The fact that drones can quickly become obsolete is not an argument for waiting to procure them. Units, developers and industry must work together already in peacetime. Otherwise we risk having plans for tomorrow's drones, but lacking the capability to use and produce them when they are needed.
No elite football team would train an entire pre-season without balls, only to buy the latest match balls once the league starts. The players need balls from day one. Not primarily to get to know the ball, but to develop the team.
Yet we risk ending up with a similar line of reasoning when it comes to drones.
NATO Secretary General Mark Rutte described on 23 September how he believes the alliance should approach rapid technological development. Because drones quickly become obsolete, he argued that it "doesn't make sense to have your shelves full with drones for when war breaks out". Instead, production capacity should be in place so that the latest technology can be manufactured when it is needed.
Rutte has a point. Buying millions of drones of a particular model and putting them in storage without a plan for use, maintenance and updates would hardly be wise. But that is not the same thing as building a functioning drone capability.
Units need equipment to train, exercise and develop their methods. Industry needs to manufacture, deliver and improve the systems in order to scale up production. Both of these take time and require experience.
At the same time, Rutte has announced that NATO (North Atlantic Treaty Organization) will train five times as many drone operators by the end of 2027. This is welcome, but the training must be followed by a procurement model that gives units ongoing access to equipment.
There is also a link here to NATO's own readiness principle, Fight Tonight. The alliance must be able to carry out its tasks with the units and resources available today, while simultaneously building tomorrow's capability. This also applies to drones and the industry that is to supply units with them.
When the drone war in Ukraine is discussed in the West, the focus often falls on new technology. But technical novelty and military significance are not the same thing. A seemingly limited change can have major implications for how a system is used.
When I visited Lviv (western Ukraine) in autumn 2025, I heard Kimberly Kagan, founder of the Institute for the Study of War, describe the relationship between technology, concept and organisation. I took away a simple summary:
Technology + concept + organisation = innovation.
A drone does not become a military capability simply because it can fly. Someone must develop how it is to be used, train personnel, and make it work together with the unit's other resources.
Ukraine's initiative Drone Line is one example. According to the country's Ministry of Defence, it is about spreading solutions that have proven to work, combining units' experience with technology and analysis, and building a more systematic use of unmanned systems.
Ukraine has also decided to link drone procurement more closely to data from units' actual use. Systems that have demonstrated effectiveness and reliability are to be prioritised, while units are given greater scope to choose equipment themselves based on their needs.
In other words, they are not waiting for the next technological generation to be finished. They are developing it by using the previous one.
This is an important lesson for NATO. Units need enough drones to train, practise and develop their operations. Planning must also allow for wear, failures and systems that are consumed during exercises and development.
The same logic applies to industry. Production capability is not a recipe that can be filed away until someone presses the start button. It is built on people, suppliers, functioning procurement, testing, quality assurance, and experience in managing change.
Having manufactured a prototype is not the same thing as being able to deliver thousands of systems. Nor does being able to manufacture one model mean that the next one can be introduced without deliveries grinding to a halt.
If we also want to reduce dependence on China, it is not enough to simply final-assemble drones in Europe. Alternative component supply chains must be built up and used. Suppliers in Sweden and other allied countries need real orders in order to identify bottlenecks and improve production before demand multiplies several times over.
It may be justified to pay for reserve capacity and stockpiles of critical components. But reserve capacity should be built on a functioning operation with a proven ability to scale up.
NATO's own military leadership has spoken of "real production lines running today". This is an important distinction. A factory that could potentially start producing in wartime is not the same thing as an industrial capability that is already functioning.
We at Sweden Dynamics believe that part of the solution can be found in an area that has long had to deal with rapid change: software development.
Drawing inspiration from Software-as-a-Service, we propose a model that we call Drone-as-a-Service. The basic idea is that procurement should encompass the ongoing development of capability, in which hardware, software and units' experience are held together.
Developers work closely with military units to test solutions, adapt the technology and verify what works. Smaller production runs allow room for trial and change. Once a model is sufficiently mature and the need is clear, larger volumes can be ordered.
For these volumes, we should make use of the civilian electronics industry's contract manufacturers, so-called EMS (Electronics Manufacturing Services) companies. These already have expertise and capacity in component supply, electronics production, testing and systems integration.
Depending on the complexity of the system, we estimate that a production line can manufacture hundreds, or over a thousand, smaller drones per month. The point is not to keep factories running at maximum capacity year-round, but to let several production facilities operate on an ongoing basis, introduce new models, and keep the component supply chain active as well.
With recurring orders, units get equipment to use, developers get feedback, and factories gain experience with both the product and the changes made to it. New versions can be introduced as part of regular operations rather than becoming separate development projects.
This would also create more predictable demand. As an illustrative calculation, we can assume an annual need of 2,000 to 3,000 mainly smaller drones per brigade for training, exercises, development and replacement. For four brigades, that corresponds to 8,000 to 12,000 drones. With additional needs for schools, joint development and other units, we arrive at a figure in the order of 10,000 to 15,000 per year.
This is not a fixed requirements calculation, but it illustrates how recurring orders could give both units and industry a foundation from which to develop. At the same time, peacetime production needs to be dimensioned so that there is the capacity to rapidly increase deliveries.
In addition, readiness stockpiles are needed. We believe planning should be based on units being able to maintain their capability for three to five months of warfare, even before production has had time to scale up. To prevent stockpiles from filling up with outdated technology, they should be rotated continuously on a first-in, first-out basis. Units would then use older systems for training and exercises while new versions are added to the stockpiles.
Production not needed for Swedish units could also be included in Swedish support packages to Ukraine. In this way, the production chain can be kept active while the systems are put to use and can be further developed based on lessons learned from the war.
Volumes are not the only thing that should guide decisions, either. Where the task allows for reuse, longer service life or higher reliability, that must also be factored in. A cheap drone is not automatically a cheap military capability.
We do not need to choose between stockpiles and factories. We need usable equipment with the units, readiness stockpiles that are rotated, and ongoing production that can scale up when needs increase.
This is also how rapid technological development should be managed: through closer feedback loops, upgradeable systems and recurring deliveries. Not by waiting for a final version that never arrives.
Fight Tonight does not mean that the technology we use today must be the same in five years' time. It means that we have a functioning capability now and can continue to develop it as systems and countermeasures change.
The latest drone cannot replace years of training and exercises that were never carried out. A production chain that is only activated once war arrives cannot deliver capability tonight.
When the match begins, the team must already be able to play. The same applies to the units, and to the industry behind them.