
AI, Drones and Cyber: Why Switzerland Must Rework Its Procurement
New systems become outdated faster than classic large-scale projects. What matters, then, is not only what the army buys, but whether it can test, integrate, upgrade and take responsibility for it.
Anyone who talks about a high-tech revolution in defence often thinks first of spectacular images: small drones above the battlefield, artificial intelligence analysing imagery, or cyberattacks on screens. For Switzerland, however, the harder question comes before any deployment. It is: can a system be procured in time, safely embedded in existing structures, updated, and kept operational in an emergency?
The Federal Council has in the meantime outlined the answer at a strategic level. The 2026 Security Policy Strategy prioritises resilience, protection and defence capability. It names hybrid threats and attacks from a distance as likely risks. The defence guidelines add that the protection of critical infrastructure, command capability, counter-drone defence, electronic warfare and the protection of networks and data should be given priority.
That is more than a list of new equipment. It is an admission that the logic of procurement has to change.
The bottleneck is the time between testing and deployment
Classic armament programmes are often planned over many years. With software, sensors, counter-drone systems or digital communication systems, that timespan can become problematic: technology keeps advancing, threats change, and a system may lose relative capability before it is widely fielded.
The Federal Council's armament policy strategy responds to this with a clear course. Procurements should be faster, more flexible and more closely aligned with technology cycles. Where it makes sense, the army should rely on systems available on the market that meet international standards. Special requirements that would have to be developed for Switzerland alone should be reduced.

This is not a minor detail buried in the administration. It affects operational readiness. A technically sound system is of little use if spare parts are missing, updates arrive too late, or it does not work reliably with partners and existing command networks. Conversely, a market-available solution can only make sense if it fits into the Swiss security architecture and is mastered both legally and operationally.
Speed requires verifiable limits
Faster procurement must not be confused with uncritical procurement. Especially with security-relevant software, sensor technology and radio links, testing, independent technical assessment and clearly defined operational limits are needed. The gain comes when these checks happen early and repeatedly – not when they have to be made up at the end of a shortened project.
Sovereignty does not mean building everything at home
Switzerland cannot manufacture every security-relevant technology itself. Its armament policy therefore makes no claim to autarky. It combines domestic capabilities with European cooperation: for goods outside the state treaty area, it aims – where economically and security-politically sensible – for a higher share of procurement in Switzerland and in Europe.
The goal behind this is more concrete than a label of origin. It is about knowledge, maintenance, secure data channels and supply chains that do not abruptly break off in a crisis. The strategy envisages, for instance, identifying critical supply chains, reducing dependencies and diversifying suppliers. It also aims to step up the international exchange of tests and evaluations.
For ordinary Swiss citizens this level is directly relevant, even if it remains invisible. An interrupted supply chain or a system that has not been updated does not affect the army alone. With digital or dual-use technologies, transport, energy supply, communications and emergency organisations can also be affected.
AI should support decisions – responsibility remains human
At the VBS Centre for Artificial Intelligence and Simulation, work is underway on applications that detect threats, prioritise them and support training. Automated image analysis and the simulation of drone swarms are also among the publicly described projects.
That can speed up situational assessment. But it does not replace political, legal and military responsibility. The published projects describe support for detection, prioritisation, planning and training. That is precisely why it must remain transparent who takes decisions, how data is protected and how faulty or manipulated AI results can be identified. Speed must not lead to responsibility disappearing inside the system.
Dual use makes the boundary more permeable
Through a joint initiative by Innosuisse and armasuisse, the Confederation funds technologies that can be used for both civilian and security-policy purposes. The focus includes robust navigation and time measurement, secure communication, and the detection and defence of drones. These topics show why defence is no longer merely a question of military platforms.
The new technology policy will therefore not be measured by how many buzzwords or individual devices are presented. What matters are shorter, traceable procurement paths, robust maintenance, secured data and clear rules for use. Only when this chain works will technical innovation turn into actual security for Switzerland.
Sources
VBS: Security Policy Strategy 2026
armasuisse: Armament Policy Strategy of the Federal Council
VBS: Innosuisse and armasuisse launch call for projects on dual-use technologies



