Ukraine’s Ministry of Defense is turning front-line drone footage into a controlled AI training asset, allowing selected defense companies to train models on annotated combat data through Avengers Labs and the Brave1 Dataroom while Ukraine keeps the improved finished systems.

The program gives domestic and allied developers access to visual and thermal data captured during real drone operations, according to Ukrainian defense officials and Brave1 materials. The datasets include aerial and ground targets seen in conditions such as night, rain, fog, camouflage and electronic warfare interference.

Companies do not receive raw battlefield footage to take away. The training takes place inside a protected Dataroom built with Palantir software and operated through Ukraine’s defense-innovation ecosystem. Ukrainian officials have described the initial focus as AI tools for detecting and intercepting hostile drones, including Shahed-type attack drones.

The arrangement gives companies rare battlefield data to improve computer-vision models. In return, Ukraine receives the resulting models, creating a loop in which combat footage is captured, labeled, secured, used for training and returned to the war effort as improved software.

Data Becomes Military Leverage

The move matters because modern battlefield AI depends less on generic model design than on access to data that reflects real combat. Synthetic imagery and clean test footage often miss the conditions that make front-line detection difficult: jamming, smoke, poor light, camouflage, sensor noise and fast-moving targets.

For Ukraine, the program could speed development of systems that reduce the burden on drone operators and air-defense crews. For allied defense firms, it offers exposure to a dataset that few militaries can provide at scale. For readers, the development signals how the war is shaping the next market for defense AI: verified operational data is becoming as valuable as hardware.

From Brave1 To Dataroom

Brave1 was launched by the Ukrainian government in 2023 to connect the state, military and defense technology companies. It has since become a main channel for testing and scaling drones, electronic warfare tools, robots and battlefield software.

In January 2026, Brave1 and Palantir announced the Brave1 Dataroom, described by Brave1 as a protected environment for using battlefield data to test and train military AI models. The Dataroom was presented as a joint effort involving the Ministry of Defense, the Armed Forces of Ukraine, a military-intelligence research institute and Palantir.

Avengers Labs builds on the Avengers AI platform, which Ukrainian officials say is used to detect, classify and track hostile targets from drone and camera feeds. Ministry-linked reporting says the system feeds Ukraine’s DELTA and VEZHA battlefield information tools.

"AI is becoming a decisive factor on the modern battlefield."

— Mykhailo Fedorov, Ukraine’s defense minister

Thermal Camera Fpv 256/384/640 9MM Thermal Imaging Camera For FPV Drones With CVBS Interface Micro Imager Module(256X192 9mm CVBS)

High-Resolution Thermal Imaging: This FPV thermal camera module is available in multiple resolutions (256, 384, 640) to deliver...

AmazonView Latest Price

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Limits Still Not Public

It is not yet clear which foreign companies will be admitted, how many have already trained models inside the system, or what contractual limits govern the finished software. Ukrainian officials have not publicly released full details on data-labeling standards, independent testing, error rates or how battlefield performance will be measured.

The military effect is also still developing. Ukrainian officials say automated detection can help operators and improve drone interception, but the performance of models trained through Avengers Labs has not been independently verified.

Model Testing Moves Frontward

The next phase is expected to center on expanding Dataroom access, adding more target classes and testing finished models against active battlefield needs. Ukraine has said the first focus is autonomous detection and interception of hostile UAVs, while keeping a human role in lethal decisions.

If the model-return arrangement works as described, Ukraine could gain a faster pipeline from combat data to deployed AI tools, while allied developers use the Dataroom to harden their systems against conditions that ordinary test ranges cannot fully reproduce.

Key Questions

What is Avengers Labs?

Avengers Labs is described as a partnership layer connected to Ukraine’s Ministry of Defense and Brave1 ecosystem that lets selected defense developers train AI models using protected battlefield datasets.

Does Ukraine give companies raw combat footage?

No. The described model keeps sensitive data inside the Brave1 Dataroom. Companies train and test models in that environment, while Ukraine retains the finished improved models.

Why is this data valuable?

It comes from real drone missions under battlefield conditions, including jamming, weather, darkness and camouflage. Those conditions are hard to reproduce with public data or simulations.

What is the first military use case?

Ukrainian officials say the initial focus is AI for detecting and intercepting enemy drones, including Shahed-type systems used in Russian attacks.

Source: Thorsten Meyer AI