Helsing’s AI-Piloted Gripen
In a landmark development for European defense autonomy, Munich-based startup Helsing, in collaboration with Saab, successfully conducted test flights on May 28 and June 3, 2025, where its AI agent Centaur autonomously piloted a Saab Gripen E fighter jet. These flights, conducted over the Baltic Sea with a human safety pilot onboard, marked the first time a European AI system directly controlled a modern combat aircraft in live airspace. Centaur managed key air combat tasks—including beyond-visual-range (BVR) target identification, threat prioritization, maneuvering, and evasion—without human intervention during the exercises.
Trained on over 1 million simulated flight hours in just 72 hours, Centaur exemplifies the speed and scale at which AI can outperform traditional military training models (a seasoned human pilot may accumulate ~5,000 flight hours across an entire career). Helsing envisions mixed human-AI cockpits as operationally viable within this decade, revolutionizing how fighter aircraft are designed, piloted, and updated—software-first, agile, and continuously learning.
In comparison, the U.S. Air Force’s “Valkyrie” and “Skyborg” programs, led by companies like Kratos and Shield AI, have also tested AI-piloted drones and crewed-uncrewed teaming, but often on unmanned platforms or in simulation environments. China, through its AVIC and military R&D arms, has showcased AI-controlled air combat simulations and hinted at autonomous J-16 and drone swarm capabilities, though live aircraft autonomy demonstrations remain less transparent. Helsing's real-world test on a crewed, export-grade fighter aircraft—using European AI—positions Europe as a credible third force in the AI arms race, behind the U.S. and China but rapidly closing the gap.
This development has major implications across political sovereignty, defense funding, and industrial economics—signaling that Europe’s defense future may be written in code, not just steel.
Base Case: AI-Augmented Cockpits Become Standard by 2030 (Probability: 60%)
In this most likely scenario, Helsing’s Centaur leads a broader adoption of AI-augmented crewed fighter jets within Europe and among allied export customers. Governments establish regulatory frameworks requiring a human-in-the-loop for lethal decisions, enabling Centaur-like systems to manage maneuvering, target classification, and electronic warfare autonomously—while leaving weapons authorization to the pilot. By 2030, at least five European nations field mixed human-AI squadrons, and Saab’s Gripen becomes the go-to export platform for countries seeking affordable AI-ready fighters. NATO standardizes protocols for manned-unmanned teaming (MUM-T), and Helsing expands into land and naval domains with similar agent-based control systems. The AI arms race intensifies but remains bounded by ethical policy structures and export controls, maintaining a balance between innovation and oversight.
Upside Case: Europe Emerges as AI Defense Leader, Disrupting U.S. Primacy (Probability: 25%)
In the upside scenario, Helsing rapidly scales its technology beyond Saab to other platforms and domains, such as Eurofighter Typhoon upgrades, autonomous ISR drones, and battlefield coordination systems. By leveraging its data simulation advantage—where it trains agents on millions of synthetic combat hours weekly—it achieves an innovation tempo unmatched by legacy U.S. defense contractors. While the U.S. remains dominant in strategic weapons and stealth, Europe becomes the software powerhouse of autonomous tactical warfare, exporting AI agents as secure payloads across air, land, and sea systems. This sparks a new industrial base in European defense-tech, rivaling Silicon Valley in funding and talent magnetism. Ethical AI governance frameworks led by the EU gain international adoption, giving Europe soft-power leverage alongside its rising hard-tech capabilities.
Downside Case: Public Backlash and Policy Gridlock Stall Deployment (Probability: 15%)
In this risk scenario, Helsing’s demo triggers global concern about the militarization of AI and “killer robot” narratives take hold. Domestic opposition in countries like Germany and Sweden pushes policymakers to freeze funding and impose strict constraints on AI use in combat, especially after a high-profile simulated test goes awry or a misidentified target incident occurs in export contexts. Meanwhile, China and the U.S. continue to advance less constrained AI programs, widening the capability gap. Saab's AI-ready Gripen sees stalled export sales due to regulatory bottlenecks, and Helsing pivots toward non-lethal defense software (e.g., logistics AI, signal intelligence), slowing the momentum of autonomous combat aviation. Europe’s AI defense edge erodes as it becomes mired in political caution and inter-EU disagreement over deployment rules.
