The Reichstag in Berlin, seat of the German Bundestag. Credit: Paul Scheelen via Pexels The Bundeswehr is assessing 30 artificial intelligence tools that could help speed up battlefield decision-making by processing data from drones and satellites. A prototype is expected in the second quarter of 2027, with full deployment planned for 2028. Vice Admiral Thomas Daum, who heads the military’s cyber command, gave the timeline in remarks reported by Reuters on Thursday.
He said the evaluation is still underway and that he is optimistic about the outcome. Most of the companies involved are German, while French firm ChapsVision is among the few foreign candidates still in the running. The name missing from the list is what makes the story interesting. Germany ruled out Palantir in April, rejecting the Maven Smart System already used by the Pentagon and several NATO members.
That left Berlin with a fairly simple question: if it did not want the American system, what would it use instead? Now there is an answer, and it comes with a timeline. The timeline is also the uncomfortable part. Maven is already deployed.
Germany’s alternative is expected to reach the prototype stage in about 18 months and full deployment roughly a year after that, assuming the programme stays on schedule. Defence projects rarely come with such certainty. Sovereign military software can reduce dependence on foreign suppliers, but it takes time to build, test and integrate. In the meantime, other armies are already using systems that Germany is still developing.
Ukraine shows why the trade-off is difficult. The amount of imagery coming from drones and satellites has grown far beyond what analysts can process manually. That makes a strong case for using the best tools available now. At the same time, the war has shown how much leverage comes with control of the software layer, particularly when the software sits close to military intelligence and targeting decisions.
Berlin has been working through that tension for some time. France and Germany pledged to build a European alternative to Palantir, only for German officials to later raise concerns that the French-led project could create another dependency. Evaluating 30 mostly German systems suggests Berlin is taking a different route: rather than betting everything on one European champion, it is building a field of domestic options. Whether that is the better approach is another question.
Thirty tools competing for the same decision-support role points to a market that has not yet settled. The difficult part will come later, when those systems need to work together. Military software programmes often struggle at the integration stage, and one of Palantir’s selling points is that its different capabilities sit within a single system. Germany is trying to get to the same point by putting several pieces together.
That gives it more control over the technology, but it also makes the project harder. It helps explain why the prototype is still 18 months away. Daum’s cyber command is also an interesting place for the programme to sit. Battlefield decision support is not purely a cyber problem, but putting the evaluation there suggests Berlin is approaching AI as part of the military’s wider data and network infrastructure before treating it as a targeting system.
That makes sense. It also explains why the starting question is which tools can process the data rather than which system a commander would actually rely on in the field. Germany took 90% of Europe’s record defence tech funding last year, and a domestic supplier base has grown significantly since 2022. Startups such as Comand AI have raised money to build command software for European forces.
A procurement programme like this is one of the clearest ways to turn that investment into something the military can actually use. The important milestone, however, will not be the prototype. It will be what happens to the 30 candidates afterwards. Germany will need to narrow the field quickly and make sure the systems that survive can work with each other and with NATO infrastructure.
If it cannot, 2028 could arrive with a collection of capable tools that still do not form a useful decision-support system. At that point, the argument for buying the American one will be easy to make again, simply by pointing at the calendar.
