Q2 2026: 38,082 office leasing deals in Dubai, up 4% quarter on quarter. It reads like a healthy, gently expanding market. It isn't — and the headline is the problem, because it averages together two things that are moving in opposite directions.
Read one level down and the picture inverts. New leases: +16%. Renewals: −16%. The two cancel almost exactly, which is how you get a flat-to-slightly-up top line. Total demand isn't growing. It's rotating. Tenants aren't staying put and re-signing where they are — they're leaving, and signing somewhere new.
That distinction is the entire story. A market where demand is flat but everyone renews is stable and boring. A market where demand is flat but a sixth of tenants refuse to renew and a sixth of leases are brand new is a market in the middle of a physical reshuffle — capital and occupiers relocating from one kind of building into another. Treating "the office market" as a single number hides that reshuffle completely, and it is why benchmarking the wrong way is the most expensive analytical error an investor can make in the UAE right now.
The rent nobody quotes
Ask what a Dubai office costs and you'll be quoted the market average: 238 AED/sqft/year. That number just did something it hadn't done since 2021 — it stopped rising. The first quarter without rental growth in five years. For anyone whose plan depends on rents ticking up every quarter, that is the signal to re-read the plan.
But the average is a blend, and the blend hides a chasm. Prime DIFC space runs to ~750 AED/sqft/year — a 3.2× spread between the middle of the market and the top, inside a single city. A gap that wide is not explained by location; DIFC and the average both sit in the same twenty-kilometre core. What separates them is product: specification, floorplate efficiency, and the age of the building. New, well-specified space commands a multiple of the blended average not because of its postcode but because of what it physically is.
The practical takeaway: the average is the wrong number to underwrite against if you are buying new. It describes the stock tenants are leaving, not the stock they are moving toward.
Demand is small-floorplate
Now look at what tenants are actually signing. 66% of all leasing deals are units under 500 sqft — and that band grew 17% quarter on quarter, faster than the market. Add the next two brackets and 80% of every deal sits under 1,000 sqft. The typical Dubai office transaction in 2026 is small: a compact, efficient floor for a compact, efficient team.
Here is the mismatch that drives everything else. Almost none of the legacy stock was designed for that demand. The older towers were built around large, full-floor corporate tenancies — the leasing model of a previous decade. A building carved for 5,000 sqft occupiers does not subdivide gracefully into 400 sqft suites: the cores, the lift waiting times, the shared services and the fit-out economics all fight it. So the small tenant who wants modern, efficient, right-sized space in an old tower is quietly miserable — and the −16% renewal line is that tenant voting the only way a tenant can, by not signing again.
Supply stays tight — then opens
If demand is rotating into new space, the next question is how much new space is coming, and when. The pipeline answers in two acts. Near term it is thin: 1.9 mn sqft completes in 2026, 1.7 mn in 2027. Then it opens — 2.4 in 2028, 3.2 in 2029, and 4.2 mn sqft by 2030, more than double the 2026 figure.
The shape of that curve is the whole opportunity, and its whole risk. For the next two years, well-specified new space meets demand that is already leaving older buildings, against a supply line that stays tight — which is precisely the condition under which pricing power sits with the landlord and new-build rents hold their premium. An asset entering today delivers into that gap. The other edge of the curve: by the back half of the decade, supply roughly doubles, and the buyer underwriting a 2029–2030 exit has to ask whether the premium that looks obvious in 2026 still holds when the pipeline opens. The timing of delivery relative to that curve is not a detail — it is the investment.
Abu Dhabi says it louder
If Dubai whispers the product-over-geography thesis, Abu Dhabi shouts it. Al Maryah Island — the ADGM financial free zone — went from 2,300 AED/m²/year in 2024 to 6,000 today, a 2.6× jump in roughly two years. The CBD sits directly next door, a bridge away, and rents there stayed flat at 1,930. Al Reem, also central, sits at 1,600. Same city, same access, wildly different outcomes.
The gap is not geography — it is the ADGM address and the Grade A product that comes with it. Two of every three office enquiries in the emirate go to ADGM. That concentration is why Al Maryah can nearly triple while its neighbours stand still. And the pull is self-reinforcing: 170,000 m² of new Grade A lands on the island by around 2027, chasing the demand that is already pointed there. Same emirate, opposite outcomes — decided by product and address, not by the map.
Two markets, two strategies
Put the four readings together — flat average rents, small-floorplate demand, a tight-then-loose supply curve, and a product-driven price gap in both emirates — and one conclusion falls out. There are two office markets sharing one city, and they reward opposite playbooks.
- Buy below DLD median, refurbish, re-let or flip
- Rent is the return — and the average is flat at 238, first time since 2021
- Renewals −16%: your tenant has options, and they're using them
- Capex isn't upside — it's the entry ticket
- New leases +16%: demand is migrating into new spec
- Small floorplates — 66% of deals — built for today's tenant
- Price to prime, not to average: 750 vs 238
- Supply stays tight until 2028; entry today, delivery into the gap
Secondary is an income play. You buy below the DLD median, refurbish, and re-let or exit. Your return is rent — and rent has stopped rising on the average, while your tenant has newer options and the renewal data says they are using them. Capex here is not upside; it is the price of admission, the money you must spend just to make the asset lettable at all. Done well it works. Underwritten as a growth story, it disappoints.
Primary off-plan is a capital growth play. You are not buying today's rent — you are buying delivery into a supply gap, the small floorplates that two-thirds of the market actually wants, priced to prime rather than to average. The risk is the mirror image of secondary's: it lives or dies on the delivery date landing inside the tight window and on the exit clearing before the pipeline opens.
Benchmarking an off-plan asset against secondary comparables. They are separate cycles with separate buyers. A secondary office competes on price. A new one competes on product. Mixing the two is how investors talk themselves into the wrong asset with the right spreadsheet — the most expensive mistake in this market. The spreadsheet looks disciplined precisely because it is built on the wrong comparable set.
What this means before you sign
Three things follow directly from the data. First, decide which market you are in before you value anything — an income asset and a growth asset are underwritten with different comparables, different return sources, and different risks. Second, if you are buying new, do not benchmark against the 238 average; benchmark against prime and against the specific micro-corridor of comparable new stock. Third, put the delivery date against the supply curve: entering into the tight window is the thesis, delivering after it opens is the risk.
Every one of those decisions turns on the same question — is the price being asked the price this market actually pays? The portal shows the asking price. DLD shows the paid price. The gap between them, by district and by type, is the difference between a disciplined entry and an expensive one.