Stark Defence's Virtus loitering munitions -- marketed as autonomous strike drones capable of engaging targets up to 100 kilometers away with a 5-kilogram warhead -- failed to hit a single target across four separate strike attempts during October 2025 live-fire trials with British forces in Kenya and German forces near Munster. During the German trial, one Virtus drone missed its target by more than 150 meters before crashing into woodland. In Kenya, a drone's battery caught fire upon impact after a failed attack run. The complete operational failure occurred in controlled demonstration conditions, not chaotic battlefield environments.
The failure is significant because of who funded the company and how fast it scaled. Stark Defence was founded in 2024 by Florian Seibel, who previously co-founded Quantum Systems but reportedly left because investors would not allow him to equip surveillance drones with weapons. In just 15 months, Stark raised $100 million total: Sequoia Capital led a $15 million round in October 2024 (valuing the company at EUR 90 million), then led a $62 million round in August 2025 that quintupled the valuation to $500 million. Investors include Peter Thiel's Thiel Capital, In-Q-Tel (the CIA's venture capital arm), the NATO Innovation Fund, 8VC, Project A, and Dopfner Capital. The company expanded to a 40,000-square-foot factory in Swindon, UK, and opened offices in Munich and Kyiv, plus acquired Berlin-based autonomous flight control startup Pleno.
Despite the total trial failure, Germany moved to award Stark approximately EUR 300 million ($317 million) as part of a EUR 900 million ($950 million) kamikaze drone contract split with Helsing and Rheinmetall, intended to supply up to 12,000 drones for a German brigade in Lithuania defending NATO's eastern flank. By February 2026, Germany formally approved a EUR 536 million Helsing-Stark drone deal, though the long-term framework was cut from EUR 4.3 billion to EUR 2 billion. Stark's response to the trial failures was unapologetic: 'We have crashed not once or twice, but hundreds of times. This is how we test, develop, and ultimately continue to deliver defense technologies like Virtus to the front lines in Ukraine.'
The pattern is a textbook case of venture capital momentum overriding demonstrated capability. A company valued at $500 million on the basis of investor prestige (Thiel, Sequoia, CIA, NATO) fails every single weapons test, then wins a EUR 300 million government contract anyway. The investor roster functioned less as validation of technology and more as a political credential -- a who's-who of transatlantic defense-tech patronage that made the company too connected to fail.