Defence needs faster engineering and production.

Advanced weapons and affordable equipment are both necessary

Some missions justify expensive and highly integrated aircraft, sensors and weapons. A typical example is the neutralization of a large, static target such as a logistics node or a command center, whose importance largely offsets the cost of a 200,000$ AASM guided bomb. Other tasks, such as repelling fragmented infantry assaults over large swathes of land, require large quantities of expendable and inexpensive equipment. Preparing for a long conflict means paying attention to both exquisite performance and affordable mass.

Building up mass requires habits that are unfamiliar to armed forces and legacy military contractors: accepting compromises in integration, and recognising that “whatever works” can be an acceptable design philosophy.

As long as equipment can reach the front line, be handled safely by troops, and prove useful against its intended targets, some loss of integration or precision may be acceptable.

Ukraine has built much of the backbone of its modern, integrated way of fighting as the equipment itself evolved. That is a considerable advantage. Western armies rely on logistics and network architectures inherited from a more monolithic doctrine. Compliance requirements should not hinder our ability to respond. Getting this balance right is difficult, which is why we want to work with military users, procurement teams and industry to establish where compromises are acceptable.

Updates need to be part of the design

Drone links and detection methods change as each side responds to the other. A design that works in one set of conditions may need a different radio, sensor or software update to remain useful.

This finding about UAV systems illustrates a general point about procurement: buying equipment also means arranging for the changes it will need after delivery.

We therefore want to develop power, communications, sensing and software as interchangeable parts. If one sensor needs replacing, we should be able to retain the other useful parts, test the revised configuration and update its assembly. This is the reason for our modular approach.

The West has competition

China is systematically studying the war in Ukraine and translating selected lessons into training, procurement and industrial development. CSIS finds an average of roughly four months between battlefield innovation and Chinese analysis, with implementation generally taking longer. The PLA remains untested in a comparable war, and exercises cannot reproduce frontline feedback. Even so, China appears to have established a serious process for learning and adapting before it has to fight. Europe cannot assume that its potential adversaries will be slower to update their equipment and methods.

Source: Lilly Min-Chen Lee and Kateryna Bondar, Learning Without Bleeding, CSIS · 7 October 2026. Redrawn and annotated by Weirding; no affiliation or endorsement implied.

Decentralized, loss-tolerant manufacturing

The typical defence manufacturing pipeline typically involves small numbers of large factories producing specialized parts. These are vulnerable to a range of contingencies, from sabotage, to, sometimes, economical woes (see the Forges the Tarbes, the only foundry able to make the CAESAR cannon barrels).

We want sourcing, assembly and quality checks to be decentralized, dynamic and close to where the equipment is used. It also needs to be loss-tolerant: a network of container-sized labs is much more costly to disrupt than a monolithic assembly line. That manufacturing approach is a proposed direction for the lab, described separately in Future Directions.

Why we want to remain a lab

We want to test alternative technical approaches within a project and stop using those that do not meet the requirement. If a sensing method becomes ineffective, the work on power, communications and software should still be useful. For deployed equipment, an update would also need to address compatibility and support.

Our current development effort is the Drone Wall. It needs to demonstrate useful detection, manageable operating costs and a design that can be updated and reproduced. Further projects should follow a defined need and make useful use of the engineering we have already done.