The mobilization was swift and symbolic. Last spring, 3,000 soldiers from the British Army’s 4th Light Brigade, historically known as the Black Rats, deployed across multiple logistical vectors—air, sea, rail, and road—from Yorkshire into the densely wooded eastern territories of Estonia. This contingent joined a larger formation of 14,000 NATO troops for Exercise Hedgehog, a massive simulation designed to stress-test the alliance’s readiness against a high-intensity Russian incursion. While heavy armor, including 69-ton Challenger tanks, Apache gunships, and advanced supersonic missile systems, dominated the physical landscape, the true tactical innovation lay in an unseen digital architecture.
British military strategists heralded the 4th Brigade’s primary advantage not as a kinetic weapon, but as an invisible, self-optimizing intelligence network. This system, dubbed Project ASGARD (Autonomous Sensor-to-Gun Adaptive Response Dynamics), represented a radical break from conventional command structures. Developed with astonishing velocity—a mere four months, contrasting sharply with the customary multi-year cycles of defense procurement—ASGARD’s core function is to fuse disparate battlefield assets, linking "sensors" (reconnaissance assets) and "shooters" (weapon platforms) into a single, shared, wireless electronic brain.
In traditional military operations, the discovery of an enemy asset, such as a tank concealed in a copse, triggers a slow, hierarchical process. A drone operator must relay the information up a centralized chain of command, where officers collectively analyze, verify, and decide on a course of action. This human-centric loop often takes minutes—a fatal delay in modern, high-speed warfare.
ASGARD, however, functions on a decentralized model, often analogized to an octopus’s nervous system, where individual extremities retain autonomy while adhering to overarching strategic objectives. During the Hedgehog exercise, AI-enabled object recognition systems aboard orbiting reconnaissance drones instantly identified targets. Instead of routing the data through a lengthy hierarchy, the image and precise coordinates were transmitted directly to nearby kinetic assets—be it a towed artillery piece, an allied main battle tank, or a loitering munition drone waiting on its launch catapult. The human operators interfacing with this digital targeting web used ruggedized commercial smartphones. The interface presented them with pre-calculated strike options, ranked by metrics like $pKill$ (probability of kill). A simple tap on the screen initiated the weapon’s irreversible, high-speed flight path toward the target.
This brutal new calculus of speed is Europe’s answer to geopolitical pressure. Sven Weizenegger, head of the German military’s Cyber Innovation Hub, notes the proximity of the threat: "The Russians are knocking on the door." Policymakers across the continent are betting heavily on automated battlefield technology to ensure that door remains shut. Estonia, a small frontline state, views this technological edge as critical. Angelica Tikk, head of Innovation at the Estonian Ministry of Defense, emphasized that AI-enabled intelligence, surveillance, reconnaissance (ISR), and the deployment of massive drone fleets are now decisive factors, allowing small nations to "punch above our weight."
The Industrialization of Lethality
The scope of this pivot is staggering. Following the operational lessons learned in Ukraine, where drone production surged from 2.2 million units in 2024 to an estimated 4.5 million in 2025, European defense planners are preparing for conflict at scale. Andrius Kubilius, the EU defense and space commissioner, estimated that defending a small state like Lithuania (population 2.9 million) against a large-scale aggression would necessitate the deployment of three million drones annually.
Projects like ASGARD are designed not only to handle these quantities but to leverage them for maximum velocity. British officials project that the AI-powered targeting web can collapse the entire kill chain—from detection to strike authorization—to under a minute. This efficiency, according to a British Army press release, is expected to make the force "10 times more lethal over the next 10 years." Following ASGARD’s anticipated completion by 2027, Germany plans to roll out its own equivalent, Uranos KI, as early as 2026.
The underlying strategic theory is that this combination of mass-produced, rapidly deployed, and algorithmically guided lethal drones will create an overwhelming asymmetric advantage, serving as "brutal, guns-and-steel, feel-it-in-your-gut deterrence," as described by Eric Slesinger, a Madrid-based venture capitalist focused on defense startups.
This rapid rearmament effort is driving a profound reshaping of the European defense industrial base, a shift spearheaded by a new class of technology-first defense firms. Twenty companies contributed to Project ASGARD, ranging from established giants like General Dynamics to agile, venture-backed startups. None, however, embody the current European military zeitgeist more completely than Helsing.
The Rise of the New Defense Titans
Founded in Munich in 2021 by a unique combination of expertise—a theoretical physicist, a former McKinsey partner, and a biologist turned video-game developer—Helsing quickly attracted significant investment, including an early €100 million commitment from Spotify CEO Daniel Ek. The company has aggressively positioned itself at the apex of Europe’s defense tech ecosystem, achieving a $12 billion valuation last June, making it the continent’s most valuable defense startup.
Helsing’s emergence coincided with a monumental policy shift. In March 2025, the European Commission called for a "once-in-a-generation surge in European defence investment," targeting AI and drones as priority areas within a new framework that could unlock nearly a trillion dollars for defense programs. Germany alone has earmarked $12 billion for drone procurement.
Crucially, Helsing rejects the traditional European model where governments dictate technical specifications through rigid, decades-long contracting processes. Instead, they operate with a "traditional tech-startup muscle," according to Chief Scientist Antoine Bordes, previously an AI research leader at Meta. "You raise money, you create technology using this money that you raised, and then you go to market with that." This agile, market-driven approach is gaining traction among European officials seeking instruments for rapid procurement.
Helsing’s portfolio spans all operational domains. In the orbital layer, they are collaborating with Loft Orbital to establish a constellation of reconnaissance satellites capable of global military asset detection and classification. In the air, they offer the HF-1 and HX-2 loitering munitions—drones that combine persistent ISR capability with explosive payloads. The company has already disclosed orders for 10,000 airframes for Ukraine, with reports indicating dozens of successful combat missions. Looking forward, the company unveiled the Europa, a four-and-a-half-ton uncrewed autonomous fighter jet slated for mass production. This sleek aircraft, resembling an upturned boning knife in concept art, is designed to penetrate heavily defended airspace under the command of a human pilot located safely beyond enemy range. Underwater, Helsing envisions drone mini-subs capable of month-long, untended surveillance missions at depths up to 3,000 feet.
The Collective Electronic Brain
The critical linchpin tying this complex ecosystem together is Altra, Helsing’s "recce-strike software platform." Altra served as the collective electronic brain in the ASGARD trials. General Richard Barrons, former commander of the UK Joint Forces Command, champions these AI-driven "kill webs" as the ultimate deterrent. He posits that any potential aggressor, considering an incursion, for instance, into Narva, Estonia, would be forced to conclude that the invading force "will be destroyed the minute it sets foot across the border."
A fully realized targeting web enables saturation attacks—a military strategy using synchronized barrages of missiles, drones, and artillery coordinated across national borders and operational domains to overwhelm adversary defenses. The goal is simple: "lethality that deters effectively," as articulated by Helsing VP Simon Brünjes. This concept mirrors the US Navy’s strategy for defending Taiwan with autonomous swarms designed to create a "hellscape" for invading vessels.
The Diminishing Human Factor
Achieving full saturation capability faces one major hurdle: the human element. Richard Drake, head of the European branch of Anduril (another key defense startup involved in ASGARD), notes that while the ASGARD kill chain is technically capable of full autonomy, current government regulations mandate a "human in the loop" for lethal decisions. Estonia’s Ministry of Defense maintains this stance, insisting on human control over lethal force.
Yet, the technical capability to bypass human oversight is rapidly maturing. Helsing’s drones currently utilize object recognition for target detection, which an operator must review. However, the aircraft achieve "last mile" autonomy during their terminal guidance phase—the final half-mile to the target—operating hands-free. This mode reportedly achieves a hit rate around 75%, according to CSIS research.
The necessity of human input is rapidly being eroded by the sheer volume of assets and the threat of enemy electronic warfare. Advances in jamming technologies, capable of severing communication links between operators and drones, force reliance on internal autonomy. Russia, for its part, is also escalating the autonomy race, upgrading its Lancet strike drones with advanced recognition systems and even testing autonomous nuclear-capable torpedoes.
To overcome the human bottleneck—where "a million drones are great, but you’re going to need a million people"—firms like Helsing and Anduril are developing "one-to-many" systems. Antoine Bordes’s team at Helsing is engineering AI to allow a single human to oversee multiple HX-2 drones simultaneously; Anduril is aiming for a single operator to marshal a fleet of ten or more. While a human is technically still involved, their capacity to intervene on individual lethal decisions within a coordinated swarm attack is drastically diminished, blurring the line between human supervision and automated engagement. UN Special Rapporteur Morris Tidball-Binz has warned that the international community is "crossing a threshold which may be difficult, if not impossible, to reverse later."
For policymakers like Weizenegger of the German Cyber Innovation Hub, the choice is stark: "If you don’t have the people, then you can’t control so many drones. So therefore you need swarming technologies in place—you know, autonomous systems… It’s about winning or losing. There are only these two options. There is no third option."
The Obsolescence Paradox
The push for "Helsing speed" reflects a continent fixated on immediate readiness. Executives emphasize the need to be "ready to fight tonight in the Baltics," a preparedness they feel is currently lacking. This sense of urgency has led European governments, such as Denmark, to prioritize rapid acquisition over optimal quality: "If we can’t get the best equipment, buy the next best. There’s only one thing that counts now, and that is speed."
To meet this demand, Helsing is establishing a network of dispersed "resilience factories" across Europe, designed to churn out drones at a wartime clip. Their first facility in southern Germany aims for 1,000 drones per month, setting the stage for major contracts like the planned €300 million German order for 12,000 HX-2s. Helsing suggests Germany alone requires a stockpile of 200,000 HX-2s to weather the initial two months of an invasion.
However, the efficacy of massed, AI-driven drones remains an open question. While drones are responsible for 70% to 80% of combat casualties in Ukraine, they have often led to a grinding, attritional conflict—a "Somme in the sky," according to military observers. This environment fosters rapid innovation, but also rapid obsolescence. As former Ukrainian government advisor Kateryna Bondar notes, drones produced today may be technologically obsolete in six months, making large stockpiles potentially futile.
Furthermore, industry claims about AI efficacy face battlefield realities. Bohdan Sas, founder of Buntar Aerospace, points out that Western companies often tout "super-fancy recognition" achieved in pristine testing environments—open fields with clear targets—which collapses in the messy reality of war, where targets are expertly hidden. Ukrainian forces have reportedly disabled the autonomous functions of Russian Lancet loitering munitions due to real-world operational failures. The reliability of operating 100,000 autonomous drones in a contested, jammed environment remains highly suspect.
The Historical Echoes of Escalation
The current trajectory, fueled by technological optimism and strategic fear, carries profound ethical and historical weight. The use of new air power technologies was perfected during the Spanish Civil War, notably in towns like Corbera, where German and Italian forces used civilian centers as laboratories for destruction. Today, Ukraine serves a similar, grim function, providing Western defense firms with invaluable, real-world combat data—sometimes for a fee—that cannot be replicated on a test range.
The human cost of drone warfare is immense; UN human rights commissions have documented how Russian drone usage is designed to spread terror among civilians, constituting a crime against humanity. A widespread drone war on Europe’s eastern border, where tens of millions reside within strike range, threatens tragedy on an unprecedented scale.
The shift toward autonomous systems raises the risk of unchecked escalation. If deterrence fails and a major conflict erupts, the ethical standards regarding human control are likely to change instantly, as suggested by Helsing executives themselves when contemplating a "full-scale war with China or Russia." General Barrons suggests that NATO’s targeting web would be used not just defensively, but immediately for deep retaliatory strikes inside Russian territory. This mentality, according to Richard Moyes of Article 36, a nonprofit focused on civilian protection, is one where military planners are "not imagining off-ramps."
The professional distance maintained by defense tech firms from the visceral realities of conflict concerns military veterans. General Barrons cautioned against the mindset prevalent in the technology sector: "I think on planet Helsing and Anduril, they’re not really fighting, in many respects. And it’s a different mindset." While technology firms argue that ethically built autonomous systems can limit noncombatant casualties more effectively than previous weapons, the ultimate certainty remains simple: lethality means death. The only question is how quickly, and how massively, these algorithmic kill chains will bring that grim outcome to bear.