Türkiye’s Artificial Intelligence Action Plan (2026-2030): A Four-Axis National Roadmap for the AI Ecosystem
Introduction: The Legal Nature and Framework of the Circular
By Presidential Circular No. 2026/9 (the “Circular”), published in the Official Gazette dated 18 August 2026 and numbered 33344, the “Türkiye Artificial Intelligence Action Plan (2026-2030)” (the “Action Plan”) was announced to the public. To be published on the official website of the Ministry of Industry and Technology (www.sanayi.gov.tr), the Action Plan requires all public institutions and organizations diligently to fulfil their respective duties and responsibilities and to provide all support and assistance needed during the implementation process.
By its legal nature, the Circular is not a regulatory act directly imposing obligations on natural or legal persons; it is an instruction and coordination instrument addressed to the administration. Nevertheless, the Action Plan to which the Circular refers sets out the four-year regulatory, investment and institutionalization agenda of the AI ecosystem as a roadmap that pairs responsible institutions, time horizons and performance indicators. In this respect, the Action Plan foreshadows the secondary regulation and incentive architecture that will directly affect the financial and real sectors, the start-up ecosystem and public procurement in the period ahead.
The Action Plan states that its preparation drew on the 2021-2025 National AI Strategy and the 2024-2025 Action Plan, the Report of the Grand National Assembly’s AI Research Commission, and the National AI Policy Document of the Presidency’s Science, Technology and Innovation Policies Board. The predecessor strategy had been put into force by Circular No. 2021/18, published in the Official Gazette dated 20 August 2021 and numbered 31574. The new document is built on the principles of human-centricity, reliability, ethical responsibility, digital sovereignty and sustainable development, and on the four core axes expressed as “Detect, Utilize, Produce and Govern”.
A participatory method is reported to have been followed in preparing the document: workshops were held with close to 500 sector professionals from public institutions, the private sector, universities and civil society organizations, and more than 2,200 action proposals were gathered from broad segments of society through a Citizen Opinion Survey. Following Circulars No. 2026/7 (the 2026-2030 MASAK Strategy Document) and No. 2026/8 (the National Green Finance Strategy and Action Plan), the Action Plan forms the latest link in a series of dated, indicator-based roadmaps issued within the framework of the “Türkiye Century” vision.
The Architecture of the Plan: Four Axes, Sixteen Actions, Five Layers and Three Components
The backbone of the Action Plan consists of four core axes that describe the AI maturity of society and institutions in four stages: Detect (building awareness, talent, the data ecosystem and the rules), Utilize (establishing infrastructure and driving adoption in the public and private sectors), Produce (financing, developing models and building physical-AI manufacturing capability) and Govern (attracting global investment, conducting AI diplomacy, assuming regional leadership and ensuring the security of models). Each axis contains four priority actions, so that a total of 16 priority actions are defined within a 4×4 framework.
The actions are built on an architecture that treats the AI ecosystem as mutually reinforcing layers. Five core layers Energy (low-carbon, scalable power), Compute Hardware (resilient GPU/chip capacity under a multi-supplier principle), Infrastructure (compute, storage, networking, data centres and AI factories), Models and Data (sectoral foundation models and export-ready packages) and Applications (solutions that create value in the field) form the backbone of the plan. Talent, Financing, and Regulation and Trust are the three cross-cutting components that feed all layers. Each action contributes to one or more points of this layered structure.
Implementation rests on four principles: speed and clarity (a small number of high-impact priorities, measurable targets, and a clear owner, budget frame and success indicator for each action); market orientation (the “scan-pilot-scale” cycle); trust and responsibility (a proportionate risk approach, transparency and independent assessment); and phased execution. Accordingly, the 2026-2027 period will prioritize establishing the core governance structures, piloting data spaces, the first public-sector applications, and the investment processes for compute capacity; the 2028-2030 period will emphasize scaling the pilots, commercializing sectoral foundation models, deepening international cooperation and realizing the investment targets.
Key Numerical Targets and the Economic Vision
The Action Plan sets out concrete, measurable and mutually reinforcing targets. On the talent side, it aims for at least 1.5 million citizens to complete the AI literacy programme within the first 12 months and 5 million in total by the end of 24 months, with workshops in all 81 provinces and at least 1,000 trainers trained in the first 6 months; it also targets the training of 10,000 advanced AI experts and 100,000 AI application professionals able to deploy solutions in their own sectors. Under the “Pioneers of AI” scholarship programme, at least 500 graduate/post-doctoral fellows are to be supported, reaching at least 1,000 active fellows in total; through the Global AI Talent and Researcher Attraction Programme and an accelerated “TechVisa”, at least 500 international experts are to be drawn to Türkiye.
On the infrastructure and economic side, the plan targets reaching 1 GW of installed AI data-centre capacity by 2030, building a Compute Portfolio supported by low-carbon energy and underpinned by a USD 10 billion investment plan, and through the “GPU for Everyone” programme opening 10 million GPU-hours of flexible access per year to researchers, start-ups and SMEs according to sectoral priorities. By mobilizing public resources and private-sector investment towards common goals, the plan envisages the creation of economic value exceeding TRY 1 trillion and the launch of regulatory sandboxes in at least five priority sectors.
The Actions, Axis by Axis
The Detect axis comprises four actions: (1) the National AI Literacy and Safe Use Programme free content and workshops delivered through existing channels (notably e-Government and the EBA (Education Information Network) education platform), public libraries and community centres, an assessment system integrated with e-Government, micro-certificates based on measurement and evaluation, and an annual “National AI Week”; (2) the National AI Talent Programme 10,000 advanced experts, 100,000 application professionals, the “Pioneers of AI” scholarships and international talent attraction; (3) the Data Ecosystem data spaces, data access/sharing and the “National AI Champions”; and (4) the Legal-Ethical Framework a proportionate risk approach, transparent use, intellectual property and the TÜRKPATENT AI Ethics Line.
The Utilize axis aims to turn momentum into actual use in the public and private sectors: (5) the Compute Portfolio 1 GW of capacity uniting the energy, hardware and infrastructure layers under one roof, low-carbon energy and a USD 10 billion investment plan; (6) GPU for Everyone 10 million GPU-hours of flexible, fair access per year; (7) AI Transformation in the Public Sector a roadmap for public institutions, AI criteria in public procurement, and embedding the “scan-pilot-scale” cycle in institutions; and (8) Sectoral AI Pilot Programmes pilots that turn ideas into field-validated products in priority areas such as health, energy and manufacturing, together with strengthening SME capacity.
The Produce axis seeks to bring together the investment environment, financing, model development and security infrastructure: (9) AI Growth Zones investment campuses with ready energy and infrastructure, centres of excellence and incubation centres; (10) the AI Financing Ladder an uninterrupted flow of resources from basic research to start-up scaling; (11) Foundation Model Development and Export turning sectoral data advantages into commercializable models and export-ready packages; and (12) Physical AI and Robotics converting the capacity of the defence industry, automotive and machinery manufacturing into a competitive advantage in autonomous systems, robotics, smart manufacturing and industrial automation.
The Govern axis aims to convert Türkiye’s geographic and institutional weight into concrete actions: (13) Attracting Global AI Investment a predictable incentive framework making Türkiye an attractive hub for international capital and hyperscaler cloud companies; (14) International Cooperation and AI Diplomacy common standards and bilateral agreements in platforms such as the OECD, G20, UN, the Organization of Turkic States and the Organization of Islamic Cooperation; (15) Regional AI Capacity a common large language model (LLM) for Turkic languages, joint model development and a regional investment bank; and (16) AI Safety and Test Environments reliability assessment, security testing and regulatory sandboxes.
Governance, Financing and Transparency Architecture
The plan’s governance rests on four functions: strategic steering, implementation tracking, ethical guidance and technical safety assessment. The apex decision and coordination body is the National AI Council, chaired by the President; the Council will approve national priorities and annual implementation plans, resolve inter-agency jurisdictional disputes, and issue an annual “National AI Policy Letter”. The Council’s secretariat and implementation tracking will be carried out by the Programme Office within the Ministry of Industry and Technology’s Directorate General of National Technology and Artificial Intelligence. Ethical guidance is entrusted to the National AI Ethics Council, while technical safety assessment is left to capacity coordinated by the Cybersecurity Directorate and strengthened in cooperation with the TÜBİTAK AI Institute.
To reduce institutional uncertainty, a responsibility-assignment matrix will be prepared for each action, identifying the responsible, accountable, consulted and informed institutions. On transparency, the National AI Progress and Transparency Portal will be the single platform that publicly tracks each action’s target–budget–time alignment, sub-tasks, realizations and status colour (green/yellow/red). Personal data and trade secrets will be excluded, and detailed internal management dashboards will be operated under role-based authorization.
At the heart of the financing architecture lies an integrated ladder of two funds. The first rung, the National AI Research Fund, will support basic and applied research and early-stage projects, providing each project with an automatic HPC/GPU credit package and priority access to the National Data Library. The second rung, the AI Growth Fund, will accelerate the Series A/B and scaling processes of maturing start-ups. The quantitative targets announced in the document are: at least TRY 10 billion in project support through the Research Fund, with at least 40% of supported projects advancing to the pilot/prototype stage; at least 20 start-ups receiving investment through the TRY 15 billion Growth Fund; and the mobilization of at least two units of private capital for every unit of public resources.
Key Points for the Private Sector, Investors and Stakeholders
Although the Action Plan imposes no direct obligations, it contains concrete opportunities and preparation items for the private sector and investors. The public sector becoming the first customer and reference point for domestic AI solutions, the introduction of AI criteria into public procurement, and the sectoral pilot programmes together create a new demand signal for suppliers. For start-ups and SMEs, “GPU for Everyone” access, the two-rung fund ladder, the AI Growth Zones and the TechVisa are instruments that ease access to financing, compute capacity and talent. For investors, the predictable incentive framework and the positioning aimed at hyperscaler cloud companies stand out.
On the compliance and risk-management side, early preparation matters. The provision of standards-based technical safety and performance assessment for high-impact AI systems before go-live and at major version updates; the wider use of Algorithmic Impact Assessments and model cards; the emphasis on the TÜRKPATENT AI Ethics Line and intellectual property; and the inclusion of personal-data-protection awareness in the literacy programme these are matters that may require companies that use and develop AI to review their internal governance, documentation and assessment capacity now. The regulatory sandboxes offer an early-participation opportunity to test and scale innovative solutions safely.
What Has Changed Compared with the 2021-2025 National AI Strategy?
The new document establishes a structurally and institutionally different framework from its predecessor. The most visible change concerns the coordinating authority: whereas the 2021-2025 Strategy was run along the axis of the Presidency’s Digital Transformation Office, the new Action Plan is positioned under the coordination of the Ministry of Industry and Technology (Directorate General of National Technology and Artificial Intelligence), integrated with the “National Technology Initiative” narrative. This points to an institutional preference that foregrounds a “producing AI” emphasis along the manufacturing, industry and defence-industry axis.
In terms of content, the new document consolidates scattered headings into a clean framework of four axes and 16 actions, constructing a journey from awareness to global leadership. The layered approach that unites the energy-hardware-infrastructure-model-application layers within a single architecture, a new production emphasis such as physical AI and robotics, the two-rung fund ladder, and a transparent monitoring portal with an independent evaluation component together reflect a more operational and measurable design than the previous period. The concepts of digital sovereignty and “sovereign AI capacity” are also placed more prominently at the centre of this document.
Assessment: A Strong Vision, Points to Watch
The Action Plan appears to be a significant institutionalization step in that, for the first time, it gathers Türkiye’s AI initiatives into a single four-axis, layered and indicator-based framework and provides for a transparent progress portal. That said, three points stand out. First, with coordination moving to the Ministry of Industry and Technology, how the boundaries of authority and coordination between the new structure and the Presidency’s Digital Transformation Office with its existing digital-transformation and data-governance mandate will be drawn will be decisive in practice.
Second, the point concerns the legal nature of the document. The Action Plan is a policy roadmap and does not put forward directly binding regulation. Nevertheless, the Plan is premised on a framework-level statutory foundation that clearly defines core principles, areas of competence and responsibilities and is flexible enough to leave room for innovation; it envisages that implementation details will be determined through secondary legislation within the remit of the relevant institutions. That foundation is supported by a graduated obligation structure based on a proportionate risk approach: simple checklists are envisaged for low-risk applications, while an expanded Algorithmic Impact Assessment and a publicly available Model Card obligation are envisaged for high-impact systems. In this respect, compared with a comprehensive, horizontal regulation such as the European Union AI Act, the Plan reflects a comparatively more flexible regulatory choice that prioritizes innovation. Indeed, in regulated sectors such as finance, health, energy, mobility and telecommunications, it envisages that existing practices concerning cloud use, data processing, outsourcing and cross-border data transfers which currently constrain the wider adoption of AI applications — will be reviewed together with the sectoral regulators, and that an action plan will be established for secure and auditable cloud use that is risk-based and provides predictability to investors. On the other hand, in the legislative work tied to the Plan, how and on what timetable the need for secondary regulation will be met — on matters such as the scope of the framework AI law, amendments concerning regulated sectors, and intellectual property should be monitored.
Third is the difference in scale between the targets and the announced public financing. Alongside ambitious targets such as 1 GW of compute power, USD 10 billion of investment and economic value exceeding TRY 1 trillion, the scales of the Research Fund (at least TRY 10 billion) and the Growth Fund (TRY 15 billion) are comparatively limited; success therefore depends largely on mobilizing private capital and on the “two units of private capital for every unit of public resources” leverage assumption materializing. The commitment to track part of the performance indicators at the outcome and impact level rather than at the activity level (numbers of participants/trainings) — is positive, and the transparency portal and independent evaluation component are strong elements that offset this risk.
Conclusion
Türkiye’s Artificial Intelligence Action Plan (2026-2030) is a roadmap that gives concrete form to the country’s AI vision and will largely shape the regulatory, investment and talent agenda of the next four years. For the private sector and stakeholders, the document should be read not as a legislative text creating new obligations directly, but as a map of priorities indicating public demand, the incentive architecture, financing instruments and the direction of the trustworthy-AI framework. Proactive preparation particularly on trustworthy AI, data governance, intellectual property and regulatory sandboxes can contribute both to managing compliance and reputational risks and to seizing the emerging opportunities. The fact that the first public progress report is to be released within 12 months of the Action Plan’s publication shows that this preparation should begin early.
This content is for general information purposes only and does not constitute legal advice. Professional advice should be sought for specific situations.