Latvia's defense minister urges NATO arms makers to disperse production before a war

The board-deck version of Raivis Melnis's case fits on one slide: split production across several sites so a single Russian strike cannot stop it ^[1]. That version leaves something out. The executives who agree with him also say one big plant is cheaper. Karmo Saar, head of sales at Krattworks, an Estonian maker of drones used by Ukraine ^[17], said that for Ukrainian companies it "would be a lot more economical, cheaper, and better to run this operation in one big facility, but they cannot do that" ^[13]. The chief executive of Ukrainian-Estonian Ark Robotics, which already makes different components at different sites, called a single site "the best way to do it" ^[11].

So the skeptic's case comes from inside the dispersal camp, and so does the answer. The Ark Robotics chief said spreading out should be the "default for defense-based manufacturing going forward," outside Ukraine as well ^[12]. Saar was blunter about his own continent. "In Europe, we do that," he said. "And when the war strikes, I think we're going to be punished." ^[14]

Ukraine's model also has a public backstop. Kyiv compensates some businesses whose property is damaged or destroyed in Russian attacks, and it is planning more support ^[7]. Melnis counted that as part of the lesson, saying Ukraine has shown how a country "can provide that reassurance if, for instance, an industry site is destroyed" ^[6]. The step he described taking for Latvia, a NATO and EU member that borders Russia ^[16], is narrower. He encourages partnerships between Latvian and Ukrainian defense companies so Latvian firms can learn how Ukrainian systems are "built for war purposes" ^[5].

Srđan Kovačević, chief executive of Croatian drone firm ORQA, which has signed up to begin working in Ukraine, stated the trade-off directly ^[15]. "This distributed nature actually reinforces resilience, even though there are economic downsides to it," he said ^[15]. None of the executives put a figure on those downsides.

Melnis set no date for any of this, and the evidence does not make it a problem for this week. His argument is about sequence. He said "the question is whether we are ready to produce the systems in different locations, dispersed, and how we organize that" ^[4]. In my view, organising it is the harder half. A European producer that adds capacity this year in one large plant gets the lower cost now and puts its risk in one building for as long as that plant runs. A producer that splits the line pays the premium now. Unless a European government copies Kyiv's compensation, it also carries the risk of a destroyed site on its own ^[7].