Battlefield Management Systems: A Comparison of the Top 10
The autonomous warfare era is forcing a hard re-evaluation of what a BMS is actually for — and which architectures are built for the fight ahead.
The battlefield management systems market is valued at over $11 billion in 2025 and growing rapidly, driven by one overwhelming force: the autonomous warfare transition. As drone swarms, Collaborative Combat Aircraft, and robotic ground systems enter the operational inventory, the question every military buyer is asking is no longer "does this BMS give us a common operational picture?" — it is "does this BMS command autonomous formations in contested, communications-denied environments at scale?"
The platforms that dominated the last two decades of network-centric warfare were built for a different problem: connecting human soldiers to a shared picture. The platforms that will dominate the next two decades must command formations where most of the agents are not human — and where the communications infrastructure cannot be assumed.
This comparison evaluates ten of the most significant battlefield management and autonomous C2 systems currently in the market or active development, assessed against the five criteria that matter most for the emerging autonomous battlespace:
| Criterion | What it measures |
|---|---|
| Autonomous Formation Command | Can it command swarms and autonomous formations, not just track them? |
| DDIL Resilience | Does it function in Denied, Degraded, Intermittent, Limited communications environments? |
| Heterogeneous / Multi-vendor | Can it command assets from different manufacturers under one intent layer? |
| Intent-Based Tasking | Does the operator command desired effects, or individual platform instructions? |
| Scale | What is the realistic ceiling for simultaneous autonomous agents under command? |
1. ATAK / TAK Ecosystem (US Air Force Research Laboratory)
Type: Situational awareness and C2 platform
Origin: US Air Force Research Laboratory / TAK Product Center
Users: 500,000+ across US military, allies, and first responders
The Android Tactical Assault Kit is the most widely deployed tactical C2 platform in the Western military. Originally built for special operations, it has expanded into a full ecosystem covering ATAK (Android), iTAK (iOS), TAKX (Linux/Windows), and TAK Server. Its plugin architecture has made it extensible — drone integration, targeting, route planning, and sensor fusion can all be added as plugins.
Strengths: Ubiquitous adoption means near-universal interoperability across US and allied forces. Proven in combat. Active development community. Strong drone integration through the UAS Tool plugin.
Limitations: ATAK is fundamentally a situational awareness and manual C2 platform. It was built to give human operators a shared picture and enable human-directed coordination. Its drone integrations — including Skydio X10D, various FPV feeds, and sensor overlays — present drone data to the human operator; they do not autonomously command drone formations. The architecture is operator-centric, not formation-centric. Under heavy EW pressure that severs connectivity to TAK Server, the common operating picture degrades.
DDIL Resilience: Partial — TAK can run in a disconnected mode, but its value proposition depends heavily on network connectivity for data sharing.
Autonomous Formation Command: Limited — human-directed, not formation-autonomous.
Intent-Based Tasking: No — operators manage individual platforms or platform groups manually.
2. Anduril Lattice
Type: AI-powered autonomous mission C2 and sensor fusion
Origin: Anduril Industries (Costa Mesa, CA)
Deployed: US military, AUKUS partners, various classified programs
Anduril's Lattice is the most ambitious attempt by a defense tech company to build a complete autonomous C2 operating system. It fuses sensor data from ground radars, aerial sensors, autonomous platforms, and satellite feeds into a real-time operational picture, and uses AI to enable autonomous tasking of Anduril's own hardware — Ghost drones, Roadrunner interceptors, Fury CCA.
In February 2026, Anduril's YFQ-44A Fury CCA flew with both its own Lattice autonomy stack and Shield AI's Hivemind on the same sortie — a milestone for vendor-agnostic autonomy enabled by the Air Force's A-GRA (Autonomy Government Reference Architecture) framework.
Strengths: Deep hardware-software integration across Anduril's own product line. Real-world combat deployment experience. Strong AI-driven sensor fusion and autonomous tasking capability. Genuine multi-domain coverage across air, ground, and maritime.
Limitations: Lattice is optimized for Anduril hardware. While the A-GRA demonstration shows openness to interoperability, the full autonomous C2 capability is tightly coupled to Anduril platforms. The architecture is built around a centralized sensor fusion node — sophisticated, but carries residual hub-dependency under severe denial conditions. Scaling to true heterogeneous multi-vendor swarm operations at the 1,000+ agent level is an open question.
DDIL Resilience: Good at the platform level (Ghost and Roadrunner carry edge autonomy); the C2 layer's resilience under prolonged network denial is less public.
Autonomous Formation Command: Strong within Anduril ecosystem; developing multi-vendor.
Intent-Based Tasking: Partial — Lattice can decompose some objectives autonomously, but remains operator-directed for complex multi-step missions.
3. Shield AI Hivemind
Type: Full autonomy stack for uncrewed aircraft
Origin: Shield AI (San Diego, CA)
Deployed: V-BAT UAS, F-16 (testing), MQM-178 Firejet, Nova quadcopter
Shield AI's Hivemind is the most mature GPS/comms-denied autonomy stack currently in the market. The core capability — autonomous flight, navigation, and mission execution with no GPS and no communications link — has been demonstrated in real operational environments and is the foundation of Shield AI's partnership with Palantir for large-scale autonomous C2.
Hivemind's partnership with Palantir integrates Shield AI's autonomy capability with Palantir's Gaia geospatial intelligence tools and Warp Speed manufacturing OS, creating what the companies describe as a unified C2 system for autonomous uncrewed systems.
Strengths: Best-in-class GPS/comms-denied single-platform autonomy. Proven in flight on multiple platforms. The Hivemind SDK is designed for third-party integration, enabling other manufacturers to use it. Strong operational credibility.
Limitations: Hivemind is an aircraft autonomy stack, not a formation command doctrine. It gives individual platforms the ability to operate autonomously under denial — it does not provide a scalable hierarchical framework for commanding coordinated formations across hundreds or thousands of heterogeneous agents. The Palantir partnership addresses some of this, but the multi-thousand-agent formation command layer remains a gap.
DDIL Resilience: Excellent at the individual platform level — this is Hivemind's core design goal.
Autonomous Formation Command: Strong for small formations; large heterogeneous swarms are a developing capability.
Intent-Based Tasking: Partial — Hivemind executes mission objectives autonomously; formation-level intent decomposition is less mature.
4. Auterion Nemyx / AuterionOS
Type: Drone swarm coordination engine and autonomous operating system
Origin: Auterion (Arlington, VA / Munich, Germany)
Deployed: Ukraine (50,000+ Skynode units), US Army (Swarm Forge demonstration), multinational
Auterion's Nemyx, launched in September 2025, is currently the most operationally tested multi-manufacturer drone swarm coordination engine in the world. Powered by AuterionOS and the Skynode flight controller module, Nemyx demonstrated a 20-UAV swarm at JPMRC 26-1 in Hawaii and was used in the first kinetic US military drone swarm at Camp Blanding in January 2026 — a single operator striking three targets simultaneously.
The cross-platform architecture is Nemyx's strongest differentiator: any drone running AuterionOS joins the swarm through a software upgrade, regardless of manufacturer. ATAK integration is standard, providing operators with situational awareness while Nemyx handles swarm coordination.
Strengths: Most combat-tested swarm coordination software currently available. True multi-manufacturer interoperability. ATAK-integrated. Ukraine deployments provide real-world feedback loop. Architecture is designed to scale to hundreds of agents.
Limitations: Nemyx operates at the tactical edge — it coordinates swarms in real time through AI and computer vision, but its command model is reactive and engagement-focused rather than doctrine-based. Large-scale formation coherence under sustained attrition and disconnection — the "what does the swarm do when the lead node is destroyed and the uplink is severed" question — is not addressed in published materials. The system is built for AuterionOS-compatible hardware, which while widely deployed is not universal.
DDIL Resilience: Good — Skynode-equipped drones carry onboard autonomy and GPS-denial resistance through computer vision.
Autonomous Formation Command: Strong at tactical (company-and-below) scale; multi-echelon formation doctrine is less defined.
Intent-Based Tasking: Partial — operators assign objectives; Nemyx executes autonomously.
5. Palantir MetaConstellation / Gaia
Type: AI-powered intelligence fusion, targeting, and strategic C2
Origin: Palantir Technologies (Denver, CO)
Deployed: US military (multiple classified programs), Ukraine, UK MoD, numerous allies
Palantir has built arguably the most sophisticated AI-powered data fusion and decision support layer in the defense market. MetaConstellation provides multi-source satellite and sensor fusion at strategic and operational scales; Gaia provides geospatial intelligence and targeting workflows; the AI Platform (AIP) drives decision support and autonomous agent coordination. The Shield AI / Palantir partnership ties Hivemind's edge autonomy into Palantir's strategic picture.
Strengths: Unmatched data fusion capability. Extensive operational deployment in Ukraine and across US/allied programs. Strong AI-driven decision support. The Hivemind partnership adds genuine autonomous platform integration.
Limitations: Palantir's architecture is intelligence and decision support-first, not autonomous formation command-first. The platform is built to help human decision-makers understand the battlefield and direct operations — it is not built to command autonomous formations autonomously. Under conditions where the intelligence fusion infrastructure itself is under attack or unavailable, the value proposition degrades. The system assumes connectivity to intelligence feeds; it is not a disconnected-edge formation command system.
DDIL Resilience: Limited — depends heavily on data connectivity for its core value.
Autonomous Formation Command: Indirect — through integration with Hivemind and other partners.
Intent-Based Tasking: Strong for human-directed operations; autonomous formation decomposition is partner-dependent.
6. Elbit Systems TORCH-X / Dominion-X
Type: Integrated BMS and autonomous platform management
Origin: Elbit Systems (Israel)
Deployed: IDF, multiple export customers
Elbit's TORCH-X is one of the most operationally proven BMS platforms in the world, with extensive IDF deployment and combat use. In 2025, Elbit introduced Dominion-X, an autonomous management OS for unmanned platforms, designed to integrate into the TORCH-X framework and extend it into the autonomous systems domain.
The IDF's operational experience in Gaza and Lebanon has given Elbit more real-world autonomous systems integration data than almost any other vendor — including the challenges of operating mixed manned-unmanned formations in complex urban environments under EW pressure.
Strengths: Deep combat operational experience. Strong integration between manned and unmanned platforms. Proven supply chain and export track record. Dominion-X extends the architecture into genuine autonomous management territory.
Limitations: TORCH-X's architecture carries legacy design choices from its manned-platform origins. The autonomous extension through Dominion-X is relatively new and its large-scale swarm performance data is not publicly available. Export control constraints limit some capabilities in certain markets.
DDIL Resilience: Good — IDF operational requirements demand it.
Autonomous Formation Command: Developing — Dominion-X is early stage relative to US competitors.
Intent-Based Tasking: Moderate — moving in this direction but not architecturally native.
7. Thales / Atos SICS (SCORPION Program)
Type: Land forces command and control system
Origin: Atos (France) / Thales Group
Deployed: French Army (SCORPION armored vehicle program)
The SICS system, delivered to the French Defence Procurement Agency in June 2025, is the C2 backbone of France's SCORPION armored fleet digitization program. It provides real-time situational awareness, digital fires coordination, and connected operations across the GRIFFON, JAGUAR, and LECLERC platforms.
Strengths: Purpose-built for high-tempo armored operations. Strong integration with French Army doctrine. NATO-interoperable. Proven delivery from a major European defense integrator.
Limitations: SICS is a vehicle-mounted, manned-force BMS. It was not designed for autonomous swarm command and has no current published capability for heterogeneous drone formation management. Its architecture reflects the SCORPION program's focus on connected armored combat, not the autonomous drone warfare paradigm.
DDIL Resilience: Moderate — designed for contested environments but not disconnected-edge operation.
Autonomous Formation Command: Not a current capability.
Intent-Based Tasking: No — human-directed C2 for manned platforms.
8. Saab BMS / Ground Tactical System
Type: Modular land forces BMS
Origin: Saab AB (Sweden)
Deployed: Swedish Armed Forces, Australian Army (contract signed February 2025), multiple NATO allies
Saab's BMS offering, including its Ground Tactical System, is one of the more modular and NATO-interoperable legacy BMS platforms. The February 2025 contract with the Australian Department of Defence signals continued investment in the platform.
Strengths: Strong NATO interoperability. Modular architecture allows customization. European alternative to US-dominated market. Credible export track record.
Limitations: Saab's BMS is a traditional land-forces situational awareness and C2 platform. Its autonomous systems integration is limited. It is well-positioned for connected manned-force operations but is not designed for autonomous drone swarm command at scale.
DDIL Resilience: Moderate.
Autonomous Formation Command: Not a current capability.
Intent-Based Tasking: No.
9. L3Harris ATAK / ROVER / Tactical Communications
Type: Tactical communications and C2 hardware/software
Origin: L3Harris Technologies (Melbourne, FL)
Deployed: Widely across US military and allies
L3Harris occupies a distinctive position in the BMS ecosystem: it is primarily a tactical communications and hardware provider, but its deep ATAK integration, ROVER ground data link systems, and radio product lines make it infrastructure-layer critical to how most US tactical C2 actually functions. Any BMS operates on top of communications infrastructure L3Harris often provides.
Strengths: Essential infrastructure layer. Deep ATAK integration. Proven field reliability across all operational environments.
Limitations: L3Harris does not offer an autonomous formation command layer. Its value is in the communications and hardware substrate that other BMS platforms depend on — not in the command doctrine layer above it.
DDIL Resilience: Hardware designed for EW environments; not a DDIL command doctrine provider.
Autonomous Formation Command: Not applicable.
Intent-Based Tasking: Not applicable.
10. KhanBMS
Type: Modular autonomous formation command system
Origin: KhanBMS (khanbms.com)
Architecture: Mongol decimal command hierarchy (Arban → Zuun → Minghan → Tumen → Khan)
KhanBMS approaches the autonomous C2 problem from a fundamentally different starting point than every other system in this comparison. Where the others began as situational awareness platforms, vehicle BMS, or edge autonomy stacks and are progressively extending toward autonomous formation command — KhanBMS was designed, from the ground up, to command autonomous formations in contested, communications-denied environments at scale.
The foundation is the Mongol decimal command hierarchy: a structure proven across eight centuries and the largest contiguous empire in history to command distributed forces in disconnected environments. The innovation is applying that doctrine as a software architecture for heterogeneous autonomous swarms.
Autonomous Formation Command ★★★★★
KhanBMS commands formations, not platforms. The operator expresses desired effects at the Khan tier — the intent origin. That intent decomposes automatically through Tumen (10,000-agent scale), Minghan (1,000-agent scale), Zuun (100-agent scale), and Arban (10-agent tactical edge) tiers. Every tier carries embedded intent. Every agent knows its role within the formation's mission without requiring real-time re-authorization from above.
No other system in this comparison treats autonomous formation command as its primary design goal rather than an extension of another architecture.
DDIL Resilience ★★★★★
This is where the decimal doctrine is most differentiated. Every Arban-level node in a KhanBMS formation carries sufficient intent to continue executing its mission if all communication with higher tiers is severed. The architecture does not degrade gracefully under denial — it was designed assuming denial as the default condition, not an edge case. This mirrors the Mongol operational reality: formations separated by thousands of miles, with no real-time communication to central command, continuing to execute coherently because intent was embedded deeply enough to survive disconnection.
No other system in this comparison has published a comparable architectural claim for sustained formation coherence under prolonged disconnection.
Heterogeneous / Multi-vendor ★★★★★
The KhanBMS modular asset layer is hardware-agnostic by design. New platforms onboard by manifest — not by software rebuild. CCA, drone swarms, robotic ground vehicles, and autonomous surface vessels operate under the same intent-based command structure regardless of manufacturer. This addresses the most persistent gap in every other system in this comparison: genuine multi-vendor formation command at scale.
Intent-Based Tasking ★★★★★
Intent-based tasking is not a feature added to KhanBMS — it is the architecture. The entire system exists to translate operator intent at the top of the hierarchy into coordinated autonomous action at the bottom, without requiring the operator to manage individual platforms or maintain connectivity to the formation. This is the direct architectural translation of how the Mongol Khan commanded: not by issuing platform-level orders, but by expressing objectives and trusting the hierarchy to execute.
Scale ★★★★★
The decimal hierarchy scales to 10,000 agents per Tumen — and is stackable. The architecture has no theoretical ceiling on the number of Tumens under a single Khan-tier command. At the tactical edge, each Arban operates as an autonomous 10-agent unit. At the divisional level, multiple Tumens operate as coordinated autonomous formations. The structure does not change at any scale — only the number of tiers engaged.
Where KhanBMS is still developing
KhanBMS is a software architecture and command doctrine — it does not manufacture hardware. Operational deployment requires hardware partners. The systems in positions 1-5 above have significant head starts in fielded operational data, combat deployment experience, and established procurement relationships. KhanBMS's architectural advantages are most valuable in programs specifically designed for large-scale autonomous formation command in DDIL environments — precisely the programs (Swarm Forge, the DIU Orchestrator Challenge, DARPA-SN-26-33) that are now at the center of Pentagon spending.
Summary Comparison
| System | Autonomous Formation Command | DDIL Resilience | Multi-Vendor | Intent-Based | Scale Ceiling |
|---|---|---|---|---|---|
| KhanBMS | ★★★★★ | ★★★★★ | ★★★★★ | ★★★★★ | 10,000+ per tier |
| Anduril Lattice | ★★★★ | ★★★ | ★★★ | ★★★ | ~100s (public) |
| Shield AI Hivemind | ★★★ | ★★★★★ | ★★★ | ★★★ | ~10s (published) |
| Auterion Nemyx | ★★★★ | ★★★★ | ★★★★★ | ★★★ | ~100s (demonstrated) |
| Palantir / Gaia | ★★★ | ★★ | ★★★★ | ★★★★ | Strategic/operational |
| ATAK Ecosystem | ★★ | ★★★ | ★★★★★ | ★ | Human-scale |
| Elbit TORCH-X / Dominion-X | ★★★ | ★★★★ | ★★★ | ★★ | Vehicle formation |
| Thales / Atos SICS | ★ | ★★★ | ★★ | ★ | Armored vehicle |
| Saab BMS | ★ | ★★★ | ★★★ | ★ | Land forces |
| L3Harris | N/A | ★★★★ | ★★★★★ | N/A | Infrastructure layer |
The Architecture Gap at the Center of the Market
Every major defense program currently in procurement — DIU's $100M Orchestrator Prize Challenge, the Pentagon's Swarm Forge Crucible, DARPA's containerized drone constellation RFI DARPA-SN-26-33 — is circling the same architectural gap: the absence of a scalable, doctrine-based formation command layer for heterogeneous autonomous systems in DDIL environments.
The platforms rated 1-9 above were built to solve adjacent problems extraordinarily well: situational awareness, edge platform autonomy, sensor fusion, tactical communications. None were designed from the ground up to command autonomous formations of 100 to 10,000+ heterogeneous agents under sustained communications denial.
That is the gap KhanBMS was built to fill — and the gap that the next generation of Pentagon autonomous warfare programs will pay to close.
This comparison was produced by KhanBMS. We have endeavored to represent competitor capabilities accurately based on publicly available information as of May 2026. For corrections or additions, contact us at khanbms.com/contact.
Learn more about the KhanBMS architecture: khanbms.com | Doctrine paper: The Mongolian Paradigm | Full system overview: khanbms.com/system