By Dr. James M. Minnich
HONOLULU — May 21, 2026

The changing character of warfare is no longer a theoretical debate about future conflict—it is an active reality unfolding on the battlefields of Ukraine. Initial Western assessments often portrayed the Russian military as rigid, hierarchical, and slow to adapt. Yet emerging evidence suggests a far more dangerous evolution. In a recent Dialogue interview, Kateryna Bondar of the Center for Strategic and International Studies (CSIS) Wadhwani AI Center described Russia’s emerging “sovereign drone ecosystem,” in which military advantage is increasingly defined not by the sophistication of a single platform, but by the speed of adaptation under wartime pressure.

As Bondar warned during the discussion, “Russians are not thinking as a Red Army anymore.” Instead, Russia is increasingly combining centralized state coordination with decentralized tactical adaptation designed to compress the timeline between battlefield experimentation and operational deployment.

At the center of this shift is a structural mismatch in how modern militaries innovate, acquire, and field capability. Traditional Western procurement systems are largely built around linear peacetime models: defining requirements, competing multi-year contracts, and conducting lengthy developmental testing to minimize risk. This framework is optimized to produce highly advanced, durable capital platforms.

Modern software-defined warfare operates differently.

As Bondar’s research highlights, battlefield systems now evolve in cycles measured in weeks rather than years. Software updates can fundamentally change operational outcomes in ways that once required entirely new equipment. In contested electronic warfare environments, rapid software iteration now improves targeting, survivability, and operational effectiveness in near real time. In this environment, adaptation speed becomes combat power.

The Garage-to-State Pipeline

Russia has increasingly bypassed traditional acquisition structures by adopting a wartime innovation model that blends civilian technology networks, volunteer engineering communities, private-sector developers, and state-directed scaling mechanisms. Innovation often originates outside formal defense-industrial channels at a decentralized “garage” level, where systems are tested, modified, and refined under live battlefield conditions.

Combat validation becomes the ultimate filter.

Systems are judged less by theoretical performance or acquisition milestones than by whether they survive and succeed under battlefield pressure. Battlefield effectiveness increasingly outweighs peacetime acquisition logic. Systems are iterated continuously under combat conditions rather than perfected before deployment. Once a platform, software package, or tactical modification demonstrates effectiveness at the frontline, the Russian state rapidly finances, standardizes, and mass-produces it at scale.

Rather than attempting to centrally design perfect solutions from the outset, the system increasingly captures organic tactical adaptation and industrializes what already works.

Software-Define Warfare

This model has accelerated the evolution from remotely piloted drones vulnerable to electronic warfare disruption toward increasingly autonomous systems capable of operating in contested environments. Technical analysis of captured Russian systems—including the V2U drone—reveals a significant architectural shift: reduced dependence on continuous operator control links combined with onboard computing sufficient to support AI-enabled perception, navigation, and targeting functions.

The contrast with traditional acquisition models is increasingly stark:

Traditional Western Model:
Exquisite Hardware Platform → Rigid, Proprietary Software Integration

Modern Wartime Adaptation Model:
Cheap Commercial Hardware → Rapidly Updated AI/Software Layer → Immediate Frontline Deployment

Rather than competing to develop cutting-edge general AI systems, Russia’s approach appears focused on applying commercially available technologies to narrowly defined battlefield tasks. Globally available semiconductors, open-weight AI architectures, and dual-use commercial technologies are integrated into software-driven strike and reconnaissance systems designed to compress the timeline between detection, decision, and engagement.

Bondar argued that AI is increasingly valued not simply for autonomy itself, but for its ability to rapidly fuse information from multiple sensors and battlefield systems. Describing a high-stress tactical environment in which a commander faces 70 separate targets with only minutes to decide, she observed, “the human brain is not capable of processing this amount of information in this amount of time.”

Training as Combat Power

The scale of this ecosystem is perhaps most visible in its approach to training and force generation.

To field these capabilities operationally, Russia has developed a structured training architecture. Moscow has announced plans to recruit approximately 160,000 drone operators, signaling a force structure increasingly centered on massed drone operations and semi-autonomous systems rather than traditional mechanized assumptions.

Private drone schools and parallel training initiatives now operate alongside formal military institutions, often adapting with startup-like speed. These organizations recruit experienced operators directly from combat units and continuously revise training curricula to reflect evolving battlefield realities, software modifications, and emerging electronic warfare countermeasures.

Training itself becomes part of the adaptation cycle.

Operators test and refine systems during instruction, creating immediate feedback loops between battlefield performance, software iteration, and force development. Training ceases to be a static prerequisite and instead becomes a dynamic accelerator of technological adoption.

The Institutional Challenge

For Western defense establishments, this presents a direct institutional challenge. Procurement systems that require months or years to certify software changes or modify programmatic requirements struggle to compete against adversaries operating on live-data feedback loops. Likewise, centralized training doctrines and rigid institutional structures may increasingly find themselves outpaced by adaptive ecosystems built around continuous variation and rapid operational learning.

Bondar cautioned against relying on outdated assumptions about future conflict. Critiquing traditional European security models centered on mechanized armor crossing borders, she warned: “Don’t expect tanks to cross the border, especially in the beginning. Expect thousands of Shaheds crossing your airspace, together with Gerberas and decoys and other systems, which will distract and kill your air defenses.”

The broader lesson emerging from Bondar’s work is clear: in an era defined by algorithmic speed, autonomous systems, and multi-domain competition, technological elegance matters less than the velocity of adaptation. Operational advantage increasingly belongs to the side that can compress the timeline between battlefield feedback, software modification, procurement, training, and deployment.

The defining competition may no longer be platform versus platform. It may be adaptation cycle versus adaptation cycle.

Dr. James M. Minnich is a professor at the Daniel K. Inouye Asia-Pacific Center for Security Studies. This article reflects the author’s analysis and does not represent the official policy or position of any institution.