Villas fell further in percentage terms only because they had risen further. If the same 75% were given back on the 2020 to 2026 rise, flats would fall 36% and villas 46%. That is an illustration, not a forecast.
Prices fell while the population grew
From 2015 to 2019 Dubai's official population rose by 1.03 million, or 44%, yet ready villas fell 36% per sq ft and flats 27%. Completions of 9,000 to 30,000 homes a year were not obviously above new household formation. Three other things coincided: the oil price fell from late 2014, mortgage loan-to-value caps came in in 2013 and 2014, and much of the population growth was labour, which does not rent villas. Today growth is faster in absolute terms: +332,000 in 2025 (Digital Dubai) and +226,885 in the first nine months of 2026 (Gulf News). Note that 2024 onwards uses a new counting method, so it is not a clean series with earlier years.
What the Land Department record counts
We counted off-plan villa sales ourselves, from raw Land Department transactions. There were 9,235 in 2025 and 6,061 in 2026 to 14 September, against 121,596 and 68,293 for apartments. DXB Interact, which reads the same register, counts 16,254 villas for 2025. The two counts differ because the sources classify villas sold as plots differently, so treat the true number as a range. Villas were 7% to 12% of off-plan sales in 2025.
The record has limits, and we state them. The Land Department's units-by-type register returns no data, so nobody outside can count villa stock or villa completions from it. There are no unit IDs, so a resale cannot be separated from a first sale; Arabian Ranches III Anya shows 643 off-plan sales against 492 villas. Townhouses are filed as villas, so only projects named as townhouses can be split out. Our copy of the project register holds only projects with a 2026 start or end date.
That copy lists 3,260 villas due in 2026 and nearly all are well advanced: 1,579 are 75% to 99% built and 1,229 are 50% to 75%. Every villa dated 2028, 2029 or 2030 in the copy is under 25% built. Damac Islands 2 is the clearest case: its phases are 79% to 84% sold and 0.2% to 0.3% built, with dates in 2030.
How many villas will actually be delivered
Dubai rarely delivers its schedule. Fitch puts completions at 56% of schedule in 2022 to 2024 (Fitch via Economy Middle East), Knight Frank at 60% in 2021 to 2025 (Knight Frank), and Cavendish Maxwell at 41.3% in the first half of 2026 (Zawya). ValuStrat counts 7,375 villas completed in the first nine months of 2026 (Economy Middle East). We cannot check that figure against the record, because the record has no villa completion count.
For 2026 we add the fourth quarter from the register's most advanced projects to ValuStrat's figure. For 2027 and 2028 we take a 14% to 18% villa share of the schedule and assume 40%, 50% or 60% of it is delivered. The results:
2026: 8,400 low, 9,000 base, 10,500 high. 2027: 6,900 low, 9,800 base, 13,300 high. 2028: 6,600 low, 9,400 base, 12,700 high.
Every case is below the 15,000 to 16,600 villa households a year that today's population growth implies at 4 people a home and a 20% villa share. Both of those inputs are assumptions. At a 10% villa share, demand falls to 7,500 to 8,300 and the gap closes. The schedule says something different: 16,400 to 35,400 villas a year, which would meet or exceed demand. That is the honest reading. The schedule would oversupply villas. The rate of delivery does not, so far.
Townhouse premium and rents to 2030
On new Ejari leases, a 3-bed townhouse rents at the same city median as a 3-bed flat: AED 150,000 a year in 2025 and 2026 to date. Its premium over a 2-bed flat has narrowed from 77% in 2024 to 63%. Per sq ft the two rent for about the same, so families pay for the extra room, not for the garden. So far in 2026, like-for-like villa rents are down 3.2% and flat rents 5.6%.
We built a simple model that adds handovers and new households each year and lets rents move towards normal vacancy. By 2030 villa rents end 10% lower in the base case, 33% lower in the bear case and 6% higher in the bull case. Flat rents move almost as much, because families switch between the two. Population growth decides the result. The villa share of handovers barely changes it. Every input other than the rent data is an assumption, and these are scenarios, not forecasts.
Counter-examples
Other cities do not settle the question. In Dublin from 2007 to 2012 apartments fell 64% and houses 56% (CSO via Public Policy Archive). In Toronto's 2023 to 2024 condo glut, condos fell a little more than houses (TD). Austin's record apartment building cut rents first (Pew). Melbourne in 2017 to 2019 is the warning: during an apartment glut, houses fell 13% and apartments only 4% (SmartCompany). In Dubai, REIDIN found villas underperformed by 15% in the cycle after a villa building wave (REIDIN). The pattern is that supply sets rents, and overpricing against rents and incomes sets how far prices fall.
What would need to be true for villas to fall hardest
1. Villa handovers rise to about two to three times today's pace of 6,000 to 8,400 a year (two to six times if measured against the announced schedule). Not happening yet.
2. Or new villa households fall by 44% to 65%, through fewer families arriving. Not happening yet.
3. Villa rents on new leases turn down. They are down 3.2% like-for-like in 2026, less than flats.
4. Villas stay dear against flats. True: ready villas sold at 1.10 times the flat price per sq ft in 2025, against 0.82 in 2020 and 0.88 at the 2014 peak.
5. Handover payments force sales, or credit tightens. Not tested here.
If you own or are buying a villa, the measure to watch is villa handovers a year against new family households. The schedule is not that measure.
Method and assumptions
Prices are Land Department sales (Sell and Sell - Pre registration), 20 to 2,000 sqm, raw medians. "Villa" is property type Villa or sub-type Stacked Townhouses. Off-plan means a Pre registration procedure. The pipeline and deliveries come from DXB Interact, read on 2 November 2026. The villa share of handovers (14% to 18%) comes from published shares, not a register count. Assumptions: 4 people per home, a 20% villa share of new households, delivery ratios of 40% to 60%, and every model input other than rents. Unverified, and flagged where used: the 890,000 and 642,000 stock figures, which are the post's; household size; the split between investors and end-users, which no source publishes; and the 2008 to 2009 price falls, which come from press reports of Asteco and Landmark rather than the original reports, so we do not rely on them.
Sources: Dubai Land Department sales transactions and project register (copy of 5 October 2026); DXB Interact supply pipeline and market history (2 November 2026); Digital Dubai; Gulf News; Khaleej Times (C&W); Cavendish Maxwell; ValuStrat; Knight Frank; Fitch; Peak Trust; REIDIN; SmartCompany; TD; Pew; Public Policy Archive (CSO).