Ukraine’s Ministry of Defense is offering domestic and allied defense companies controlled access to millions of annotated frames from real combat drone missions through Avengers Labs, a platform inside the Brave1 defense-innovation cluster. The arrangement matters because battlefield computer-vision systems depend on rare, real-world training data, and Kyiv is tying access to a requirement that improved models return to Ukraine.

According to Ukrainian defense and digital-transformation officials, the program works through a protected Brave1 Dataroom built by the Ministry of Defense, the Ministry of Digital Transformation, the Armed Forces, a military-intelligence research institute and Palantir. Companies do not receive raw combat footage; they work inside the controlled environment on structured visual and thermal datasets.

The dataset is described as including aerial and ground targets recorded across tens of thousands of drone sorties, including camouflaged armor, night scenes, fog, rain and feeds from multiple sensors. Ukraine says more than 100 Ukrainian companies already have Dataroom access, while international developers are being brought in through Avengers Labs.

The ministry’s operational Avengers platform is separate from the partnership layer but connected to its purpose. By the ministry’s account, Avengers detects, classifies and tracks hostile targets in near real time, flags about 12,000 enemy units per week and feeds the VEZHA streaming module inside DELTA, Ukraine’s battlefield-management system.

Combat Data Becomes Strategic Currency

The program points to a shift in defense AI competition: the hardest asset may be verified operational data rather than model design alone. Computer-vision systems trained on clean or synthetic images can fail when weather, camouflage, electronic warfare, motion blur and sensor limits change the picture. Ukraine's position is that its front-line footage captures those conditions at scale.

For defense companies, the Dataroom offers a way to test algorithms against conditions that cannot be easily recreated in a lab. For Ukraine, the exchange creates a path to convert combat experience into finished software while keeping sensitive footage inside a Ukrainian-controlled environment.

The model also gives Kyiv leverage with partners. If the most useful training data remains inside Ukraine, outside firms may need to accept Ukrainian terms to improve battlefield vision systems for drones, counter-drone systems and fixed sensors.

From Army of Drones to Dataroom

Mykhailo Fedorov, formerly Ukraine's digital-transformation minister and associated with the country's Army of Drones effort, became defense minister in January 2026, according to the source reporting. He has described Ukraine as holding an unusually large body of annotated battlefield imagery built from real drone missions.

The Avengers Labs effort sits inside Brave1, Ukraine's defense-innovation cluster, and is linked to wider efforts to automate detection while keeping a human involved in lethal decisions. The stated focus includes onboard machine vision for front-line drones, operation in GPS-denied or jammed airspace, and improved detection or interception of Shahed-type drones.

"a body of battlefield data unmatched anywhere in the world"

— Defense Minister Mykhailo Fedorov

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.

Access Rules and Results Remain Opaque

Several details are not yet public or independently verified. The exact access rules for foreign firms, pricing or exchange terms, export-control limits and vetting standards have not been fully detailed.

Ukraine's claims about dataset scale and weekly target detections are attributed to ministry reporting. It is also not yet clear how models trained in the Dataroom perform across units, weather, jamming conditions and new Russian tactics once deployed at scale.

Model Trials Move Inside Dataroom

The next test is whether Avengers Labs can convert controlled access into fielded models for detection, navigation and counter-drone missions. Ukrainian officials have described goals that include onboard machine vision across front-line drones, GPS-denied navigation and automated Shahed interception with human authorization retained for any use of force.

Defense companies seeking access will be watched for whether they accept the model-return requirement. Allies will also watch whether Ukraine can protect sensitive footage while moving faster than Russian adaptation in drone warfare.

Key Questions

What is Avengers Labs?

Avengers Labs is a partnership platform run by Ukraine's Ministry of Defense inside the Brave1 defense-innovation cluster. It lets approved defense developers train AI models on annotated combat imagery inside a protected Dataroom.

Is Ukraine selling raw combat footage?

No. The program is described as controlled access inside a secure environment. Companies train and validate models there, while raw footage remains inside the Dataroom.

Why is the dataset valuable?

Real battlefield computer-vision data is hard to obtain because it comes from drones operating under combat, weather and electronic-warfare conditions. Those conditions are difficult to reproduce with clean training images or synthetic data.

Does this make drones fully autonomous weapons?

The stated focus is detection, tracking, navigation and counter-drone support. Ukrainian officials describe human control being retained for lethal decisions, but the exact deployment rules are not fully public.

What remains unverified?

Independent confirmation is limited for the dataset's full size, the exact number of detections, the terms for foreign companies and how well trained models perform once moved from testing into live operations.

Source: Thorsten Meyer AI