Drones

The Regulatory Inflection

Drone regulation is following the smartphone adoption pattern with a 5-7 year lag. Phase one was restriction - blanket bans on commercial operations, mandatory visual line-of-sight requirements, airspace exclusion zones. Phase two, now underway, is framework development - the FAA's Part 107 rules, the EU's drone regulation framework (delegated/implementing regulations under EU 2019/947), and ICAO's Unmanned Aircraft Systems Traffic Management (UTM) standards. Phase three, which China has already entered and the West will reach toward the end of the decade, is normalisation - drones as routine infrastructure, integrated into airspace management systems, insured and regulated like any other commercial vehicle.

The key regulatory catalyst is BVLOS (beyond visual line of sight) approval. Current US rules require a human operator or visual observer to maintain visual contact with the drone, which limits commercial operations to roughly a 1-mile radius and makes delivery, pipeline inspection and agricultural spraying economically marginal. The FAA published its Part 108 BVLOS NPRM in August 2025; a June 2025 executive order set a 1 February 2026 deadline for a final rule, but the agency blew past it - reopening the record in January 2026 on detect-and-avoid and the contested drone-vs-crewed right-of-way question, leaving the rule unpublished through mid-2026. This single change is estimated to unlock tens of billions in annual US economic value, and Zipline, Wing, Amazon Prime Air and DroneUp have each built infrastructure predicated on it.

The European framework is more advanced in structure but slower in practice. The EU's U-Space regulation establishes a digital traffic management system for drones, but implementation varies dramatically by member state - Switzerland and Estonia are operational while France and Germany are still building the infrastructure. China's build-first, regulate-retroactively approach has given it a multi-year operational head start. The regulatory gap between China and the West is now a competitive gap, and it is widening.

The Cost Asymmetry Engine

Autonomous drones are not an application of existing technology - they are the mechanism through which artificial intelligence and autonomous systems enter the physical world at scale. The economics are so lopsided they restructure entire systems of force projection, logistics and infrastructure management. A $500 first-person-view drone can disable a $10 million main battle tank. A $2,000 delivery drone can replace a $50,000 annual truck route. A $5,000 agricultural sprayer can cover ground that previously required a $300,000 piloted aircraft. These are not marginal efficiency gains - they are order-of-magnitude cost asymmetries that collapse the economic logic of incumbent systems.

The military-commercial feedback loop is the accelerant. Ukraine's war proved the doctrine: commodity FPV drones, adapted with 3D-printed tail fins and improvised warheads, became the dominant anti-armour weapon on a modern battlefield within 18 months, and Ukraine now claims a manufacturing base above 8 million units a year. This forced every NATO defence ministry to fund counter-drone systems, autonomous swarm research and AI-guided navigation. That defence R&D spending directly funds autonomy capabilities that transfer to civilian applications. Better obstacle avoidance developed for military drones becomes the sense-and-avoid system for delivery drones. Swarm coordination algorithms designed for battlefield ISR become the fleet management layer for agricultural spraying. The military proves the technology works under the hardest conditions; the commercial sector scales it under the easiest.

The regulatory framework is following, not leading. The FAA's Part 108 BVLOS (beyond visual line of sight) rule, proposed in August 2025, would unlock the US commercial drone economy the way the Telecommunications Act of 1996 unlocked broadband - but it blew past its 1 February 2026 deadline and remains unpublished, stalled on the drone-vs-crewed right-of-way question. The EU's U-Space framework is already operational in several member states. China, characteristically, never waited - DJI Agras drones spray over 100 million acres of Chinese farmland annually, and drone delivery is routine in Shenzhen. The pattern is identical to smartphones: initial restriction, then framework development, then normalisation. We are at the framework stage in the West and the normalisation stage in China. The gap matters.

The Supply Chain Vulnerability

DJI controls roughly 70% of the global civilian drone market and an estimated 80%+ of the US consumer segment. This is not a market share - it is a dependency. DJI drones are used by police departments, fire services, infrastructure inspectors, agricultural operators and filmmakers worldwide. The company's vertically integrated supply chain - from the flight-controller silicon to the gimbal stabilisation software - gives it cost advantages that Western competitors cannot match at current production volumes.

US restrictions have now escalated from the 2019 Entity List placement to the December 2025 FCC action that swept DJI and effectively all foreign-produced UAS onto the Covered List, blocking new equipment authorisations. DJI is contesting that action in the Ninth Circuit, but the practical effect is an expensive, necessary diversification of the Western drone supply chain. The problem is that alternatives remain behind on cost. Skydio, the leading US autonomous-drone maker, announced a $3.5 billion US manufacturing expansion in April 2026 to scale domestic capacity, but still ships at volumes far below DJI's millions of units. Autel Robotics, Chinese-owned but with US operations, occupies an ambiguous middle ground. European maker Parrot pivoted to defence and government and exited consumer. The component dependency runs deeper than final assembly: DJI's proprietary flight controllers, its O-series video transmission and its obstacle-avoidance stack are embedded in the workflows of hundreds of thousands of commercial operators. Switching costs are real and high.

The strategic question is whether the West can build a competitive drone industrial base before the next conflict requires it. Current trajectories suggest a multi-year gap before Western manufacturers reach cost and capability parity with DJI in commercial segments. In defence the picture is better - Anduril, Shield AI, L3Harris and AeroVironment produce capable tactical systems - but at price points several times higher than their Chinese equivalents. The drone supply-chain problem mirrors the semiconductor one: a critical technology concentrated in a geopolitically exposed location, with no quick fix.

State of the art (2026)

The drone story in 2026 is bifurcated. In war, Ukraine's distributed industrial base can now build over 8 million FPV drones a year, with fibre-optic variants immune to jamming running at 30,000-50,000 units a month and interceptor drones downing the bulk of Russian Shaheds over Kyiv. In the West, the Pentagon's Replicator announced its first Replicator 2 buy in January 2026 but has fielded only hundreds, not the promised thousands. On supply chains, the December 2025 NDAA trigger lapsed and the FCC swept new DJI and foreign-made drones onto its Covered List, freezing fresh imports. The commercial unlock, the FAA's Part 108 BVLOS rule, blew past its February 2026 deadline and remains unpublished, stalled on right-of-way.

Ukraine as the world's first wartime drone manufacturing ecosystem

Ukraine has built something unprecedented: a wartime drone industrial base from scratch in under four years. Before 2022 it had no meaningful drone manufacturing. By 2025 it received roughly 3 million FPV drones and claims capacity above 8 million a year, produced across roughly 500 mostly small manufacturers. The government's Army of Drones programme, run by the Ministry of Digital Transformation, has created a procurement-to-frontline pipeline that iterates faster than any peacetime defence acquisition system.

This matters beyond the war itself. Ukraine is developing doctrine, manufacturing processes and operator training at a pace no Western military can match because no Western military is at war. The feedback loop is measured in days, not years: an operator identifies a need, a manufacturer builds a prototype, it is tested in combat, and the design is iterated the following week. This compressed cycle has produced innovations — fibre-optic drones immune to jamming (now running tens of thousands a month), interceptor drones against Shaheds, AI-assisted targeting — that would take Western primes years to field through traditional procurement.

The investment thesis is that Ukraine's drone companies (Brave1-backed startups, Escadrone, Kvertus and others) will become post-war exporters. Countries seeking proven combat systems will prefer suppliers that have iterated in the most demanding operational environment in history. Ukraine could become the Israel of drones — a small country whose combat-proven military technology exports become a major economic driver.