Turkey Launches EVREN, Its National AI Platform Built by the Defence Industry Presidency
Turkey’s Presidency of Defence Industries (SSB) has opened its national artificial intelligence platform EVREN to users. The platform was built to reinforce Turkey’s technological independence and data sovereignty, and it targets a broad audience ranging from defense industry companies and technology firms to academics, entrepreneurs and students.
EVREN brings data, AI models and high-performance computing power together in a single platform. All prompt, response and usage data is processed on high-performance GPU infrastructure located inside Turkey, keeping sensitive data from being transferred to foreign cloud services. Access to EVREN is provided at evren.ssyz.org.tr through e-Devlet (Turkey’s e-government) identity verification.
EVREN at a Glance
- Developed by: Presidency of Defence Industries (SSB), Turkey
- Access: evren.ssyz.org.tr (via e-Devlet identity verification)
- Models available: 11 open-weight large language models
- Active users: 7,500+
- Access model: Contribution-based credit system
- Unlimited API access: Until 1 November 2026
Key figures behind Turkey’s EVREN AI platform.
Background
Turkey’s public investment in artificial intelligence has accelerated in recent years. Under the National AI Action Plan, the government has pushed to develop domestic AI solutions for critical sectors. Defense, given its sensitive data-handling requirements, has emerged as a priority area, and EVREN is the first comprehensive public platform built specifically to meet that need.
EVREN’s compute workload runs entirely on GPU servers located inside Turkey. (Representative image: NOIRLab/NSF/AURA)
Details
Rather than charging users directly, EVREN runs on a contribution-based credit model. Users earn credit by uploading datasets, labeling data, or sharing trained models. Those credits can then be spent on the platform’s high-performance GPU pool for model training and large language model inference.
| Feature | Details |
|---|---|
| Computer vision support | End-to-end data labeling and model training for object detection, segmentation and classification |
| Inference layer | 11 open-weight large language models; API architecture with real-time response support and automatic model routing |
| API launch offer | Unlimited API calls until 1 November 2026, with no deduction from users’ credit balance |
| User base | Reached 7,500 active users within a short period |
| Roadmap | Domestic large language, vision and speech models; strengthening compute capacity through distributed GPU management |
What It Means for Turkey
EVREN marks a concrete step in Turkey’s strategy to reduce external dependency in artificial intelligence. Processing sensitive data on domestic GPU infrastructure, without routing it through foreign cloud services, gives defense companies, universities and technology firms a meaningful data-security advantage. The contribution-based credit model is designed to turn users into active participants in the ecosystem rather than passive consumers, aligning with Turkey’s goal of growing both the user and developer base of its AI sector. The platform is intended to become a shared, Turkey-centered development infrastructure for the defense industry, universities, technology companies, entrepreneurs and developers alike.
Frequently Asked Questions
What is EVREN?
EVREN is a national AI platform developed by Turkey’s Presidency of Defence Industries, combining data, AI models and high-performance computing in a single infrastructure.
Who can access EVREN?
A broad user base — defense industry companies, technology firms, academics, entrepreneurs and students — can access the platform at evren.ssyz.org.tr through e-Devlet identity verification.
Is the platform paid?
No. EVREN runs on a contribution-based credit model instead of direct paid access; users earn credit by uploading datasets, labeling data or sharing trained models.
Where is the data processed?
All prompt, response and usage data is processed on high-performance GPU infrastructure located inside Turkey, with no transfer to foreign cloud services.
How many language models are available?
The platform’s inference layer covers 11 open-weight large language models.
Until when is unlimited API access available?
API calls made until 1 November 2026 will be offered without any deduction from users’ credit balance.
Conclusion
With EVREN now live, Turkey has taken a concrete step toward building AI infrastructure centered on its defense industry. The platform is expected to expand with domestic large language, vision and speech models, while a distributed GPU management approach is planned to further strengthen compute capacity. These steps are expected to contribute to the goals set out under Turkey’s National AI Action Plan.
Related Articles
Sources
- Presidency of Defence Industries (SSB) press release, 24 September 2026
- Presidency of Defence Industries — official website
- EVREN AI platform
