The writer is the Digital Team Lead at TRT World.

We often view the artificial intelligence race as a contest between two countries: the US and China. The US is home to the world’s leading AI companies, advanced chip-design capabilities and deep capital markets. China, meanwhile, represents the other major pole of the race, with its vast domestic market, manufacturing capacity and state-backed technology investment.

Just behind these two superpowers, a quieter yet increasingly important competition is taking shape. Countries such as Türkiye, France, Germany, Spain, Saudi Arabia and Qatar cannot easily compete directly with the US or China in a frontier-model race that requires hundreds of billions of dollars in investment. But in the age of AI, power is not determined solely by who can build the largest model. Energy, data centers, computing capacity, data, human capital and the ability to deploy AI across the real economy are also becoming strategic assets. Türkiye’s recently launched national AI platform, EVREN, should be viewed within this broader transformation.

It would be misleading to describe EVREN simply as “Türkiye’s ChatGPT.”Developed under Türkiye’s Secretariat of Defense Industries (SSB), the platform brings together different stages of AI development within a shared infrastructure – from data preparation and labelling to model training, computer vision applications and the deployment of large language models.

The platform currently provides access to 11 open-weight large language models. Users can earn credits by uploading datasets, labelling data or sharing trained models, and then use those credits to access high-performance GPU infrastructure for model training and inference.

At the heart of this infrastructure are 64 Nvidia H200 GPUs, each equipped with 141 GB of HBM3E memory. Together, the system provides roughly 9 TB of high-bandwidth GPU memory and 47.5 petaflops of FP16 Tensor Core computing capacity. The fact that advanced AI chips such as the H200 are subject to US licensing and export controls for certain countries illustrates how this hardware has become not only a technological resource, but also a geopolitical one.

The critical concept here is computing power – or “compute.”EVREN has already reached approximately 15,000 active users. The next phase envisages integrating domestically developed language, vision and audio models while expanding distributed GPU infrastructure.

This is particularly important for universities, researchers and technology startups. One of the biggest barriers to developing advanced AI systems today is no longer simply access to knowledge or talent, but access to sufficient computing power. EVREN is therefore more than a software project. It is an attempt to build a shared AI development infrastructure within Türkiye.

Perhaps an even more important dimension of EVREN is data sovereignty. Today, millions of people and companies using foreign AI services send data to infrastructure located in other countries. With EVREN, user prompts, responses and usage data are intended to be processed on high-performance GPU infrastructure located in Türkiye. This distinction becomes particularly important for defense, government and other critical sectors.

In the coming years, another question will increasingly accompany the question of “Where is our data stored?”: Where is the AI computation taking place, and who controls the infrastructure performing it? AI sovereignty is increasingly being shaped by three fundamental components: data, models and compute.

EVREN should also be viewed within the context of Türkiye’s new AI action plan. Announced publicly in June and entering into force on Aug. 18, 2026, the 2026-2030 AI Action Plan sets out the country’s roadmap under four pillars: “Recognise, Benefit, Produce and Govern.”The targets are ambitious. By 2030, Türkiye aims to reach at least 1 GW of AI and data-center capacity, mobilize at least $10 billion in investment, predominantly from the private sector, and develop a workforce of 10,000 advanced AI specialists and 100,000 AI application professionals. The plan also aims to provide AI literacy training to 5 million people within two years.

Other initiatives include a “GPU for Everyone” program for researchers, startups and SMEs, the creation of a National Data Library, sector-specific AI models and the development of physical AI and robotics capabilities. Taken together, these initiatives reveal where EVREN fits into the bigger picture: Türkiye wants to become not merely an economy that consumes AI, but one capable of managing its own data, computing infrastructure and certain AI models.

Another dimension of the AI boom is often overlooked. Artificial intelligence is no longer just a software story. It is increasingly an energy and infrastructure story. Large-scale data centers require not only GPUs, but also reliable electricity, powerful grids, cooling systems and high-capacity fiber connectivity. Türkiye’s ability to achieve its 1 GW data-center target will depend heavily on whether it can provide this infrastructure.

Türkiye has spent years developing its role as an energy corridor between Europe and Asia through projects such as TANAP. Adding to that planned electricity transmission projects, expanding renewable capacity and nuclear power expected to come online through Akkuyu, and a more interesting question emerges: Could Türkiye evolve from an energy corridor into an energy-and-compute corridor between Europe and Asia? That is not yet a reality. But if the 1 GW data-center target is achieved, Türkiye could potentially translate part of its geographic advantage into digital infrastructure.

Türkiye does not need to replicate the general-purpose AI model race being pursued by companies such as OpenAI, Google DeepMind and Anthropic, where the required investment can run into hundreds of billions of dollars. A more realistic strategy may be to specialize in areas where Türkiye already has potential competitive advantages: defense and autonomous systems, industrial AI and robotics, energy, logistics, agricultural technology, and AI models for Turkish and other regional languages.

The same logic applies to other middle powers. Rather than every country attempting to build its own miniature Silicon Valley, the future AI ecosystem may become more distributed – with Gulf capital and energy, Türkiye’s engineering and manufacturing capacity, and Europe’s research and industrial base complementing one another. Türkiye’s location at the intersection of these regions could itself become an advantage.

EVREN’s computing infrastructure uses Nvidia H200 GPUs, while a significant share of the large language models currently available on the platform are global open-weight models. But technological sovereignty does not necessarily mean producing every chip, model and piece of software within national borders. The more important question is how much strategic room for maneuver a country retains in critical technologies. Can you keep your data within your own jurisdiction? Can you run models on infrastructure you control? Can your researchers access sufficient computing power? Can you develop models for your own language and industries? These are more useful questions through which to assess EVREN’s significance.

Building a platform and creating a globally competitive AI ecosystem are, of course, very different things. Türkiye still faces significant challenges, including access to advanced chips, financing, energy and data-center infrastructure, frontier AI research capacity and retaining highly skilled talent.

EVREN’s success should therefore ultimately be measured not by the number of users it attracts, but by the results it produces. How many domestic models are developed? How many startups use EVREN’s computing capacity to build products? How much research is commercialized? How much AI technology developed in defense is transferred into the civilian economy? And, most importantly, can Türkiye evolve from a country that merely uses AI into one that also exports AI technologies and products?

Instead of beating the US or China in the AI race, the objective is to become a significant AI production, computing and deployment hub at the intersection of Europe, the Gulf, the Balkans, the Caucasus and Central Asia.

EVREN is not that destination in itself. But if it is scaled effectively, widely used by the private sector and universities, and supported by investments in energy, data centers and human capital, it could become an important part of the infrastructure needed to get there. Because in the age of artificial intelligence, sovereignty will no longer be measured solely by control over territory, energy or physical infrastructure. The ability to control data, computing power and AI infrastructure will increasingly shape the economic and strategic power of nations. That is where EVREN’s real significance lies.

*Opinions expressed in this article are the author's own and do not necessarily reflect the editorial policy of Anadolu. news_share_descriptionsubscription_contact contact_the_ombudsman is_there_an_error_in_this_story