Friday, 11 September 2026
Türkiye’s new artificial intelligence programme contains an unusually large real-estate proposition hidden inside what initially appears to be a technology strategy. The country wants to establish at least 1 GW of computing and data-centre capacity, attract a minimum of USD 10 billion of predominantly private investment and create at least five dedicated AI development zones by the end of 2028. For the commercial property industry, those targets matter because artificial intelligence ultimately requires physical infrastructure. Large computing facilities need land, enormous quantities of electricity, resilient fibre connections, cooling infrastructure and access to locations where development can proceed quickly. Türkiye’s AI ambitions could therefore create a new competition between cities and regions for one of the world’s fastest-growing categories of infrastructure investment.
The proposed AI development zones are at the centre of that opportunity. Rather than expecting investors to assemble land, electricity and telecommunications infrastructure independently, the government wants to create locations capable of accommodating large data-centre and AI investments with much of the essential infrastructure already addressed. Faster planning and approval processes are also intended to reduce the time between an investment decision and construction. The concept could substantially change how international data-centre investors assess Türkiye. Conventional data-centre development frequently becomes a search for electricity before it becomes a search for property. A site can have excellent transport links, available land and attractive development costs, but without sufficient grid capacity it has little value to a large computing operator.
Türkiye’s 1 GW objective consequently makes electricity one of the most important property questions arising from the strategy. Providing that amount of capacity will require considerably more than constructing server buildings. Grid connections, substations, transmission infrastructure and potentially new generation capacity will all influence where projects can realistically be developed. This means the eventual locations of the country’s planned AI zones could be more important to property investors than the headline national investment target. Areas with available power, strong telecommunications networks and sufficiently large development sites could gain an advantage, while locations already struggling with electricity constraints may find it difficult to accommodate energy-intensive computing projects.
Istanbul will inevitably be part of the discussion because of its position as Türkiye’s principal corporate, financial and technology centre. But the scale of the programme raises the possibility that significant computing infrastructure will need to spread beyond the country’s largest business market. Data centres do not necessarily need to occupy the most expensive commercial locations. Large AI facilities can potentially operate in secondary cities or industrial areas if they have sufficient electricity, reliable fibre connectivity and appropriate physical security. That could give Türkiye an opportunity to create new technology investment locations rather than concentrating the entire industry in Istanbul.
The government is also seeking international capital. Türkiye wants global cloud and technology companies to consider the country as a regional computing location, supported by investment incentives and greater certainty surrounding the development and operation of large technology infrastructure. The USD 10 billion investment ambition provides an indication of the scale being considered, although it should not be interpreted as committed capital. The target depends largely on private investors deciding that Türkiye can provide competitive conditions for AI infrastructure and technology businesses. For property investors, that distinction is important. Government policy can make sites available and improve infrastructure, but private capital will ultimately determine how quickly the proposed development programme becomes physical real estate.
The potential opportunity also extends beyond hyperscale data centres. Türkiye plans regional technology centres where researchers, smaller businesses and technology companies could gain access to advanced computing resources and development facilities. Such locations could create demand for laboratories, research premises, offices and specialised technical buildings alongside the main computing infrastructure. Türkiye’s existing industrial base makes another part of the strategy particularly relevant to commercial property. The government wants artificial intelligence development to extend into robotics, autonomous technology, defence and advanced manufacturing. These industries require very different buildings from conventional software companies.
A company developing autonomous machines may need engineering space, laboratories, workshops and testing facilities. Robotics businesses can require combinations of office and light-industrial accommodation. Advanced manufacturers may need production halls with substantial power requirements and specialist technical infrastructure. AI investment could therefore begin influencing several property sectors simultaneously. Data centres would represent the most capital-intensive component, but laboratories, technology campuses, advanced factories, engineering facilities and conventional offices could form part of the wider ecosystem.
This is where the proposed development zones could have their greatest long-term effect. A successful computing campus rarely operates entirely in isolation. Contractors, equipment suppliers, engineering companies, telecommunications providers and technology businesses can develop around major infrastructure investments. If Türkiye succeeds in attracting international cloud providers and large AI investors, some of the resulting property demand could consequently appear outside the boundaries of the original data-centre sites.
The programme could also create a new type of competition for industrial land. Logistics developers traditionally assess sites according to motorway access, population, labour availability and proximity to major consumption centres. AI infrastructure introduces another hierarchy in which electricity and fibre capacity can outweigh many of those conventional considerations. A large site beside a motorway is valuable to a logistics operator, while for a data-centre developer a less obvious location with abundant electricity and multiple fibre routes may be considerably more attractive.
That could affect land values around suitable infrastructure nodes if the government’s programme begins attracting investment at scale. It could also place AI projects in competition with manufacturing, logistics and other electricity-intensive industries for grid capacity and development land. Environmental performance will become part of that equation. Türkiye’s programme calls for expansion of computing capacity while paying attention to energy efficiency and lower-carbon infrastructure. As AI computing becomes increasingly electricity intensive, the availability and source of power will become closely connected with the investment credentials of individual locations.
The timetable is ambitious. The initial period through 2027 is intended to establish the institutional structure, prepare computing investments and begin practical AI projects. From 2028 onwards, the programme is expected to move increasingly towards expansion, commercial deployment and international investment. At least five AI development zones are targeted by the end of 2028. Their locations, infrastructure specifications and development models will therefore be important indicators for the Turkish property market over the next two years.
There are still significant uncertainties. A national target for 1 GW of capacity does not mean that 1 GW of data centres has already been financed or commissioned. Likewise, the USD 10 billion figure represents an investment objective rather than signed transactions. The ultimate property impact will depend on whether private investors commit capital and whether suitable power and infrastructure can be delivered quickly enough. But Türkiye has provided something that data-centre investors increasingly require: a clear indication that computing infrastructure is being treated as part of national economic development rather than simply another category of commercial construction.
For the property industry, the next stage will be geographical. The key question is no longer simply whether Türkiye wants a larger artificial intelligence industry. It is where the electricity, land and fibre needed to support that industry will come from. Once the locations of the AI zones become clearer, so will the potential winners. Cities capable of combining large development sites with reliable power, telecommunications infrastructure, skilled workers and efficient approvals could attract an entirely new category of international investment.
Türkiye’s AI strategy may have been written as a technology programme, but achieving its biggest targets will require something considerably more physical: land, electricity and buildings. That could make the country’s artificial intelligence ambitions one of its most important emerging real-estate development stories.
Source: CMS
