The Presidency of Defence Industries in Türkiye has officially introduced EVREN, a national artificial intelligence platform designed to fortify the country's military technology sector. This new digital ecosystem integrates high-performance computing power with specialized data labeling services to support a growing community of users.
Securing sensitive data within Türkiye's borders
The launch of the EVREN platform represents a strategic push toward technological sovereignty, a trend increasingly seen among nations looking to decouple their defense intelligence from foreign-controlled cloud services. According to a statement from the Presidency of Defence Industries, the platform processes all prompt,response, and usage data on high-performance GPU infrastructure located physically within Türkiye.
By maintaining this localized infrastructure, Türkiye aims to prevent critical defense information from being transferred to international cloud providers. This approach addresses the growing global concern regarding data sovereignty, ensuring that the sensitive datasets used for military AI training remain under national jurisdiction.
A credit-based ecosystem for 7,500 active users
EVREN has already attracted 7,500 active users, including a diverse mix of defense firms, tech companies, academics, and students. Rather than functioning as a traditional one-way service, the platform utilizes a unique credit-based model to encourage community participation. As reported by the source, users can earn credits by uploading datasets, labeling data, or sharing their own trained models.
Users on the EVREN platform can then reinvest these earned credits to fund model training or large language model inference operations via the platform's GPU pool. This circular economy is designed to transform users from mere consumers into active contributors, helping to build a richer, more robust ecosystem of domestic data and models.
Unlimited API access through November 2026
To lower the barrier to entry for developers, the EVREN platform has introduced an aggressive open-access period. Specifically, API calls made before the November 1, 2026 deadline will incur no costs and face no usage limits. This window allows for rapid experimentation and integration without the immediate pressure of operational costs.
Technical capabilities within the EVREN architecture include support for computer vision tasks like object detection and segmentation. Furthermore, an inference layer featuring 11 open-weight large language models has already been activated, allowing existing systems to connect seamlessly via the platform's API architecture and automatic model routing.
The challenge of developing domestic LLMs and audio models
Several critical questions remain regarding the long-term scalability and depth of the EVREN ecosystem. While the Presidency of Defence Industries has stated that the platform will eventually incorporate domestically developed large language, image, and audio models, the specific timeline and capabilities of these upcoming models remain unverified. It is also unclear which specific 11 open-weight models are currently included in the inference layer.
Additionally, while the platform utilizes a distributed GPU management approach to scale resources, the source does not specify the total capacity of this infrastructure or how it will compete with the massive, centralized computing clusters maintained by global tech giants.. The ultimate success of EVREN will depend on whether the domestic models can achieve the sophistication required for high-stakes defense applications.
Comments 0