By Bosphorus News Türkiye Desk
Türkiye's 2026-2030 Artificial Intelligence Action Plan sets out a large expansion of computing infrastructure, financing and public-sector use, aiming to mobilize at least $10 billion in private-sector-led investment and raise artificial intelligence and data center installed power to at least 1 gigawatt by the end of the decade.
The government says the 16-action plan could generate more than 1 trillion lira in economic value by combining public resources with private investment. Its first phase, covering 2026 and 2027, concentrates on computing capacity, data access, regulation and early public-sector deployment, with larger infrastructure projects and commercial scale-up scheduled for 2028-2030.
Türkiye Targets 1 GW of AI and Data Center Power
By 2030, Türkiye wants at least 1 GW of installed AI and data center power, 1 exabyte of storage and 100 terabits per second of national backbone capacity, alongside at least 10 terawatt-hours a year of low-carbon electricity purchases tied to the expansion.
Renewable and nuclear power are included in the low-carbon category, and long-term power contracts for the new capacity are expected to carry a minimum 60 percent low-carbon electricity requirement. Energy supply will also influence where new AI Growth Zones are built, with grid connection capacity, projected demand and access to renewable and nuclear generation feeding into site selection.
At least $10 billion in investment is expected to come largely from the private sector, supported through data center and high-technology incentive programs.
A national Computing Portfolio would combine public reserve capacity, domestic data centers and international cloud services, with public institutions encouraged to use shared capacity instead of building separate computing infrastructure. The "GPU for Everyone" program would draw on the same pool to serve startups, small and medium-sized enterprises, researchers and public projects.
The initial target is at least 2 million GPU-hours a year, with at least 40 percent reserved for technology startups, rising to at least 20 million GPU-hours annually by the end of 2028. The document uses several GPU-hour benchmarks: its executive summary refers to 10 million a year, while the detailed "GPU for Everyone" action starts at 2 million and sets the higher 2028 target.
Two Funds Worth 25 Billion Lira
Financing will run through two funds, starting with a National Artificial Intelligence Research Fund designed to provide at least 10 billion lira for basic and applied research. Supported projects would receive GPU credits and priority access to national datasets alongside conventional funding.
A second mechanism, the 15 billion lira AI Growth Fund, is aimed at companies entering Series A and Series B financing or other scale-up stages, with the government seeking to support at least 20 AI companies and mobilize at least two units of private capital for every unit of public money committed.
Physical investment will be concentrated in AI Growth Zones, planned as sites where power, land, telecommunications and permitting are prepared in advance for data centers and AI facilities. The first zone is scheduled to begin operating by the end of 2027; by the end of 2028, the government wants at least five zones declared and three operational, alongside three Regional Centers of Excellence and six local access points.
Public Procurement Becomes Part of the Strategy
The plan puts government purchasing power directly into the AI market, targeting at least 2 percent of public investment program budgets for AI projects. Ministries would identify potential uses, test them through pilots and move successful systems into wider deployment within six to 12 months.
The first year calls for at least 30 pilots across 10 ministries and pilot-to-scale programs covering at least 10 public services, with the e-Government platform among the first areas marked for deployment. Planned uses include personalized service guidance, preliminary application checks and intelligent document assistance.
Companies whose products perform successfully inside government could receive a digital badge, allowing those deployments to be used as export references for Turkish technology companies.
AI Training Extends Beyond Engineers
The workforce targets extend well beyond specialist training, with the government aiming to have 5 million people complete an AI literacy program within two years through online programs and workshops across all 81 provinces.
At the professional level, the plan targets 10,000 advanced AI specialists and 100,000 people able to apply AI in fields including engineering, health, law, finance, agriculture and public administration. Basic AI courses are expected to be introduced at least 50 universities by the end of 2027.
The document also addresses jobs likely to be altered or displaced by automation. Türkiye plans to map occupations exposed to AI substitution and put at least 1 million workers through occupational transformation assessments during the plan period, while a separate recruitment scheme seeks to bring at least 500 international AI researchers, entrepreneurs or instructors into Türkiye through measures including TechVisa, research infrastructure access and settlement support.
Türkiye Wants Models It Can Export
Model development is built around sectors where Türkiye expects domestic data and industrial experience to support commercial products. The plan calls for mission-oriented research, sectoral foundation models and export-ready packages, including work in health, agriculture, energy, disaster management and advanced industrial applications.
Physical AI has a separate track covering autonomous systems, industrial and agricultural robotics, health applications, disaster response and humanoid robots. The document also proposes moving experience developed in Türkiye's defense and manufacturing sectors into civilian autonomous and robotic systems.
Türkiye Eyes Regional Data Centers, Turkic-Language Model
Türkiye proposes joint data center investments and shared computing models with members of the Organization of Turkic States, Gulf countries and other nearby partners, alongside cooperation on model development, training and data sharing.
The regional program includes a shared large language model for Turkic languages, with the first version targeted for production by the end of 2027. Cooperation with other regional groupings would extend to sectoral applications, common datasets and capacity-building programs.
The plan also envisages work through the Organisation for Economic Co-operation and Development, Group of 20, United Nations and regional organizations on AI standards, investment and governance.
Domestic oversight would sit with a National Artificial Intelligence Board chaired by the president, supported by an ethics board, technical security assessments and a public transparency portal. High-impact systems used in health, education, employment, social assistance, lending, biometric identification, law enforcement and justice would face algorithmic impact assessments and model-card requirements.
Regulatory sandboxes are planned in at least five priority sectors, while the wider governance program includes security testing and controlled environments for new AI systems before they are scaled or brought to market.
Annual implementation programs will still have to specify financing needs, ministry budgets, private-sector contributions and detailed timetables. The first deadlines come before the broader 2028 expansion, with the computing strategy, GPU access program, data-sharing framework, governance bodies and initial public deployments assigned to the opening phase of the plan.
Sources: Türkiye's Ministry of Industry and Technology, Türkiye Artificial Intelligence Action Plan 2026-2030, Bosphorus News review and reporting.
Read original document in Turkish here.