Critical Infrastructure: The New Battlefield

From energy facilities in Europe to airports and data centers around the world, critical infrastructure is increasingly finding itself in the crosshairs of modern warfare. As artificial intelligence makes cyberattacks faster and cheaper, and drones become more capable and harder to detect, governments face a growing challenge: how to protect the systems societies rely on every day.

In Germany, that threat has become increasingly tangible.

In September, German police found improvised explosive devices near a substation serving the Jänschwalde power plant in Brandenburg, with several electricity lines damaged and service briefly interrupted.

And just weeks earlier, an explosive-laden drone was discovered near a Ukrainian cargo aircraft at Leipzig/Halle Airport. The drone was equipped with explosives and a detonator, triggering a major security operation and the diversion of flights. German authorities later described the incident as a hybrid attack.

The threats are not limited to Europe.

During the war between the United States, Israel and Iran, data centers in the Gulf were also hit. DW reported that attacks on facilities in Bahrain and the United Arab Emirates caused disruption to banking, payment platforms and other digital services, highlighting the growing strategic importance — and vulnerability — of data centers.

Data centers have become a critical part of modern society, supporting everything from financial services and healthcare to communications and artificial intelligence. As more of the global economy depends on digital infrastructure, attacking the physical systems behind it can have consequences far beyond the facility itself.

For Eran Hoffman, VP of R&D at Israeli cybersecurity company DREAM, the scale and sophistication of cyberattacks are increasing globally.

“We can see the attacker scale happening all over the world, not only in Europe. Scale has increased and the sophistication of the attacks got really high.”

Artificial intelligence is adding another dimension to that threat.

Hoffman says AI has lowered the barrier to launching sophisticated attacks, while also reducing their cost.

“Today in the AI era the barrier of creating sophisticated attacks is much lower than in the past... every day it's getting cheaper and smarter and the attackers can leverage it.”

That means attackers no longer necessarily need significant resources to cause disruption. According to Hoffman, vulnerabilities can come from sophisticated software exploits, but also from much simpler problems such as human error and weak passwords.

The motivation can vary too. Attacks may be carried out for financial gain, while states or other actors may seek to cause disruption or damage critical services such as water and electricity.

Fighting AI with AI

DREAM is developing an AI-based cybersecurity platform designed to detect, investigate and respond to threats across different layers of an organization's infrastructure.

Hoffman describes the system as providing a “360” view by combining external information, internal data and real-time signals. The goal is to identify small signs of suspicious activity that might take a human much longer to detect.

Once a threat is identified, speed becomes critical.

“Every second counts and you need to respond with minimum impact... AI solution can happen really fast, even when you have a human in the loop that makes the final decision.”

Hoffman argues that organizations need to keep pace with the technology being used by attackers.

His warning is simple: defenders should not “come with a knife to a gun fight.”

And the same race is taking place in the physical world.

The drone threat from above

Affordable drones are becoming increasingly available, with capabilities ranging from consumer-grade systems to sophisticated platforms that can fly at low altitude and operate without easily detectable signals.

Avshalom Amossi, CRO of Ondas Defense, says the challenge is already being seen by customers across Europe and elsewhere.

“It's a complex problem... there's availability of affordable drones, varies in capabilities, some are... quite hard to detect.”

Ondas' approach is to combine different types of sensors and use AI to determine what a detected drone actually represents.

Is it a police drone? A commercial aircraft? A harmless hobbyist? Or a potential terrorist threat?

The system is designed to help determine not only what is flying, but where it is coming from, how serious the threat is and what response is appropriate.

That response can take different forms, including jamming, spoofing, taking control of a drone or physically intercepting it.

One of Ondas' systems, Iron Drone Raider, is designed to provide that last layer of protection. The autonomous interceptor can launch from a container, lock onto an incoming drone and deploy a net to bring it down, a method designed to minimize damage in sensitive environments such as airports.

The company's systems can also combine different responses when dealing with multiple drones or more complex attacks.

The objective is not simply to detect a drone, but to create a complete chain: detect, classify, decide and intercept.

From protecting sites to protecting countries

For now, much of the focus is on individual critical sites such as airports and other sensitive facilities.

But Amossi sees the technology eventually expanding beyond individual locations.

He projects that by around 2028, these capabilities could reach the state level, allowing countries to exercise greater control over what flies at low altitude across much larger areas.

That would represent a significant shift — from protecting a single airport or power facility to creating a broader low-altitude security network.

The challenge is particularly complicated because the same airspace can contain commercial drones, police aircraft, helicopters, hobbyist drones and potentially hostile systems.

And the response cannot always be the same.

Jamming, for example, can create problems around civilian aviation. Ondas therefore also works with alternatives that can take control of certain drones without relying on jamming or spoofing.

A race against an evolving threat

Whether the threat comes through a computer network or from the sky, the underlying problem is similar: attackers are becoming faster and more adaptable.

For cyber defenders, AI can help analyze enormous amounts of information and respond in seconds. For physical security, AI can help turn multiple sensor inputs into a decision about what to do with an incoming drone.

But the technology is constantly evolving on both sides.

Hoffman describes cybersecurity as a race over speed, warning that AI models are becoming “smarter, faster, cheaper” and that the scale of the threat is growing.

Amossi makes a similar point about drones: the technology is available, but systems have to continuously learn from new threats and previous interceptions. Ondas says it incorporates experience from deployments across numerous countries into that learning cycle.

For both companies, the message is that critical infrastructure can no longer be protected by a single layer of security.

The battlefield is increasingly digital and physical at the same time — and the infrastructure at the center of modern life is becoming a strategic target.

As Amossi puts it, the challenge is complex, but the technology to address many of the threats already exists.

The question now is how quickly governments and infrastructure operators will deploy it.