EXCLUSIVE The wars of the future are written with drones, data, and debts: how can Romania use the 16.6 billion euros from SAFE

Modern technology has already radically changed warfare – from cheap drones and "smart" munitions to AI, cyber warfare, and massive data exploitation – and the conflicts in Ukraine and the Middle East show that the future will be defined by volume, digital connectivity, and decision-making speed, not just by tanks and planes. For Romania, the credit line offered through SAFE is a historic opportunity, but also a risk: the money can anchor the army in an outdated industrial model or can accelerate a leap towards a force compatible with a battlefield saturated with drones, sensors, and cyber attacks.

How technology has changed the principles of warfare

From mass to network, from rare platforms to cheap swarms

Aerial, naval, and ground drones have become the "air infantry" of the 21st century: they are produced in tens of thousands of units, operate almost continuously, and compress the "discover–strike–assess" cycle from tens of minutes to a few dozen seconds.

Recent conflicts show that large, slow, and expensive platforms – such as Bayraktar TB2 drones – quickly become vulnerable in a saturated environment of air defense and electronic warfare; instead, small locally adapted FPV UAVs are hard to neutralize and decisively more cost-effective.

Sensory persistence and battlefield transparency

The integration of sensors (optical, thermal, radar, acoustic, commercial satellites) produces an almost continuous operational picture – "seeing everything, everywhere, all the time" – drastically reducing maneuvering space and surprise for classic forces.

Civilian devices – phones, applications, social networks – are exploited for localization, identification, and influence, transforming the informational and digital space into a front as important as the physical contact line.

Remote warfare and reducing human risk

Unmanned systems (UAV, USV, UGV) move people back in the command chain, but take them out of the front line: observation, fire correction, precision strikes, mining and demining, and evacuating the wounded can be done with robotic platforms.

The structural trend is "risk transfer": powerful military actors seek to strike the adversary deep into territory – energy, logistics, industrial infrastructure – through drones and long-range missiles, reducing direct contact and the political cost of human losses.

AI, decision speed, and "compression of the kill chain"

Artificial intelligence is already used for filtering data from sensors, identifying targets, coordinating drones, and suggesting strike prioritization, reducing the time from detection to strike to under a minute in some units.

As algorithms take over more classification and recommendation functions, the risk is that human oversight becomes merely formal, amplifying issues of accountability, errors, and unintended escalation.

Expanding domains: cyber, space, informational

Cyber operations against critical infrastructures, communications, transportation, and navigation (e.g., GPS interference) are already a standard element of competition between states, including below the threshold of armed conflict.

Military dependence on satellites and commercial space networks transforms low Earth orbit into a contested space, vulnerable to jamming, spoofing, or attacks on the constellations upon which both armed forces and the civilian economy depend.

Blurring the civil-military boundary and "total defense"

The massive use of civilian infrastructure and technology (cloud, IT platforms, commercial satellites, telecommunications) in military support makes these "dual-use" items perceived as military targets, increasing the risk to the population.

Models of "total defense" – in which civil society, tech companies, and citizens are explicitly integrated into the defense architecture – amplify resilience but also vulnerability, complicating the application of the principle of distinction in humanitarian law.

How wars are already being fought: key lessons from drones, cyber, and AI

Drones as the central vector of current conflicts

In Ukraine, both sides are using millions of small drones for reconnaissance, artillery fire correction, FPV strikes, logistical interdiction, and deep attacks on critical infrastructure and air bases.

Ukraine has partially compensated for naval inferiority with kamikaze USVs and long-range aerial drones, striking Russian ships and ports hundreds of kilometers away, including in the Black Sea and the Sea of Azov.

The (relative) death of large platforms and the "mass vs. defense" era

Large, expensive drones with high radar profiles and vulnerable radio links have been quickly eliminated in a saturated environment of air defense and electronic warfare; what survives are small, cheap, locally produced, and expendable platforms.

The cost-effectiveness ratio is disruptive: an FPV UAV costing a few hundred dollars can destroy a tank or armored vehicle worth millions, forcing the adversary to consume very expensive interceptors and air defense systems.

Counter-drone: the invisible war in the electromagnetic spectrum

Jamming, spoofing, emitter detection, and striking command centers have become as important as classic air defense, and both sides are investing in fiber optics for drones, resilient links, and autonomous modes to bypass jamming.

C-UAS systems are emerging: multi-spectral sensors, mobile jammers, directed energy weapons, interceptor drones, and proximity munitions, but the fundamental pressure remains economic – countermeasures must be cheaper than the targets.

AI as a force multiplier for ISR, targeting, and communication

In Ukraine and other theaters, AI is used for automatic analysis of drone imagery, pattern recognition (e.g., logistics, artillery, command points), and prioritizing targets for artillery and long-range strikes.

The trend is towards "human-machine teaming": humans operating platforms with a high degree of autonomy, including concepts like "loyal wingman" in aviation or swarms coordinated by algorithms, massively increasing the volume of targets a force can engage simultaneously.

Cyber and space: permanent fronts, not auxiliary

Electronic warfare and cyber operations reduce the accuracy of guided weapons, degrade military and civilian communications, and can temporarily paralyze adversarial logistics without a single shot fired.

Commercial satellite connectivity (LEO constellations) has become critical for command and control, and attacks (including non-kinetic) on it can have systemic effects on states dependent on digital networks.

Humanitarian consequences: more exposed civilians, harder-to-enforce rules

The use of civilian digital infrastructure for military purposes, as well as the "total" involvement of society in the defense effort, makes the distinction between combatant and civilian increasingly difficult to operate in practice.

The acceleration of decision-making through AI and large data volumes risks reducing the space for human judgment and proportionality assessment, which puts additional pressure on compliance with international humanitarian law.

SAFE and Romania's options: what we buy, what we build, what we risk

The SAFE program puts on the table up to 150 billion euros at the European level, in the form of favorable loans, for the acquisition of capabilities considered priorities: munitions and missiles, artillery, land combat capabilities, small drones and anti-drone systems, critical infrastructure protection, cyber, military mobility. Romania has approved an allocation of about 16.6 billion euros – the second largest in the EU – based on the national investment plan in defense.

What Romania wants and can finance through SAFE

According to government and MoD documents, eligible projects include:

ground air defense systems and surveillance radars;

armored transport vehicles, combat vehicles, and warships;

drones and UAV systems, plus munitions and missiles of various calibers;

investments in military infrastructure and the development of the national defense industry.

SAFE acts as a multiplier: loans are granted long-term and at low cost, but require national co-financing, coherent plans, and ideally, projects in cooperation with other member states, with design and production on EU territory.

Strategic risks

Capture by the old industrial model: if the money goes predominantly into large, few, and expensive platforms – heavy tanks, large ships, complex systems with complicated logistics cycles – Romania will remain vulnerable in a war of volume, drones, and rapid production.

Technological and financial dependence: SAFE are loans, not grants; a flawed procurement structure can burden the budget and anchor the army in proprietary solutions that are hard to maintain, without a real transfer of know-how to the local industry.

Doctrinal decoupling: modernizing equipment without changing doctrine, organization, and training (e.g., integrating drones–artillery–infantry–cyber) risks producing an army that is "new on paper, old in practice".

What Romania has to learn from the transformation of warfare

1. To think "industrial" in terms of volume, not just capability

The lesson from Ukraine is that "mass" – thousands of drones, guided munitions, sensors, C-UAS systems – matters just as much as complex platforms, and the innovation cycle must be on the order of weeks, not years.

Romania should use SAFE to stimulate domestic or regional production of small drones (reconnaissance, FPV, logistics), munitions, electronic components, and C-UAS systems, integrating the local industry into European value chains, not just as a low-tier subcontractor.

Practical implication:

A significant portion of the 16.6 billion euros should be directed towards: assembly lines for Class I UAVs, dual-use (civil-military) R&D centers, production of smart munitions and modular, rapidly replicable EW solutions.

2. To prioritize a complete drone–anti-drone ecosystem

SAFE explicitly mentions funding for small drones and anti-drone systems; for a state on the eastern flank, with varied terrain and vulnerable critical infrastructure, this should be a pillar of the national plan.

Romania needs to build a functional "triangle": production and acquisition of UAVs (reconnaissance, strike, logistics), development of multimodal C-UAS (jamming, kinetic, sensors), and digital integration with artillery and maneuver troops.

Practical implication:

equipping each brigade with organic UAS units (aerial, ground), with its own logistics chain and front maintenance capacity;

a national network of training centers for drone operators and EW, including programs for reservists with a technological profile.

3. To invest in a data-based ISR and command–control architecture

Without a sensor infrastructure, resilient communications, storage, and analysis (including AI) of data, drones become expensive gadgets; their value comes from integration into a system that can quickly detect, identify, and strike targets.

Romania needs to connect SAFE investments in sensors, radars, UAVs, and C-UAS with existing digital infrastructure and with defensive and offensive cyber capabilities, ensuring redundancy and protection for critical networks.

Practical implication:

modernizing national C2 and NATO interoperable (including for brigades on the eastern flank) with software adapted to massive UAS management;

data fusion centers with explainable AI and clear human oversight mechanisms, to maintain control over lethal decision-making.

4. To prepare society and the legal framework for "total defense" digital

From the ICRC's perspective, the increasing integration of civilian infrastructure and non-state actors in conflict complicates the application of humanitarian law and increases risks for civilians; Romania, as a democratic state, cannot afford to ignore these dilemmas.

A realistic concept of national defense must include clarifying the role of IT companies, telecom operators, and critical infrastructures in crisis situations, with legal guarantees for civilian protection and respect for the principle of distinction.

Practical implication:

updating the regulatory framework for cyber-defense and mobilizing civilian digital infrastructure for military purposes, with clear legality filters;

public education programs on informational resilience, data protection, and behavior in conflict scenarios (cyber attacks, disinformation, service interruptions).

5. To link SAFE investments to long-term political and industrial objectives

SAFE is temporary (until the end of the decade), but the debt and the configuration of the defense industry will remain long after; Romania's plan must be coherent with the national security strategy and the industrial vision until 2040.

With one of the largest allocations, Romania can become a regional hub for certain niches – for example, munitions, components for drones, C-UAS systems, or maintenance – if it conditions SAFE projects on technology transfer and consistent involvement of the local industry.

Practical implication:

clearly defining 3–5 "specializations" in which Romania wants to be a net supplier in the EU and anchoring SAFE projects in these directions;

correlating the SAFE plan with NATO programs (including on the eastern flank) and with other European instruments (EDIRPA, EDIP, ASAP), to avoid overlaps and maximize interoperability.

Analytical conclusion

War is no longer just a confrontation of brigades, but a conflict between technological ecosystems: whoever can produce, adapt, and integrate drones, sensors, AI, cyber capabilities, and countermeasures faster will dominate the tactical and strategic field. For Romania, SAFE is not just a military "credit card", but a test of strategic lucidity: if the money will be transformed into smart volume, combat digital infrastructure, competitive local industry, and adapted doctrinal frameworks, Bucharest can move from being a "security consumer" to a credible provider on the eastern flank; if not, the risk is to buy weapons for yesterday's wars, leaving future generations with a large debt and an unprepared army for tomorrow's wars.

Analysis conducted with the support of Perplexity