Dubai has assembled many of the components of a next-generation property market: digital government infrastructure, global capital, major developers, PropTech hubs, artificial intelligence and regulated experiments in tokenised ownership. The more difficult task now is to make those components work as one market.
Dubai’s real-estate story is usually told through scale. That is understandable. Transaction volumes continue to rise, new districts are launched at extraordinary speed, international capital remains highly visible and property has become one of the clearest expressions of Dubai’s broader economic expansion. In the first quarter of 2026 alone, real-estate transactions reached AED252 billion, 31% higher than a year earlier.
Behind those numbers, however, something more structural has been taking place. Dubai has been building an institutional and technological architecture around property that few other markets have attempted at comparable speed. Dubai Land Department has digitised significant parts of registration, valuation and market administration. DIFC and DLD have established the Dubai PropTech Hub. Technology communities in free zones, accelerators, financial institutions, major developers and an expanding population of startups are working on almost every part of the property lifecycle.
DIFC’s recent PropTech 2033 research mapped more than 800 PropTech business models globally and, more importantly, described the industry’s direction as a movement from isolated digital tools towards AI-native, system-level infrastructure. That distinction is important because Dubai is beginning to possess many of the components of such an infrastructure without yet having fully connected them.
What has become increasingly interesting to me, both through my earlier work around housing finance and now through building an AI investment-intelligence company, is not what Dubai lacks technologically. The city is hardly short of technology. The more important gap lies between the technologies.
Dubai has become exceptionally good at digitising property. The next opportunity is to make real-estate capital more intelligent.
A sophisticated market whose reasoning remains fragmented
There is already substantial intelligence inside Dubai’s property ecosystem, but it is distributed among businesses solving different parts of the problem.
Property Finder uses machine learning and market data to support valuation and forward-looking pricing. Bayut has developed automated valuation and market-analysis capabilities. Mortgage and home-finance platforms are digitising a process that traditionally required significant human intermediation. Keyper works on property-management infrastructure. SmartCrowd and other platforms have expanded fractional access to property. PRYPCO Mint, in cooperation with Dubai Land Department and other government partners, has taken another part of the market into regulated tokenised ownership. The Dubai PropTech Hub itself brings developers, technology companies, investors and regulators into the same innovation environment.
Realla.AI, the company I am building, starts from another point in the chain. The product today applies AI underwriting to an individual property, combining market information, financial analysis, risks and scenarios into a decision-ready investment report. In the immediate commercial context, that can help an agent or investment professional analyse a property more consistently and present the investment case more clearly to a client.
But while building the product, my view of the opportunity changed. The report is useful, but the more valuable asset is the structured reasoning generated behind it: assumptions, comparisons, risks, decisions and, eventually, outcomes. Once accumulated, that information can become institutional memory. The system begins to remember not only what an investor bought, but what was rejected, why it was rejected and whether the original reasoning proved correct. From there, underwriting can evolve into portfolio intelligence and ultimately into an intelligence layer between market information and capital allocation.
This is where the distinction between data and intelligence becomes important. A market may know the last transaction price, current asking price, rental history, developer, construction status and payment schedule of an asset and still leave the investor with the hardest part of the work: deciding whether that asset deserves capital relative to every other opportunity available at the same moment.
Today, much of that reasoning still has to be reconstructed from portals, developer materials, DLD records, mortgage offers, spreadsheets, valuation reports and individual experience. The information infrastructure is becoming highly sophisticated. The reasoning infrastructure remains fragmented.
Primary sales and secondary liquidity are not the same market
That distinction matters particularly in Dubai because the headline phrase “Dubai real estate market” conceals different mechanisms of price formation.
The primary and off-plan market is exceptionally efficient. Developers can define a product, launch it internationally, mobilise brokerage networks, combine pricing with multi-year payment plans and sell substantial inventory long before completion. Off-plan transactions have recently represented roughly 70% or more of Dubai residential activity.
This is not inherently a weakness. The ability to fund and distribute new construction is one of the mechanisms that has enabled Dubai to expand at extraordinary speed.
But the second transaction works differently.
Once the original launch, marketing campaign and payment-plan advantage recede, an owner wanting to sell must find another investor or end-user willing to accept the economics of the asset as they exist at that moment. The property competes not only with other ready properties, but with the next generation of new launches supported by fresh developer incentives.
This is why primary-market demand and secondary-market liquidity should not be treated as the same thing.
The difference is especially important for newer developments. A property can be successfully sold once without yet having developed a deep market for the second and third transaction. Pricing gaps between new and ready properties vary sharply by community and product, so reducing the issue to a single discount percentage would be misleading. The structural point is more important: developer-led primary price formation and open secondary-market price discovery are different processes.
As Dubai matures from a development-led market into a deeper investment market, the quality of the second transaction becomes increasingly important. Investors eventually need to know not only whether an asset can be bought, but how quickly it can be sold, to whom and at what price.
Tokenisation has proved access. Liquidity is the harder problem.
The first phase of Dubai’s real-estate tokenisation initiative demonstrated genuine demand. The first property offered through PRYPCO Mint attracted 224 investors from 44 nationalities, with 70% entering Dubai real estate for the first time. A subsequent tokenised property was fully funded in one minute and 58 seconds.
Those results matter because they show how dramatically technology can lower the traditional entry threshold to real estate.
The more consequential stage began when DLD enabled resale through a secondary marketplace. At that point, the experiment moved beyond digital ownership and began testing whether a new ownership structure could support a functioning market after the original sale.
This is where tokenisation is often oversimplified.
The ability to settle ownership quickly is not the same as liquidity. A digital marketplace can operate continuously, and settlement can take minutes once a buyer and seller are matched, but none of that guarantees that a willing buyer will exist at the moment an investor wants to exit.
Dubai already has significant virtual-asset infrastructure around this experiment. VARA’s public register contains dozens of licensed service providers across exchange, brokerage, custody, investment, lending and related activities. That is an important foundation, but those licences should not be confused with dozens of liquid secondary markets for tokenised real estate. They are infrastructure. The economic problem of continuous demand remains different.
In my view, this is where AI and tokenisation begin to become much more interesting together than separately.
Tokenisation can make ownership divisible and digitally transferable. AI can make the underlying asset understandable at machine speed.
An investor choosing between a full apartment, a fractional interest and several competing developments still needs to understand valuation, rent economics, future supply, developer execution, financing, downside scenarios and expected liquidity. If that reasoning can be generated continuously from reliable market information, the asset begins to acquire something it has not historically possessed: a persistent layer of investment intelligence.
The technical ability to transact then begins to meet the economic ability to decide.
The next market will have to become machine-readable
For that transition to develop beyond isolated investor applications, developers themselves will need to become active participants.
A large part of new-build real estate is still distributed in formats designed primarily for human sales: brochures, renders, launch presentations, websites, PDFs and conversations with brokers. These are effective for people. They are much less useful to an AI system acting on behalf of an investor.
Such a system needs structured access to live inventory, actual pricing, payment schedules, construction progress, escrow information, expected handover, service charges, incentives, financing, resale conditions and the developer’s delivery history.
In practical terms, projects will have to become machine-readable.
The precedent already exists elsewhere. E-commerce became algorithmically efficient when products stopped existing only as web pages and became structured objects that search engines, marketplaces and software could compare directly. Real estate is likely to undergo a similar transition.
A developer in an AI-native market may eventually present each project to two audiences simultaneously: the human investor and the investor’s machine.
That changes the economics of competition. Marketing will remain important, but the quality and transparency of structured investment information will increasingly influence whether an asset survives comparison against hundreds of alternatives.
Once enough of the market becomes structured, the investor can also behave differently. Instead of manually searching for a property, an investor could define a mandate: a given amount of capital, target return, acceptable construction risk, diversification constraints, preferred locations and expected holding period.
- a given amount of capital
- target return
- acceptable construction risk
- diversification constraints
- preferred locations and expected holding period
Software could then compare primary and secondary inventory, evaluate developers, model financing, analyse tokenised and conventional ownership structures and rank alternatives according to the investor’s own parameters. Human responsibility would remain essential, particularly for large capital allocations, but much of the work between investment intent and a defensible decision could become continuous rather than rebuilt for every deal.
That is the point at which AI stops being primarily a sales or productivity technology and begins to participate in capital allocation.
A fragmented industry will not integrate itself
There is, however, a structural obstacle.
Real estate remains one of the world’s most fragmented industries. Developers, brokers, banks, portals, property managers, registries, technology companies and investors operate under different commercial incentives and on different systems. Each has good reasons to optimise the part of the value chain it controls. Common standards therefore tend to emerge slowly, even when everyone broadly agrees about the direction of travel.
This is where Dubai’s institutional model becomes particularly relevant.
The Dubai Real Estate Sector Strategy 2033 already sets an ambitious direction for the market, including higher transaction volumes, a substantial increase in overall market value, stronger international investment, greater use of AI, improved transparency and more integrated investor experiences. DLD’s REES initiative and the Dubai PropTech Hub are already expressions of the same philosophy.
The next useful intervention may therefore be less about creating another innovation programme and more about establishing a common operating environment.
That does not mean government should build a single super-platform or replace private companies. The more valuable role is to define the conditions under which private systems can interoperate: trusted data standards, machine-readable project records, authorised APIs, identity and permissions frameworks, standards around AI-generated investment information, and connections between ownership, financing, tokenisation and secondary-market infrastructure.
Private companies would continue to compete above those standards. A portal would not need to become a tokenisation platform. A developer would not need to become a mortgage company. Realla would not need to become a property marketplace.
The purpose of a common architecture would be precisely the opposite: to allow each participant to specialise while making the information created by one participant usable by others.
This, in my view, may eventually be more important than adding another hundred PropTech startups to the ecosystem.
Beyond PropTech
The more deeply I work on this problem, the less convinced I am that the next category should simply be described as PropTech.
PropTech remains essential, but it covers an enormous range of activities: construction, brokerage, marketing, mortgages, leasing, management, building operations and ownership.
The layer beginning to emerge around investment has a more specific economic function. It converts property information into comparable reasoning about value, risk, liquidity and capital allocation, and ultimately connects that reasoning with execution.
I think of this capability as Real Estate Capital Intelligence.
It is not a single product, and no single company is likely to own it. It is more plausibly an ecosystem: trusted public infrastructure from DLD, financial and innovation connectivity through DIFC, structured inventory from developers, large reservoirs of market data from portals, financing infrastructure from banks and mortgage platforms, new ownership mechanisms from tokenisation businesses, and specialised intelligence companies working on underwriting, liquidity and capital decisions.
- trusted public infrastructure from DLD
- financial and innovation connectivity through DIFC
- structured inventory from developers
- large reservoirs of market data from portals
- financing infrastructure from banks and mortgage platforms
- new ownership mechanisms from tokenisation businesses
- specialised intelligence companies working on underwriting, liquidity and capital decisions
Dubai is unusually well positioned to assemble such a system because many cities possess only part of this combination. Some have large property markets. Others have deep capital pools or sophisticated technology ecosystems. Dubai has these elements together with a government that has repeatedly demonstrated a willingness to redesign market infrastructure when the existing model no longer matches the city’s ambitions.
If those pieces become interoperable, Dubai could move beyond being one of the world’s most digitised real-estate markets. It could become one of the first in which property, financing, ownership, liquidity and investment intelligence increasingly operate inside the same architecture.
That would be a more consequential transformation than using AI to generate better property descriptions, qualify more leads or accelerate individual transactions.
The real opportunity is larger: to give a market that already moves extraordinary amounts of money a more intelligent mechanism for understanding where that money should move next.
