Half a Million Hours of Ukraine Drone Footage Now Training AI

Real combat footage enters AI training market

A Virginia-based AI startup has released what it claims is the largest collection of real-world combat drone footage available for artificial intelligence training, drawing from more than half a million hours of video recorded during the ongoing Ukraine conflict.

Enabled Intelligence announced the dataset addition to its EView library, marking the first time Ukraine full-motion video has been made commercially available in a pre-labeled, validation-ready format. The company specializes in data labeling and AI model development for government and commercial clients.

Why it matters

The availability of authentic battlefield footage represents a significant shift in how defense AI systems are trained. Unlike simulated environments or controlled testing scenarios, this dataset captures unpredictable real-world conditions—varied weather, diverse terrain, and actual combat dynamics—that synthetic data cannot replicate. For defense contractors and military organizations developing autonomous targeting systems, the difference between training on simulated versus authentic operational footage could determine whether AI performs reliably in actual deployment.

What's in the dataset

The collection spans multiple sensor types including electro-optical, synthetic aperture radar, infrared, and foreign-language audio. According to Peter Kant, Enabled Intelligence's CEO and founder, the imagery has been labeled across three primary categories: aerial object detection, vehicle classification, and ground activity.

"This is the first Ukraine full-motion video in our EView library," Kant told DefenseScoop. "What sets it apart is that it's real—not simulated, not a controlled environment."

The pre-labeled nature of the dataset addresses one of the most time-consuming bottlenecks in AI development. Organizations can bypass months of manual annotation work and move directly to model training and deployment.

Applications across defense and commercial sectors

Kant indicated the dataset would prove particularly valuable for aerial applications, allowing different drone types to "quickly be AI-enabled." Defense use cases include intelligence gathering, offensive and defensive operations, and logistics missions in contested areas where human presence poses excessive risk.

On the commercial side, the CEO pointed to item delivery and remote sensing as near-term applications that could benefit from training on authentic operational footage.

The dataset is currently available to approved users in the United States, Ukraine, and NATO member nations. Kant did not disclose the specific sources of the video footage or identify current government customers using the collection.

Company background and government contracts

Founded in 2020, Enabled Intelligence provides data acquisition, precision labeling, and custom model engineering to U.S. military and government agencies, as well as clients in healthcare, financial services, and energy sectors.

The National Geospatial-Intelligence Agency awarded the company a single-award contract in 2025 valued at up to $708 million over seven years. The Sequoia data-labeling-as-a-service contract supports computer vision algorithms used in intelligence, surveillance, and reconnaissance operations, including foundational work for NGA's Maven program components.

"Ukraine has produced more real-world drone footage than any conflict in history," Kant said. "That data is only valuable if someone has done the hard work of making it usable."

These details were first reported by DefenseScoop.