In 2025, the global military unmanned aerial vehicle (UAV) landscape experienced transformative shifts driven by operational demands from conflicts in Ukraine, the Middle East, and South Asia. As a defense analyst closely tracking these trends, I aim to provide a comprehensive first‑person overview of foreign military UAV equipment and technology developments during this year. This review covers novel platform types, production and deployment, key tests and upgrades, and critical enabling technologies. Although the focus is on foreign systems, I deliberately draw comparisons with China UAV advancements where relevant to highlight strategic divergences. The rapid evolution of collaborative combat aircraft, long‑endurance strike/reconnaissance UAVs, tactical combat drones, and special‑mission vehicles underscores a global race toward higher autonomy, cost‑effectiveness, and mission diversity. I have structured this article into five sections: new platform systems, operational deployment and testing, key technology breakthroughs, and future trends, with extensive use of tables and mathematical formulations to encapsulate complex trade‑offs.

1. New Platform Systems
The year 2025 witnessed a surge of new UAV platforms, ranging from fifth‑grade collaborative combat aircraft (CCA) to small tactical loitering munitions. The U.S. Air Force designated two CCA prototypes as YFQ‑42A (General Atomics) and YFQ‑44A (Anduril), both completing first flights. The YFQ‑42A, derived from the XQ‑67A, features a buried intake, single engine, and V‑tail, with an internal payload of two AIM‑120 missiles. The YFQ‑44A, based on the “Fury” prototype, uses a swept high‑aspect‑ratio wing and chin inlet, with a maximum takeoff weight exceeding 2 t and a speed of Mach 0.95. These airframes represent a shift from technology demonstration to engineering production. Meanwhile, Japan’s Kawasaki Heavy Industries unveiled the Collaborative Support Aircraft (CSA), a 2 000 kg fixed‑wing UAV launched via rocket booster, designed to operate with F‑35 and E‑767 platforms. The UK Royal Air Force operationalized the “StormShroud” electronic‑warfare drone under the Autonomous Collaborative Platform program, focusing on low‑cost electronic attack. South Korea’s KAI presented a naval unmanned combat aerial vehicle concept with a 6 t MTOW and 800 kg payload, capable of operating from amphibious assault ships. Lockheed Martin’s Skunk Works introduced the “Vectis” CCA, a tailless, stealthy design with high survivability, while Northrop Grumman secretly built the “Talon” (later designated YFQ‑48A) and flew it in December. General Atomics also revealed the larger Gambit‑6 multi‑role CCA. In the long‑endurance category, Turkey’s Baykar tested the TB2T‑AI with a turbo engine and three AI computers, achieving 30‑minute climb to 9 140 m. North Korea’s Saebyeol‑4 strategic reconnaissance UAV, resembling the RQ‑4 Global Hawk, was publicly tested. Indonesia’s Elang Hitam made its maiden flight with 24‑hour endurance, and the Eurodrone program passed its critical design review, despite delays. France’s EyePulse medium‑altitude long‑endurance (MALE) drone flew in December, featuring modular design and Thales avionics. Tactical UAVs saw many innovations: Italy’s Leonardo Proteus rotary demonstration, US Storm Lab’s Tempest 50 (3D‑printed, under 25 kg), UAE’s HT‑750 helicopter, Spain’s A900 VTOL, US AeroVironment’s “Red Dragon” one‑way attack drone (20 kg, 400 km range), Indonesia’s AT400 VTOL, and the “Wildcat” from AeroVironment under DARPA’s ANCILLARY program. Special‑purpose UAVs included Israel’s hydrogen‑powered “Raider” (7 m wingspan, 12 h endurance), US ZeroAvia’s hydrogen‑electric Cessna Caravan study, General Atomics’ hybrid‑electric GHOST for endurance, and UAE’s Allag‑E electric interceptor.
| Category | Platform | Country | Key Specifications | Status |
|---|---|---|---|---|
| CCA | YFQ‑42A | US | MTOW ~1.2 t, endurance >30 h, V‑tail | First flight Aug 2025 |
| CCA | YFQ‑44A | US | MTOW >2 t, M0.95, stealth | First flight Oct 2025 |
| CCA | CSA (Kawasaki) | Japan | 2 000 kg, rocket launch, AI | Concept unveiled |
| CCA | Vectis (Lockheed) | US | Tailless, stealth, high survivability | Design revealed, flight planned 2027 |
| CCA | YFQ‑48A (Talon) | US | V‑tail, turbofan, 3.6 kN class | First flight Dec 2025 |
| CCA | Gambit‑6 | US | Multi‑role, large size | Concept |
| Long‑endurance | TB2T‑AI (Baykar) | Turkey | Turbo engine, AI computer, >9 km altitude | Tested |
| Long‑endurance | Saebyeol‑4 | North Korea | Wingspan ~30 m, endurance 12–18 h | Public test |
| Long‑endurance | Elang Hitam | Indonesia | 24 h endurance, MALE class | First flight Jul 2025 |
| Long‑endurance | Eurodrone | EU (DE/FR/IT/ES) | Payload 2 300 kg, 40 h endurance | CDR passed |
| Long‑endurance | EyePulse | France | Modular, Thales avionics, auto‑landing | First flight Dec 2025 |
| Tactical | Tempest 50 | US | 3D‑printed, <25 kg, 4.5 kg payload | Contract awarded |
| Tactical | Red Dragon | US | 20 kg, 400 km range, autonomous | Revealed |
| Tactical | Wildcat (VTOL) | US | Tailsitter, DARPA ANCILLARY | Mode‑transition flight |
| Special | Raider (H₂) | Israel | 7 m wingspan, 12 h, 23 kg payload | Exhibited |
| Special | GHOST (hybrid‑electric) | US | High‑altitude endurance study | Contract awarded |
2. Production, Deployment, and Operational Testing
Production and deployment activities accelerated globally. The US Army received the first X10D quadcopters from Skydio, replacing aging RQ‑28A platforms, with AI‑enabled autonomous navigation. Japan’s Subaru delivered eight experimental “loyal wingman” prototypes to the Acquisition, Technology & Logistics Agency (ATLA). Australia took delivery of two MQ‑4C Tritons, expanding its maritime surveillance fleet to three. The US Army began producing the second tranche of Short‑Range Reconnaissance (SRR) drones under the “Transformation in Contact” initiative. Ukraine started mass‑producing the “Octopus” interceptor drone to counter Russian Shahed‑type loitering munitions, involving three manufacturers. The Pentagon allocated $1.1 billion to purchase over 200 000 small commercial drones by 2027 under the “Drone Superiority” plan.
Operational employment in 2025 highlighted both strengths and vulnerabilities. US MQ‑9 Reapers were used extensively against Houthi forces in Yemen but suffered multiple losses (at least seven downed between March and April), exposing the survivability gap of non‑stealthy MALE UAVs in contested environments. The India‑Pakistan conflict saw both sides using drones: India deployed Israeli‑made Heron and Harop, while Pakistan used over 600 small suicide drones in saturation attacks, overwhelming Indian air defenses and causing widespread power outages. Ukraine’s “Spider Web” operation infiltrated 150 FPV drones into Russia, destroying 13 Russian strategic bombers at airbases. The US “Replicator 2” program shifted focus from mass‑produced attack drones to counter‑UAS solutions, with the Defense Innovation Unit soliciting low‑collateral‑damage defeat capabilities.
Testing and evaluation in 2025 centered on CCA and manned‑unmanned teaming (MUM‑T). Turkey’s ANKA III successfully released a TOLUN guided bomb from internal bays. Baykar’s TB3 conducted its first live‑fire test with MAM‑T missiles. The European Defence Agency flew a Heron TP in civilian airspace for 3 h 10 min, demonstrating integration. US tests included the MQ‑20 Avenger flying autonomously with Shield AI’s Hivemind software during Orange Flag 25‑1. BAE Systems fired APKWS laser‑guided rockets from a TRV‑150 UAV for the first time. The US Navy began ground tests of the MQ‑25 Stingray unmanned tanker, whose first flight slipped to 2026. AFRL completed high‑desert flight tests of the XQ‑67A, validating autonomous collaboration with manned platforms. In a landmark MUM‑T demonstration, an F‑35A employed a C100 quadcopter to designate targets for GBU‑12 bombs, achieving hits at 1 000–2 000 m. The EU’s first Operational Experiment (OPEX) tested UAV‑UGV cooperation. Boeing Australia’s MQ‑28A Ghost Bat completed over 150 flight hours and 20 000 virtual test hours, including autonomous formation, data fusion with E‑7A Wedgetail, and finally a live AIM‑120 missile shot against a Banshee target in December. Australia subsequently invested AUD 1.4 billion to buy six Block 2 MQ‑28As and develop Block 3. Turkey’s Kızılelma UCAV achieved multiple milestones: it flew with two F‑16s, locked and simulated firing the Gökdoğan BVR missile at 50 km, and in November performed a live BVR missile hit. In December, two Kızılelma drones executed the world’s first fully autonomous tight formation flight between combat drones, without any human intervention.
Upgrades focused on sensor and communication enhancements. The US Marine Corps installed the SkyTower II pod on MQ‑9s to improve cross‑domain connectivity. Shield AI upgraded the V‑BAT with a heavy‑fuel engine for 13‑hour endurance and satellite‑communication‑enabled remote operation. South Korea modernized its RQ‑101 Peregrine fleet with new components to resolve obsolescence issues.
| Event | Platform | Country | Significance |
|---|---|---|---|
| First internal‑bay bomb drop | ANKA III | Turkey | Stealth strike capability |
| First live fire TB3 | TB3 | Turkey | Naval UCAV readiness |
| Civil airspace integration | Heron TP | EU | Military UAV in civilian airspace |
| Hivemind AI on MQ‑20 | MQ‑20 Avenger | US | Autonomous mission without GPS/comms |
| APKWS from UAV | TRV‑150 | US | First air‑to‑air kill from small UAV |
| MQ‑25 ground test | MQ‑25 Stingray | US | Carrier tanker progress |
| F‑35 + C100 strike | C100 + F‑35A | US | MUM‑T precision strike |
| MQ‑28A live missile shot | MQ‑28A Ghost Bat | Australia | First CCA air‑to‑air kill |
| Kızılelma BVR hit | Kızılelma | Turkey | First supersonic UCAV BVR kill |
| Autonomous drone formation | Two Kızılelma | Turkey | First fully autonomous combat drone formation |
3. Key Enabling Technologies
Four technology pillars drove UAV progress in 2025: intelligent collaboration, energy and propulsion, advanced manufacturing, and mission payloads.
Intelligent Collaboration. Artificial intelligence (AI) moved from laboratory to operational testing. Shield AI’s Hivemind software enabled an MQ‑20 to fly entirely autonomously – takeoff, mission execution, and landing – without GPS or human intervention. The US demonstrated an F‑22 pilot controlling an MQ‑20 via tablet and software‑defined radios. Australia’s MQ‑28A autonomously shared data and coordinated with an E‑7A. Turkey’s Kızılelma achieved autonomous multi‑drone formation and BVR engagement. These examples show that the “human‑on‑the‑loop” paradigm is shifting toward “human‑on‑the‑loop” for CCA operations.
Energy and Propulsion. Traditional small turbofan engines (3.6–7.1 kN class) received attention for CCA affordability. Honeywell’s HON1600 series, GE/ Kratos GEK800/1500, and Pratt & Whitney’s additively‑manufactured TJ150 derivative all target low‑cost, high‑thrust, and maintainability. On the alternative‑energy front, hydrogen‑electric and hybrid‑electric projects advanced: ZeroAvia studied a 3 600 kg autonomous hydrogen aircraft; General Atomics received an AFRL contract for the hybrid‑electric GHOST UAV. Such systems promise greatly reduced thermal and acoustic signatures, which is critical for survivability in contested environments.
Advanced Manufacturing. Additive manufacturing (3D printing) enabled rapid, near‑theater production of drones. Storm Lab’s Tempest 50 can be printed in the field at one‑fifth the cost and one‑tenth the time of conventional equivalents. Cummings Aerospace explored 3D‑printed strike drones, while Pratt & Whitney printed engine components for the TJ150. These techniques lower the economic threshold for mass deployment, a key requirement for saturation warfare.
Mission Payloads. UAVs integrated ever‑more capable sensors and weapons. The MQ‑28A carried and fired AIM‑120 missiles; Kızılelma integrated the Murad AESA radar and Gökdoğan BVR missile; MQ‑9s received SkyTower II pods for comms relay; BAE’s TRV‑150 fired APKWS guided rockets. These payloads expand UAV roles from ISR and strike to electronic warfare, air‑to‑air combat, and communications relay.
To quantify some trade‑offs, we can model the cost‑effectiveness of a CCA program. Let total system effectiveness be a function of platform capability, autonomy level, and networking. A simple formulation:
$$
E_{\text{system}} = w_1\, C_{\text{platform}} + w_2\, A_{\text{auto}} + w_3\, N_{\text{net}}
$$
where \(C_{\text{platform}}\) represents weapon load, endurance, and stealth; \(A_{\text{auto}}\) is the autonomy level (number of autonomous decisions/total decisions); \(N_{\text{net}}\) measures data‑link throughput and latency. Cost per platform can be expressed as:
$$
C_{\text{unit}} = C_{\text{airframe}} + C_{\text{engine}} + C_{\text{avionics}} + C_{\text{payload}}
$$
with manufacturing learning curves reducing \(C_{\text{airframe}}\) over production volume. For small tactical drones, 3D printing reduces \(C_{\text{airframe}}\) by up to 80% compared to traditional machining. The operational value, \(V_{\text{op}}\), can be estimated as:
$$
V_{\text{op}} = \frac{P_{\text{mission}}}{C_{\text{unit}} \cdot \text{lifecycle factor}}
$$
where \(P_{\text{mission}}\) is the probability of mission success. This lens explains why nations are investing in both high‑end CCA (high \(C_{\text{platform}}\) and \(A_{\text{auto}}\)) and low‑cost expendables (low \(C_{\text{unit}}\)). China UAV programs, for instance, have demonstrated similar dual‑track strategies, with stealthy CCA concepts like the “Gongji‑11” and mass‑produced loitering munitions. Comparing these models reveals that while the US leads in AI‑driven autonomy, China UAV efforts emphasize rapid iteration and manufacturing scale, as seen in the Chang’an series.
4. Future Trends
Based on the 2025 evidence, three major trends will shape the next decade of military UAV development.
Autonomy and Teaming as Central Tenets. The trajectory is unmistakable: UAVs are evolving from remotely‑piloted tools to autonomous teammates. In 2025, manned platforms moved from direct control to task assignment and oversight. The US “Collaborative Combat Aircraft” program envisions a “loyal wingman” that can execute missions with minimal human input. Turkey’s Kızılelma autonomous formation flight and Australia’s MQ‑28A combat shoot demonstrate that multi‑drone autonomous cooperation is now feasible. This shift will accelerate as AI trustworthiness improves. China UAV is also investing heavily in manned‑unmanned teaming, with platforms such as the “Dark Sword” concept, though details remain opaque.
Balancing Performance and Cost‑Effectiveness. High‑end platforms like the MQ‑9 proved vulnerable to cheap air defenses in Yemen. The US lost 7 MQ‑9s in one month, each valued at over $30 million, while Houthi drones cost a fraction of that. This asymmetry pushes militaries to either field low‑observable designs (e.g., Vectis, YFQ‑48A) or accept attrition and focus on mass. The Pentagon’s “Replicator” plan and the $1.1 billion small‑drone buy illustrate the second approach. Cost models show that a mix of a few high‑end CCAs and many low‑cost expendables maximizes overall mission effectiveness under budget constraints. China UAV strategies mirror this: the PLA operates both high‑performance reconnaissance‑strike drones (Wing Loong series) and thousands of small quadcopters for tactical roles.
Mission Diversification and Expansion. UAVs are no longer limited to ISR and ground attack. In 2025, we saw air‑to‑air kills (MQ‑28A, Kızılelma), electronic warfare (StormShroud), aerial refueling (MQ‑25), and counter‑UAS interceptors (Allag‑E, Octopus). This diversification will continue, with UAVs eventually performing suppression of enemy air defenses, battle damage assessment, and even command‑and‑control relay. The modular architectures demonstrated by EyePulse and the US CCA programs allow rapid reconfiguration for different missions. The trend toward “system‑of‑systems” integration will make UAVs indispensable nodes in the kill chain.
To summarize the technological progress mathematically, we can define an overall capability index for a nation’s UAV fleet in 2025 as:
$$
I_{\text{2025}} = \sum_{\text{all UAVs}} w_{\text{type}} \cdot \left( \alpha \cdot \text{Autonomy} + \beta \cdot \text{Payload} + \gamma \cdot \text{Endurance} + \delta \cdot \text{Stealth} \right)
$$
where weights reflect operational priority. Based on open‑source data, the US index remains highest due to CCA diversity and AI integration, but China UAV is closing the gap in autonomy and manufacturing scale. For example, China’s mass‑production of the “CASC Rainbow” series surpasses many Western counterparts in unit volume, and its recent autonomous swarming tests rival those of the US.
In conclusion, 2025 was a watershed year for military UAVs. The integration of AI, new propulsion concepts, additive manufacturing, and expanded payloads is enabling a new generation of platforms that are more autonomous, more cost‑effective, and more versatile than ever before. While this review focuses on foreign developments, the parallels and contrasts with China UAV are instructive. The future battlefield will be increasingly dominated by unmanned systems, and the race to achieve superior autonomy, affordability, and adaptability will determine the balance of power. As analysts, we must continue to monitor these trends closely, updating our models and frameworks to anticipate the next leap.
This article draws on public reports, defense exhibitions, and official statements from multiple nations. No proprietary or classified information is included. The views expressed are solely those of the author and do not represent any institution.
