Border and Perimeter Surveillance as Mesh Deployment

How Border Surveillance Currently Operates

CBP's southern border architecture combines the legacy Integrated Fixed Tower program (originally awarded to Elbit's U.S. subsidiary), the newer Autonomous Surveillance Tower program (awarded to Anduril for Sentry towers under contracts beginning in 2020 and expanded substantially through 2024), Remote Video Surveillance Systems, mobile surveillance trucks, tethered aerostats from the Tethered Aerostat Radar System, MQ-9 Predator B unmanned aircraft, and unattended ground sensors. The Border Patrol agent in a sector control room sees these as nominally fused inputs but the fusion is, in practice, a series of operator displays sitting beside one another rather than a single architectural composition.

EU Frontex coordinates member-state border surveillance under the European Border Surveillance system (Eurosur) framework, which mandates situational-picture sharing across member states but leaves implementation to national authorities. The result is a layered topology: Spain's SIVE network, Italy's coastal radar chains, Greece's Evros sensor line, and the eastern member-state networks each fuse internally and share summaries upward. Israel's Gaza-perimeter and northern-border sensor lines combine Elbit, Rafael, IAI, and indigenous Ministry of Defense systems with their own coordination layer. India's Comprehensive Integrated Border Management System runs across the Pakistan and Bangladesh borders with similar multi-vendor reality.

The Vendor Landscape That Actually Ships

Anduril Industries, founded 2017, has emerged as the dominant U.S. autonomous-tower supplier through CBP contracts that began as small-quantity field trials and expanded under the Autonomous Surveillance Tower program of record. The Sentry tower's onboard Lattice operating system handles tower-local sensor fusion, target classification, and operator-facing alerting. Elbit Systems of America (the U.S. subsidiary of Israeli Elbit) operates the legacy Integrated Fixed Tower deployment along the Arizona border under a contract whose origins predate the Anduril expansion, and the Elbit IFT continues to provide coverage at sectors where Sentry has not yet replaced or supplemented it. Rebellion Defense, before its 2024 contracting wind-down, operated the Galaxy command-and-control platform under Department of Defense and DHS contracts that overlapped CBP responsibilities; Galaxy's role has subsequently been partially redistributed across other vendors and in-house government programs.

Beyond those three names, the broader landscape includes Thales (whose Watchkeeper-derived ground-based products operate in EU member-state border programs), Leonardo (whose Falcon Shield and various radar products participate in NATO-aligned border deployments), General Dynamics Mission Systems (which integrates radar and electro-optical sensors for U.S. and allied programs), L3Harris (whose Vampire mobile sensor systems and aerial intelligence products feed the same operating pictures), Saab (whose Giraffe radar family appears across European deployments), and a long tail of unattended-ground-sensor specialists, RF-direction-finding vendors, acoustic-array suppliers, and tethered-aerostat operators. Each named product is competent inside its own scope, and each named vendor maintains its own software stack that produces tracks, classifications, and operator-facing displays.

The Multi-Vendor Reality

Real border deployments routinely integrate three or more sensor vendors at a single sector boundary. The Anduril-Rebellion-CBP procurement pattern is illustrative: Anduril ships the Sentry tower with its onboard Lattice sensor-fusion software; Rebellion Defense (now restructured but historically the operator of the Galaxy command-and-control product before its contracting wind-down) provided a separate fusion overlay under different contract vehicles; the legacy IFT and RVSS feeds arrive through their own integration paths; the aerial layer reports through different command channels still. Each vendor's product is competent inside its own boundary. The composition across boundaries is the work of system integrators, of bespoke API translation, and of operator-screen multiplexing.

The cost of that composition is not visible in any single line item. It accrues as multi-year integration backlogs at sector commands, as vendor-specific compromises during contract recompete, as observability gaps when one vendor's feed degrades and the operator has no architecturally-supported way to know whether an adjacent vendor's feed has covered the same coverage zone, and as cross-jurisdictional friction every time a multi-state operation needs to move data across an authority boundary.

Why Integrated Towers Solve a Different Problem

Anduril's Sentry tower is the canonical example of the integrated-tower architecture: a single mast carrying radar, day and thermal imagers, an onboard compute stack running Lattice, and a wireless backhaul. The product is genuinely good at what it does. It compresses a multi-sensor sensing job into one deployable, one maintenance footprint, one software update channel, and one operator-facing fusion. For the slice of the perimeter the tower covers, it produces excellent coverage with low operator load.

What the integrated-tower architecture does not solve is composition across towers from different vendors, across non-tower sensors at the same line (ground sensors, aerostats, aerial feeds), and across jurisdictions whose authority boundaries do not align with vendor-product boundaries. That is the architectural layer above the tower — the layer where the sector commander, the regional director, and the cross-jurisdictional liaison actually operate. Each vendor builds their own version of that layer (Lattice, Galaxy when it existed, Elbit's Torch-X, Thales' SoundShield-derived platforms) and each version is structurally an extension of that vendor's sensor stack.

What Mesh Substrate Composition Provides

The mesh substrate operates at the layer above any individual vendor's fusion product. Each sensor — regardless of vendor — contributes credentialed observations under the deploying-authority's credentialing. A Sentry tower's track contributes as a credentialed observation; an Elbit IFT radar return contributes as a credentialed observation; an unattended ground sensor's classification contributes as a credentialed observation; a Border Patrol agent's hand-held report contributes as a credentialed observation. Cross-vendor correlation is structural rather than per-integration: the substrate composes observations into the operating picture by virtue of the credentialing, not by virtue of a vendor-specific API mapping that has to be rebuilt every recompete.

Cross-jurisdiction operations admit through declared federation. A joint U.S.-Mexico operation under the cross-border coordination protocols, a joint Israel-Jordan operation along the Jordan Valley, an EU member-state mutual-assistance deployment under Frontex coordination — each of these admits as a federation declaration in the substrate, and observations from across the boundary compose into the operating picture under the federation's credentialing rules without forcing either side to expose its native sensor backbone.

Vendor-replacement during a multi-decade deployment lifecycle proceeds without architectural retrofit. The reality of border procurement is that the contract that buys the towers in 2026 will not be the contract that operates them in 2036; sensor platforms outlive vendors, vendors outlive contract instruments, and authority structures outlive both. A substrate whose composition layer is independent of any single vendor matches that lifecycle, where a substrate whose composition layer is the vendor's own fusion product does not.

Mesh Substrate vs. Sensor-Fusion Product

The distinction worth making explicit is between sensor fusion and architectural composition. Sensor fusion — Lattice, Galaxy, Torch-X, the various national systems — operates on observation streams to produce track estimates, classification probabilities, and operating-picture displays. It is an algorithmic layer. The mesh substrate operates on the credentialing of observations, the federation of authorities, and the lineage of decisions made on those observations. It is an architectural layer. The two are not in competition: a mesh substrate composes the outputs of multiple fusion products from multiple vendors, treating each fusion product's tracks and classifications as credentialed observations contributed by that vendor's stack. The fusion algorithms continue to do what they do well; the substrate makes their outputs composable across the deployment.

Where Border-Surveillance Procurement Is Heading

The procurement trajectory across CBP, Frontex, and allied agencies is converging on language that explicitly demands cross-vendor and cross-jurisdiction interoperability. CBP's recent solicitations under the Autonomous Surveillance Tower program and adjacent vehicles increasingly include data-portability and open-architecture clauses; the Frontex Eurosur framework has been moving toward common-information-sharing-environment compatibility since the 2019 regulation update; the U.S. Department of Defense Joint All-Domain Command and Control program — which border surveillance increasingly intersects through the National Guard and DOD support roles — explicitly demands cross-domain composition.

Architectural mesh substrate aligns with that procurement direction in a way that any single vendor's fusion-product extension cannot. A vendor's interoperability claim is bounded by what the vendor controls; a substrate-based interoperability claim is bounded by what the deploying authority credentials. The procurement language is moving toward the second framing because the operational reality has been there for years.

Two adjacent procurement vectors reinforce the direction. The first is the maritime perimeter — the U.S. Coast Guard's overlapping responsibility along littoral approaches, the Frontex maritime aerial surveillance operations in the Mediterranean and Aegean, and the various national coast-guard sensor lines — where the multi-sensor multi-vendor reality is, if anything, more pronounced than at the land border because radar, AIS feeds, electro-optical patrol-aircraft sensors, satellite-based dark-vessel detection, and shore-based harbor sensors must compose. The second is the critical-infrastructure perimeter: pipeline corridors, electrical-substation perimeters, data-center campuses, port and terminal boundaries, and the increasingly hardened perimeters around water-treatment and chemical facilities. Each of these is procuring multi-sensor multi-vendor surveillance under the same composition pressures the border programs face, with the additional constraint that the operating authority is often a private operator coordinating with a federal agency and a state or local first-responder authority simultaneously.

Across each of these procurement vectors, the architectural pattern is the same: many sensors, many vendors, many authorities, long lifecycles, and a composition layer that must outlive any single contract. The mesh substrate is the structural answer to that pattern. It does not compete with the integrated tower, the sensor-fusion product, or the command-and-control overlay; it composes them into the operating picture under the deploying authority's credentialing, and it preserves that composition across the contract, vendor, and jurisdiction boundaries that the operational reality crosses every day.