First, the thing everyone gets wrong about these two numbers
The 24,537 homes delivered in the first half of 2026 did not come from projects registered in 2026. They came from projects registered around 2020 to 2022, which have spent four to six years being built. The register we are reading covers projects registered this year, and those are the 2028 to 2030 delivery wave. So a low construction percentage across the 2026 cohort is not evidence that anything has gone wrong — a project registered in March is supposed to be at zero in August. Anyone telling you otherwise is either confused or selling something. What the register is genuinely useful for is the shape of the wave rather than its health today: how many units are queued, when their developers have told the government they will finish, and how much building has happened against those dates. That is a forward-looking picture nobody else publishes, and it is the only honest way to read a record half-year. A record is a measurement of the past. The register is the closest thing to a measurement of the future.
Homes delivered in H1 2026 were registered around 2020–2022. The 2026 register is the 2028–2030 wave. Do not confuse the two.
The scale of what is queued
364 projects are registered with the Dubai Land Department for 2026, covering 87,405 units with a declared value of AED 105.1 billion. Set that against the 24,537 homes delivered in the whole of the first half of this year and the proportion becomes clear: this single registration year holds roughly three and a half times the units that Dubai's best half-year since 2008 managed to hand over. Those units do not all arrive at once, and that is the point of looking at the filed dates. Completion years break down as 37 projects in 2027, 138 in 2028, 114 in 2029, 61 in 2030, 11 in 2031 and 3 in 2032. The distribution has a clear peak, and it is 2028 — more than a third of everything registered this year is contractually filed to finish in that one year, with 2029 close behind. Combine the 2028 and 2029 cohorts and you have 252 of 364 projects, roughly 69% of the register, arriving inside a two-year window. If the filed dates hold, the second half of this decade will see delivery volumes that make the current record look modest. Whether they hold is the whole question.
252 of 364 registered projects — 69% — are filed to complete across 2028 and 2029.
How much of it has actually started
This is where the register earns its keep, because construction percentage is filed alongside the completion date and almost nobody reads the two together. Of the 364 projects, 263 were logged at exactly 0% construction on 21 August 2026, and 328 were under 1%. That is 90.1% of the cohort. Exactly one project on the entire 2026 register was more than half built. As established above, that is normal for a cohort registered this year and it is not a scandal. It does, however, put a number on what 2028 requires. 138 projects are filed to complete in 2028. The vast majority of the register they belong to has not meaningfully begun. For those dates to hold, those developments need to go from a standing start to finished handover in roughly twenty-eight months — and Dubai's own delivery record suggests that four to six years is the realistic span for a project of any size, which is why this year's completions were registered around 2020 to 2022. We are not predicting failure. Individual towers do get built in two to three years, particularly smaller ones, and several of these are small. But when a third of a registration year clusters on a single completion date and that cohort is collectively at under 1% construction, the sensible expectation is that the 2028 peak will spread into 2029 and 2030 rather than land as filed.
263 of 364 projects sit at exactly 0%. One project on the whole register is more than half built.
The slowdown in launches is real, and visible in the filings
The coverage this week notes that new launches are slowing while deliveries rise. The register supports that, and it is worth showing the shape rather than asserting it. Grouping the 364 projects by their filed construction start date gives 56 in January, 43 in February, 62 in March, 42 in April, 35 in May, 29 in June, 31 in July and 20 in August, then 17 in September, 11 in October, 5 in November and 13 in December. The first quarter carries 161 projects; the third carries 68. One caveat before you read too much into the tail: these are filed start dates, not registration dates, and projects registering later in the year will continue to add to the final months. The late-2026 figures will therefore fill in somewhat. But the front half of the curve is complete and already told: registrations with a first-quarter start outnumber those with a third-quarter start by well over two to one. Taken with the record deliveries, this is the classic shape of a market moving from expansion into absorption — a lot of supply arriving, less new supply being started behind it. For a buyer that is a mixed picture rather than a bad one. More completed stock competing for tenants puts pressure on rents in the districts receiving it. Fewer new launches eventually tightens the pipeline again, which supports prices for anything already built.
161 projects filed a Q1 start. 68 filed a Q3 start. The slowdown is in the filings, not just the commentary.
What this actually means if you are buying
Three practical conclusions, none of which require you to take a view on where the market is going. First, treat any handover date on a project registered this year as a projection rather than a commitment, and treat a 2028 date with particular care given how crowded that year is and how little construction stands behind it. Get the contractual completion date into your sale agreement, find out what the penalty for delay actually is, and model your own finances against a later date than the brochure quotes. Second, if you are buying for rental income, look at what is scheduled to complete in your specific district rather than across Dubai. A record year of deliveries is irrelevant to you if none of them are in your area, and severe if several thousand are. Third, and most usefully, use the construction percentage as a filter when comparing two projects with the same handover date. A development at 15% built with a 2029 date is telling you something quite different from one at 0% with the same date, and that comparison is free, public and takes about a minute to check. It is the single most underused number in Dubai off-plan buying.
Two projects, same handover date, different construction percentages. That comparison is free and almost nobody makes it.