Introduction
In May 2026, Helsing, a Munich-based AI defence company, announced plans to raise funding to an estimated $18 billion with backing from tech billionaire Daniel Ek. This funding solidified its status as one of Europe’s most valuable defence startups. However, just months earlier in November 2025, an internal German Defence Ministry presentation revealed that Helsing’s flagship drone, the HX-2, had a mere 25% launch success rate during frontline testing in Ukraine (Kyriasoglou et al., 2026). Helsing’s ability to secure multi-billion-dollar investment despite its drones’ battlefield failures points to a structural disconnect between Europe’s defense procurement policies and functional equipment.
From Software to Battlefield
Helsing was initially founded as an artificial intelligence software company in Munich in 2021. The company focused on AI systems that collect and process data from battlefield sensors and weapons to provide real-time tactical insights to military personnel. Helsing’s first drone, the HF-1, was developed with Ukrainian manufacturers, followed by the HX-2, an AI-guided loitering munition with an approximate range of 100 kilometers. This model was promoted as GPS-free and resistant to electronic warfare jamming. By May 2026, Helsing was Europe’s most valuable defence tech startup, valued at $18 billion, and was seen as the European counterpart to US firms like Anduril, known for advanced autonomous weapons systems. A significant aspect of this positioning is its ownership structure: the company remains over 80% European-owned despite its latest funding round being led by US investors. Helsing highlights that this ownership is in alignment with Europe’s broader push for defence industrial sovereignty under the European Defence Industrial Strategy (Levingston et al., 2026)
The invasion of Ukraine has accelerated demand for battle-ready aerial technology, transforming the conflict into the most demanding drone warfare environment in history. Unlike controlled trials, Ukraine subjects systems to advanced and real-time battlefield conditions. These include a series of evolving Russian electronic warfare strategies (Slusher, 2025). The enemy in these conditions is real, not simulated, and combat conditions are unforgiving. For Helsing, whose HX-2 was designed to operate in GPS-denied and electronically contested areas, Ukraine serves as the ultimate test.
Battlefield Reality vs. Marketed Capability
In 2024, Helsing signed a contract to provide Ukraine with 4,000 strike drones of its first model, the HF-1. However, according to an internal presentation prepared by Germany’s Defence Ministry in November 2025, 40% of HF-1 shipments remained unused in Ukrainian inventory. According to a publication by Bloomberg, the report further noted that the drone had drawn widespread criticism from Ukrainian forces, who deemed it “too expensive and ineffective” (Kyriasoglou et al., 2026). To address these shortcomings, the contract was amended to supply Helsing’s newer flagship model, the HX-2, although this presented new problems (Helsing, 2024).
As part of its recent drive toward military modernisation, the German Bundeswehr planned a purchase worth €267.7 million from Helsing. Internal documents from the German Defence Ministry, however, told a different story. They revealed that the HX-2 model had little success on the battlefield. A separate Politico investigation found that the HX-2 recorded a mission success rate of just 36%, hitting its target only five times out of 14 in the Donbas region (Lunday & Petersen, 2026). The same internal presentation further documented that during frontline testing by Ukraine’s 14th Regiment, only 25% of HX-2 drones were able to successfully launch. Additionally, both the HF-1 and HX-2 had been marketed to include three AI components: terminal guidance, midcourse guidance and visual target acquisition. Yet the internal document revealed that delivered units lacked all three capabilities (Kyriasoglou et al., 2026). These vulnerabilities were readily exploited by Russian electronic warfare capabilities, which successfully disrupted communication links between operators and the remaining drones in the field. As a result, Ukraine halted further orders from Helsing in January 2026. The promise of a European-led, AI-driven glimpse into the future of drone warfare was not enough to offset the equipment’s poor battlefield performance.
On the 19th of January 2026, Helsing responded directly to the Bloomberg report, calling it “misleading” and stating that the company was unaware of the internal German Defence Ministry presentation. The company denied the reported takeoff failure rate, asserting that the HX-2 had been successfully tested and approved for frontline use. Helsing further claimed that more than six units within the Ukrainian army had issued concrete requests to order the HX-2, and that one regiment had specifically requested over 1,000 additional drones for combat missions. The company added that the HX-2 had been listed in Ukraine’s official central ordering system, and cited near 100% hit accuracy rates during controlled trials conducted in Germany, the United Kingdom and Kenya, a claim that stands in direct conflict with both the Bloomberg and Politico investigations into its battlefield performance (Helsing, 2026).
The Procurement Gap
Helsing’s failures are a symptom of a larger structural issue in European defence procurement. A parallel example is Berlin-based Stark Defence, which faced a highly publicised trial failure after its Virtus drone failed to hit a single target across four attempts during live-fire exercises with British and German forces in October 2025 (Pitel and Clover, 2025). Despite this, Stark was selected by the Bundeswehr as one of three companies to supply armed autonomous drones, receiving an initial contract worth €300 million. More established defence manufacturers such as Rheinmetall were notably excluded from this process. Stark and Helsing show a worrying pattern: Germany and larger EU defense frameworks are funding unproven startups over established manufacturers with working products.
One explanation lies in the EU’s defence procurement criteria. The European Defence Industry Programme (EDIP) explicitly incentivises member states to procure from European-owned small to medium enterprises and startups (Council of the EU, 2025). Rheinmetall, despite being headquartered in Germany, had only 26% European institutional ownership in 2025, barely more than its 24% North American investor base (Rheinmetall AG, 2025). This ownership profile makes it a structurally weaker fit under EDIP’s European preference rules. Another factor reinforcing this decision-making process is investor interest in artificial intelligence and its military applications. Both Helsing and Stark’s contracts include “innovation clauses” requiring continuously upgraded technology. Marketed as makers of AI-driven drone systems, they appeal to a military that needs to modernize quickly in response to the war in Ukraine (Pitel, 2026). Helsing lost a customer in January but still managed to close an $18 billion investment four months later. This suggests that market and political confidence in these firms is now decoupled from battlefield results.
Europe is at a critical time of geopolitical instability. Recent US intelligence assessments, reported by The Wall Street Journal in August 2026, indicate that Russian President Vladimir Putin could attempt to test NATO’s resolve within the next few years through a limited attack on an allied country, ranging from a cyberattack to a small-scale incursion deliberately calibrated to fall short of triggering a full Article 5 response. Intelligence officials assess that the key targets would likely be the Baltics and Poland, with the window for such a move estimated between late 2026 and 2029 (Seligman and Kubota, 2026). Against these odds, it is imperative that European-procured defense equipment performs to its expected standards. A drone system that cannot withstand jamming or even become airborne risks becoming a key liability for exploitation on the battlefield. If Europe’s defense industrial base is being tested for the kind of provocation these assessments describe, the margin for procurement decisions driven by ownership rules or investor sentiment narrows considerably.
Conclusion
While Helsing cannot be evaluated solely on the unsuccessful deployment of its earlier drone models, the contracts granted to it by the Bundeswehr seem to represent an investment in the company’s future potential rather than its current output. This strategy may seem worthwhile in a time of relative peace; however, with rising tensions on the battlefield in Ukraine, states should invest in companies with battle-ready equipment. This is a standard Helsing has yet to meet, as shown by Ukraine’s decision to halt further orders in January 2026. At a valuation of $18 billion, Helsing is symptomatic of a shift in European defense procurement strategy. The ultimate test of this strategy will come on the battlefield. As tensions rise in Ukraine, with recent U.S. intelligence assessments warning Putin could test NATO’s resolve in the coming years, Europe needs field-tested, battle-ready equipment (Seligman and Kubota, 2026).
Bibliography
Kyriasoglou, C., Rudnitsky, J., & Leonard, J. (2026, January 19). Ukraine holds off on new Helsing drone orders after setbacks. Bloomberg. https://www.bloomberg.com/news/articles/2026-01-19/ukraine-holds-off-on-new-helsing-drone-orders-after-setbacks
Council of the EU (2025) European Defence Industry Programme: Council gives final approval. 8 December. https://www.consilium.europa.eu/en/press/press-releases/2025/12/08/european-defence-industry-programme-council-gives-final-approval/
Helsing (2024) ‘Helsing unveils intelligent strike drone for mass and precision.’ 2 December. https://helsing.ai/newsroom/helsing-unveils-intelligent-strike-drone-for-mass-and-precision
Helsing (2026) ‘Statement from Helsing.’ 19 January. https://helsing.ai/newsroom/statement-from-helsing
Levingston, I., Pfeifer, S., Kirchfeld, A. and Pitel, L. (2026) ‘Helsing nears $1.2bn raise at $18bn valuation’, Financial Times, 11 May. https://www.ft.com/content/0ca32fa3-cf28-4105-b1f9-ec5bc0c9158c?syn-25a6b1a6=1
Lunday, C. and Petersen, L. (2026) ‘Ukrainian battlefield data undercuts Germany’s big drone bet’, Politico Europe, 29 January. https://www.politico.eu/article/ukrainian-battlefield-data-undercuts-germanys-big-drone-bet/
Pitel, L. (2026) ‘Daniel Ek and Peter Thiel-backed start-ups win German military drone contracts’, Financial Times, 10 February. https://www.ft.com/content/eea27ca8-75e4-4d8a-90e7-bbf98d90ece1?syn-25a6b1a6=1
Pitel, L. and Clover, C. (2025) ‘Drone start-up backed by Peter Thiel crashed and burned in armed forces trials’, Financial Times, 31 October. https://www.ft.com/content/ebdfcf20-5b64-4680-8392-d0455257fede?syn-25a6b1a6=1
Rheinmetall AG (2025) Shareholder Structure 2025. Düsseldorf: Rheinmetall AG. https://ir.rheinmetall.com/investor-relations/share/shareholder-structure
Seligman, L. and Kubota, Y. (2026) ‘U.S. Intel Finds Putin Could Test NATO’s Resolve With Limited Incursion’, The Wall Street Journal, 2026 https://www.wsj.com/world/europe/u-s-intel-finds-putin-could-test-natos-resolve-with-limited-incursion-e3b02e2c
Slusher, M. (2025) Lessons from the Ukraine Conflict: Modern Warfare in the Age of Autonomy, Information, and Resilience. Washington, DC: Center for Strategic and International Studies. https://www.csis.org/analysis/lessons-ukraine-conflict-modern-warfare-age-autonomy-information-and-resilience