In May 2025, the multinational field exercise Hedgehog 2025 (also known as Siil 2025 in Estonia) brought together more than 16 000 personnel from 12 member states of the North Atlantic Treaty Organisation. With scenarios centered on a simulated incursion across Estonia’s eastern border, the drills were intended to test European deterrence and defense concepts under high pressure (Defence24.com, 17.02.2026).
One episode from those manœuvres quickly became emblematic of a deeper structural challenge for allied militaries. A small contingent of Ukrainian drone operators, acting in the role of an opposing force (OPFOR), outpaced and “neutralized” a larger NATO grouping — including a British brigade and Estonian units — within the span of a day. Using a mix of unmanned aerial systems (both reconnaissance and loitering munitions) and a battlefield management tool known as Delta, the team was credited with the simulated destruction of 17 armored vehicles and about 30 additional engagements, effectively rendering two battalions combat-ineffective in the context of the exercise.
What distinguished this outcome was not merely the presence of drones, but the velocity at which sensor data was collected, fused and acted upon. Ukrainian operators drew on years of front-line experience against Russian forces, and on systems designed to compress the kill chain — from detection through to strike — into a matter of minutes. NATO units participating in the exercise, by contrast, repeatedly exposed vulnerabilities associated with traditional manœuvre patterns, limited concealment and slower decision cycles when faced with persistent aerial ISR (intelligence, surveillance and reconnaissance) pressure (Ukrainian Pravda, 18.02.2026).
Exercise observers did not characterize the outcome as a definitive statement on NATO’s future combat effectiveness. Staff involved stressed that the purpose of such war games is to identify weaknesses and inform doctrinal, organisational and capability development. Nevertheless, the results underscored a broader strategic trend: modern battlefields increasingly prioritize information dominance and rapid decision-making, attributes that legacy C4ISR (command, control, communications, computers, intelligence, surveillance and reconnaissance) architectures were not originally designed to deliver at scale.
The Failure in C4ISR Integration
At the core of the Hedgehog experience was a systemic lag in how allied forces integrated and exploited battlefield data. Large formations operated under assumptions that — for decades — had underpinned mechanized warfare: that position, concealment and tempo could be maintained without holistic, real-time situational awareness. Against an adversary exploiting low-cost, networked sensor platforms and rapid data fusion, these assumptions proved costly in simulation (Medium, 13.02.2026).
Traditional C4ISR systems in many Western militaries remain built around hierarchical information flows, where sensor feeds are aggregated and disseminated through structured channels that prioritize security and classification. In an era where tens of unmanned platforms can populate a battlespace and transmit data simultaneously, such architectures introduce friction that opponents with more agile data practices can exploit (Defence24.com, 17.02.2026).
The Hedgehog scenario revealed three specific operational deficiencies:
- A lack of continuous, decentralized data sharing between units and command echelons.
- Insufficient edge processing to translate raw sensor inputs into actionable insights without delays.
- Doctrine and training that did not fully internalize the implications of an ISR-dense environment.
Toward a More Decentralized, Data-Intensive C4ISR Approach
The lessons from Hedgehog suggest a need for a fundamental re-orientation of allied C4ISR concepts. Instead of treating intelligence as a segregated output to be processed at command centres, future approaches must prioritize the continuous flow of raw data and its transformation into patterns, predictive cues and real-time decision support.
Two complementary strands are central to this evolution:
1. AI-Enabled Data Fusion and Pattern Analysis
Machine learning and artificial intelligence are essential not simply for processing information more quickly, but for recognizing complex patterns in large datasets that exceed human analytical bandwidth. Systems such as Delta — which ingest UAV feeds, ground sensors and other ISR inputs and synthesize a common operational picture — highlight the direction needed for future doctrinal adaptation. AI’s role is not limited to automation; it also supports predictive analytics that can anticipate threat movements or suggest courses of action based on historical and real-time data correlations.
However, maintaining and evolving such capabilities require access to large volumes of raw data from diverse sensors, not only aggregate summaries. Without this raw input, AI models risk over-fitting to narrow scenarios and failing to generalize to novel operational contexts.
2. Digital Twins and Synthetic Data Environments
Digital twin technology — virtual models of physical systems or operational environments — offers a way to stress-test doctrine and technologies under controlled yet highly realistic conditions. By integrating historical war data with synthetic scenarios, digital twins allow analysts and commanders to explore system behaviors at scale, evaluate potential outcomes and refine force designs before real-world deployment.
Work on synthetic environments enables wargames to include thousands of autonomous agents, realistic terrain models and multi-domain interactions without the logistical limitations of physical exercises. When linked with machine-learning pipelines, these environments can generate synthetic training data that enriches AI models and enhances their predictive and interpretive capabilities.
Organizational and Doctrinal Imperatives
Technical transformation alone will not suffice. The Hedgehog experience pointed to deeper cultural and organisational imperatives. NATO allies must cultivate:
- Distributed decision-making, where front-line units have both the authority and the tools to act on real-time data.
- Interoperable data standards, ensuring cross-national forces can share and interpret information without friction.
- Training and education, realigning curricula to emphasize cognitive agility and multi-domain awareness.
As one after-action reflection from exercise participants indicated, recognizing these needs is an important step; converting them into new doctrine and practice is the challenge that lies ahead (Defence24.com, 17.02.2026). The apparent architectural paradox between a data-intensive, AI-enabled C4ISR architecture and decentralized, real-time situational awareness can be resolved through a layered, cascaded design. Central C4ISR and AI processing nodes manage large-scale data aggregation, model training, and cross-theatre analysis, while mission-proximate edge servers operate with compressed, mission-relevant datasets and embedded AI reasoning capabilities, supporting localized command and control functions. Continuous and resilient interconnection among operational actors and systems remains essential to ensure coherence, synchronization, and shared understanding across the network.
Lessons Learnt
The Hedgehog 2025 exercise was not a failure in the classic sense, but a revealing assessment of how rapidly warfare is changing and how slowly large military coalitions adapt. The simulation highlighted that dominance in unmanned systems and real-time data integration can redefine the tempo of operations. In response, allied C4ISR approaches must embrace decentralized data flows, AI-driven pattern analysis and immersive synthetic environments such as digital twins. Only by doing so can future exercises and, should it ever be necessary, actual operational deployments better reflect the demands of 21st-century conflict.
Stay Tuned
Stay connected with our ongoing publications on Electronic Warfare and Radar Technology.
