CIHBw explains how startup projects reach soldiers and end in a recommendation. A successful pilot still needs an owner for permanent introduction.
Germany’s Cyber Innovation Hub offers startups an unusually concrete account of what happens after a pitch: a possible project with users, followed by a recommendation on introduction. The commercially important detail is the final decision. A company should design the collaboration so that evidence produced during a pilot can support the next buying step.
CIHBw’s startup page describes a focus on market-available dual-use and defence technology, projects lasting up to twelve months and direct work with soldiers. It asks for a problem-and-solution presentation of no more than fifteen slides. At the end, the Hub recommends for or against introduction; the Bundeswehr decides whether to introduce the technology permanently. Its listed focus areas include software-defined defence and AI, electronic warfare and drones. CIHBw’s startup process
A useful pitch should explain what the user currently does, where that process falls short and what a bounded evaluation would establish. A software business might propose measuring the effort required to complete a support workflow. A hardware supplier could focus on whether users can operate and maintain a product under the expected conditions. The proposal should make success observable without relying on a broad claim that the technology is transformative. The CR14 cyber range evaluation case examines the scope and reusable evidence a defensive-product experiment should produce.
The company also needs to state what it requires from the collaboration. Access to users, a representative environment, integration support and feedback are different inputs. Naming them helps the Hub understand whether the project is practical. It also prevents the company from budgeting a simple product demonstration when the real work requires extensive engineering support.
The pilot should produce material that another organisation can use: a clear account of the requirement, the evaluated configuration, the results and the remaining gaps. That matters because an introduction decision can involve people who did not attend the demonstrations. Evidence should remain useful after the original project team moves on.
A commercial team should also ask how support, licensing and scale would change if the product were adopted more widely. A pilot involving a small number of users can conceal expensive manual work. The company needs to distinguish a reusable product capability from custom effort that only makes sense if separately funded. That distinction helps determine whether successful evaluation can become a sustainable business relationship. The FFI ICE worx case focuses on the adaptation a commercial product may still need for a defence customer.
The public page explains an engagement route, but it does not promise every applicant a project or guarantee a procurement following a recommendation. Founders should budget preparation and participation accordingly. The meaningful milestones are a confirmed fit with a focus area, a scoped collaboration and evidence supporting a decision about the next phase.
The Hub's KI-PROcure case provides a concrete example outside the more visible hardware demonstrations. It describes work with procurement startup GovRadar on software that helps prepare purchasing documents and compare product characteristics. The project expanded into the Bundeswehr medical service, with participating hospitals and medical-material organisations. The Hub says pharmaceutical databases were integrated and users began testing the application locally. The Hub's KI-PROcure case
That account makes the adoption question more specific. The value of procurement software depends on the documents staff must produce, the information available to them and the work needed to review the result. An impressive text-generation demonstration would address only part of that process. A useful assessment would ask whether the resulting document can be checked, corrected and used within the purchasing organisation's established responsibilities.
The named medical users also explain why a successful demonstration in one office may require further work in another. Vocabulary, product catalogues and review practices can differ. The commercial significance of a pilot is partly its ability to reveal those differences while the scope is still manageable. That is a stronger basis for estimating an extension than assuming every additional department will use the same configuration without adaptation.
The current startup page asks about technology maturity, the company and its commercial stage, while its pitch guidance includes pricing and a product demonstration. It describes a ten-minute virtual pitch followed by questions. The form also requires an EEA presence, which may be a headquarters, subsidiary, reseller or another form of company there. These are features of the published intake process, alongside the requirement to match current focus areas. The startup intake and pitch guidance
Together, those questions suggest that the Hub is assessing more than an interesting research idea. A pilot needs a company able to explain what exists now, who will work on it and what commercial arrangement could support the work. A clear price model is useful even when the final scope has not been agreed: it reveals whether the business expects payment per user, per location, for implementation work or for an ongoing service.
For a hypothetical document-management startup, the most informative commercial comparison might be between a small initial user group and a later departmental deployment. The first could involve substantial setup relative to licence revenue; the second could require support staffing and additional administrative permissions. Explaining that difference helps a pilot team understand which costs are temporary learning costs and which would recur as usage expands.
The Hub's startup page reports more than 200 innovation projects since 2017 and more than 50 innovations in the Bundeswehr. It separately mentions over 200 startups taking part in its process in 2025. These figures describe different populations and periods. Dividing the innovation total by the project count would not establish a reliable success rate for a new applicant, and the startup figure cannot be substituted for the number of completed pilots.
The more useful evidence is project-level progression: who used the product, what task they attempted and what happened after the trial. A published case can demonstrate that a particular workflow reached users without revealing a final procurement value or the scale of permanent deployment. Commercial planning benefits from retaining that intermediate evidence, rather than forcing every case into a binary category of success or failure.
For the company conducting a pilot, this means that the final account should make the decision legible to someone who did not attend the demonstrations. It should connect the observed user benefit with the cost of continuing the service and the remaining organisational work. The prospective adopting office needs to understand both what improved and what responsibility it would assume by retaining the product.
KI-PROcure is useful precisely because its public account names a real administrative activity and describes expansion into another user community. That gives other startups a concrete model for reporting progress: explain the task and the change in its execution, then connect that evidence to a bounded next decision.
As of 6 September 2026, the current page provides a submission route and describes the Hub’s process without a universal application deadline. Companies should review its latest focus material before preparing the deck. The strongest commercial case connects an existing capability to a user problem, defines a manageable experiment and makes the requirements for permanent adoption explicit. That turns a pilot into a decision tool rather than an isolated demonstration.
Current CIHBw startup page reviewed 6 September 2026. Its project process is attributed to the Hub; permanent introduction is a Bundeswehr decision.
Suggest a correction