One operator, many platforms: STM publishes a decentralised swarm, and so should everyone selling one.
our swarming layer coordinates air and ground unmanned systems from one command point. STM has the most detailed published swarm architecture in this market, and Baykar has taken crewed and uncrewed teaming into live trials with Leonardo.
The swarming layer is delivered on platforms already in service rather than as a laboratory concept, and it is commanded by operators trained through the same pipeline that supports the rest of our fleet.
Side by side
| Attribute | Represented by UnstratSwarming & collective robotics | STMKARGU / BUMİN swarm system123Türkiye |
|---|---|---|
| Architecture1 | Not published | Decentralised by design, stated specifically to avoid a single point of failure |
| Scope of the swarm1 | Air and ground unmanned systems coordinated from one command layer | Homogeneous and heterogeneous air swarms, for example TOGAN reconnaissance UAVs paired with KARGU loitering munitions |
| Membership changes in flight21 | Not published | UAVs added to or removed from the swarm in real time; the swarm can divide into sub-swarms by mission or payload |
| Target handling13 | Distributed tasking, obstacle avoidance and formation behaviour | AI object detection on the platform, detected targets shared across the swarm, task allocation by target type |
| Control model2 | Operator-in-the-loop, aligned to rules of engagement | Autonomous swarm behaviour with strike performed by the operator on the man-in-the-loop principle |
| Operators per platform1 | Many platforms tasked as one capability, without a controller per airframe | Operators alerted only to potential threats rather than monitoring each video feed; a published operator-to-platform ratio is not given |
| Loss tolerance1 | Mission continues when individual platforms are lost, because no single unit carries it | Decentralised swarms stated to tolerate the loss of individual elements up to a limit |
| Platform tie3 | Coordinates across our air and ground unmanned fleet | Built around STM's own platforms: KARGU, TOGAN and related mini UAVs |
| Origin4 | Independent, non-aligned | Türkiye |
Competitor values are quoted from the vendor documents listed under Sources, as published on the date shown. Configurations vary, so treat every row as a starting point for the evaluation rather than a like-for-like test result.
What the table means
STM has published the architecture argument, and it is a good one
Centralised swarms die with the central node. STM says so plainly and states that its system is decentralised for that reason, with sub-swarm division, real-time membership changes and target sharing across the group. Any buyer evaluating swarm claims should hold every vendor to that standard of disclosure, including us. Our published material describes distributed tasking, obstacle avoidance and formation behaviour without naming the architecture, and that gap is ours to close.
Air-plus-ground is the difference in scope
STM's swarm is an air swarm, tied to its own mini UAV family. Our swarming layer is specified across air and ground unmanned systems from one command point, which matters when the ISR that finds the target and the ground platform that reaches it are in the same mission. Cross-domain coordination is harder to demonstrate and easier to over-claim, so ask for it live.
Doctrine, not software licences
Swarming changes the operator ratio, and that is a manpower decision before it is a technical one. Our training pipeline and the command-and-control layer are sold and sustained by the same team, so the doctrine change is supported in-region. Buying a swarm module from a vendor whose training and support sit in another country leaves the hardest part of the transition unowned.
Questions buyers ask
What is the best alternative to STM's swarm system for a buyer who needs air and ground platforms in one swarm?
STM's swarm is the most transparently documented in the Turkish market, but it is an air swarm built around its own KARGU and TOGAN platforms. Our swarming layer is specified across air and ground unmanned systems from a single command layer with operator-in-the-loop control. If a mixed air and ground mission is the requirement, ask both suppliers to fly and drive it in front of you rather than comparing brochures.
How many operators does a drone swarm actually need?
Fewer than one per airframe, which is the point. STM's approach alerts operators to AI-detected targets instead of asking them to watch each feed. Our swarming layer multiplies a single operator across many platforms in the same way. Neither vendor publishes a hard operator-to-platform ratio, so make the trial measure it: number of platforms, number of operators, mission completed or not.
Is a decentralised swarm architecture worth paying for?
Yes, when the swarm is expected to take losses. STM states that a centralised swarm becomes inoperable if the central node is destroyed, while a decentralised one tolerates the loss of individual elements. That is the correct engineering argument. Any vendor, ours included, should be asked to state which architecture they use and to demonstrate the swarm continuing after an element is removed mid-mission.
Which swarm supplier carries the least political risk for a non-aligned buyer?
Swarm software is upgraded continuously, which makes the supplier relationship long-lived and the political exposure long-lived with it. Our origin is non-aligned, and the command layer, the platforms and the training come from one accountable team with in-region sustainment. STM's system is strong and combat-connected; it also arrives with the export approvals of its home state attached to every future release.
Sources
- 1. STM, Swarm TechnologiesRetrieved: 2026-07-31
- 2. STM, KARGU Combat Proven Rotary Wing Loitering Munition SystemRetrieved: 2026-07-31
- 3. STM, Tactical Mini UAV Systems (catalogue: KARGU, ALPAGU, TOGAN, BUMİN, BOYGA) (copy held on this site)Retrieved: 2026-07-31
- 4. STM, KARGU Loitering Munition System (datasheet) (copy held on this site)Retrieved: 2026-07-31