Pull to refresh
Logo
Türkiye launches defense-focused national AI platform EVREN

Türkiye launches defense-focused national AI platform EVREN

New Capabilities

National AI platform keeps training data on domestically hosted GPUs

Today: Türkiye launches national AI platform EVREN

Overview

Updated 2 hours ago

Türkiye's defense procurement agency opened a national AI platform on September 27 that keeps sensitive data inside the country. EVREN, built by the Presidency of Defense Industries, gives Turkish firms, academics, and students a single place to label data, train models, and run inference on domestically hosted GPUs.

The platform is a bet on data sovereignty. Every prompt, response, and dataset is processed on high-performance NVIDIA H200 hardware inside Türkiye, avoiding the servers of foreign cloud providers. It drew 7,500 active users within days of opening.

Why it matters

Turkish defense and research teams can now train and run AI models without sending sensitive data to foreign cloud providers — a capability the country lacked.

Questions about this story

Free account needed to ask — your question is kept and asked for you right after sign-up. Answers are public.

No questions yet — be the first to ask.

Key Indicators

7,500+
Active users at launch
Reached within days of opening, with one report putting the figure at 8,000 the same day.
11
Open-weight large language models
Available through an inference layer with automatic model routing and real-time response support.
64
NVIDIA H200 GPUs
High-performance compute hosted in Türkiye, managed through a distributed GPU approach.

Voices

Curated perspectives — historical figures and your fellow readers.

Ever wondered what historical figures would say about today's headlines?

Sign up to generate historical perspectives on this story.

People Involved

Organizations Involved

Timeline

2 events Latest: Today
  1. Türkiye launches national AI platform EVREN

    Today Launch

    SSB opens EVREN with GPU infrastructure, 11 open-weight LLMs, and 7,500 active users.

  2. EVREN passes 8,000 active users

    Today Milestone

    Active users cross 8,000 within hours of launch, per local reports.

Scenarios

1

EVREN reaches 25,000 users and adds a domestic Turkish LLM

Likely Resolves by Q2 2027

Discussed by: SSB statements and Daily Sabah coverage of platform expansion plans

SSB has said the platform will be developed with domestically built large language, image, and audio models. If EVREN keeps its early momentum and the free API period drives adoption, user counts and homegrown model availability are the visible proof. SSB's own announcements would confirm both.

2

Named Turkish defense firms shift AI workloads to EVREN

Possible Resolves by Q1 2027

Discussed by: SSB's stated goal of serving defense industry firms; TRT World coverage

EVREN's stated purpose is to give defense firms a sovereign place to build AI. A concrete sign of success would be a specific company announcing it trained or deployed a model on the platform, moving work off foreign cloud services. Such an announcement would signal real adoption rather than sign-ups.

3

EVREN adoption stalls once free API access ends

Unlikely Resolves by Q1 2027

Discussed by: The contribution-based credit model's friction; November 1 free-period cutoff

API calls are free until November 1, 2026, after which users must earn or spend credits. If the free period drove sign-ups that do not convert into sustained use, active counts would fall. The credit model requires users to contribute data or models to keep working, which may deter casual adopters.

Historical Context

2 moments from history that rhyme with this story — and how they unfolded.

October 2019

GAIA-X sovereign cloud initiative (2019)

Germany and France launched GAIA-X, an initiative to build a European cloud infrastructure that reduces dependence on American providers like Amazon and Microsoft. Backers wanted European data stored and processed under European rules.

Then

GAIA-X grew into a standards and certification body with hundreds of members.

Now

Adoption lagged behind the US hyperscalers it was meant to rival, showing how hard sovereign infrastructure is to scale.

Why this matters now

EVREN pursues the same data-sovereignty goal as GAIA-X, but inside a single country with defense-grade requirements and a credit-based contribution model.

July 2023

Meta opens weights of Llama models (2023)

Meta released Llama 2 with open weights, letting organizations run large language models on their own servers instead of sending data to a vendor. This made on-premise AI practical for government and defense users.

Then

A wave of open-weight models followed, giving institutions a closed-model alternative.

Now

Open weights became the default route for organizations that cannot send sensitive data to cloud vendors.

Why this matters now

EVREN hosts 11 open-weight models, the same class of model that made private, on-premise AI deployment possible in the first place.

Sources

(6)