Everything here is built from public sources: vendor websites, press releases, stock-exchange filings, official EU and US procurement documents, and government audit reports. Every factual claim carries a numbered footnote with a dated source. Where I give my own reading, I say so ("I read this as...") and I name what would prove it wrong.

The chapter follows a seven-part spine: the technology and mission (Playing Ground), the market structure (Terrain), the companies (Contenders), their strategic moves (Their Plays), the money (War Chests & Casualties), the patterns I see forming (Winning & Losing), and what to watch next (The Campaign Ahead).

Disclosure. The author is Head of Product (Fractional) at AXIA, an insider-risk / data-loss-prevention software vendor. AXIA does not compete in maritime domain awareness. This briefing uses only publicly available material and reflects the author's personal view, not AXIA's position. It is not investment advice.

Maritime Domain Awareness (MDA) is the discipline of building a wide-area, persistent understanding of what is happening on, under, and around the world's oceans — and, crucially, of finding the activity that someone is trying to hide. The International Maritime Organization defines it as "the effective understanding of anything associated with the maritime domain that could impact security, safety, the economy or the marine environment."¹ In plain terms: who is sailing where, are they who they claim to be, and is their behavior normal or suspicious — answered across an entire exclusive economic zone (the 200-nautical-mile band of sea a coastal state controls, abbreviated EEZ), a strategic chokepoint, or open ocean. The buyer is a navy, a coastguard, a border or fisheries-enforcement agency, a customs body, or a multilateral institution such as the EU's maritime-safety agency (EMSA) or border agency (Frontex). The product these buyers want is software that fuses many sensor feeds into one trustworthy picture and then flags the vessels behaving badly — the ones that have switched off their transponders ("gone dark") or broadcast a false identity ("spoofing").

The stack reads bottom-up: a cheap, ubiquitous cooperative signal (AIS) that adversaries can defeat, sitting beneath progressively more expensive non-cooperative layers (radio-frequency, radar, optical) whose entire purpose is to catch what the cooperative layer misses.

The scope boundary that governs this whole chapter. Four adjacent things get loosely called "maritime surveillance," and conflating them produces nonsense competitive analysis, because each has a different buyer, a different sales channel, and a different set of competitors. This chapter is about exactly one of them.

First, MDA is not Vessel Traffic Services (VTS). VTS is port- and waterway-scale traffic management — the maritime equivalent of air-traffic control — standardized under international instruments and sold to port authorities as shore-radar-plus-console systems by large engineering houses. Kongsberg Norcontrol, for instance, states it has "delivered over 250" such solutions "since delivering the world's first VTS system to Teesport in 1979."² VTS watches vessels that want to be seen, in a small area, under safety regulation. MDA watches vessels that do not want to be seen, across a vast area, for security. A single company can sell both, but they are different products bought by different people. I read VTS incumbency as one of the ways large defense and engineering firms claim an "MDA" label without actually competing in the analytics layer this chapter is about — a distinction worth holding onto when the bundling question (§N.2) comes up.

Second, MDA is not maritime patrol aircraft, drones, or coastal-radar hardware. Those are platform-and-sensor programs — Thales, for example, launched a "full turnkey solution" called CoastShield in November 2024 that integrates sensors into a command system.³ The MDA software may run on top of such platforms, but the platform itself is a separate category bought as capital equipment or as a flying-hours service.

Third, MDA is not commodity trade-and-compliance maritime intelligence. This is the commercial twin of national-security MDA: the same raw data (vessel transponder signals, satellite imagery) and the same core analytics (dark-activity detection, spoofing detection), sold to banks, insurers, and commodity traders for sanctions screening and freight analytics rather than to governments for enforcement. Several vendors — Windward, Kpler, Spire's maritime heritage — straddle both, precisely because the technical stack is shared even though the procurement stack is not. That straddle is a recurring theme of the chapter, but the commercial-compliance buyer is adjacent terrain, mapped and not contested.

The rule, then: a "contender" in this chapter competes in the analytics-and-fusion software layer for government MDA missions, or supplies the sensor data that layer consumes. Traffic-control consoles, patrol aircraft, and pure commodity-trading analytics are mapped around the edges but are not the battlefield.

The sensor stack, and why it has the shape it has. Every layer of the MDA stack exists to compensate for the failure of the layer beneath it. The foundation is the Automatic Identification System (AIS) — a transponder that broadcasts a ship's identity, position, course, and speed. International safety law (SOLAS Chapter V) mandates AIS aboard essentially all large vessels on international voyages and requires it be kept "in operation at all times" except where protective rules allow otherwise.⁴ That mandate is what makes AIS the cheapest wide-area picture of the seas that has ever existed: most ships announce themselves, continuously, for free. But the same mandate carries an escape clause, and a transponder is trivial to switch off or to feed false data — which is exactly what sanctioned and criminal vessels do. Lloyd's List Intelligence documents "the evolving practice of AIS spoofing in vessel identity manipulation" in the shadow fleet,⁵ and the investigative outlet Follow the Money reports the Russia-linked shadow fleet "goes darker" with deliberate transponder switch-offs.⁶ So the defining problem of modern MDA is the dark-and-spoofed-vessel problem — and AIS being mandated, ubiquitous, cheap, and adversarially unreliable is the single fact from which every other layer of the stack follows.

Above AIS sits its closed governmental twin, LRIT (Long-Range Identification and Tracking), which reports ship identity, position, and time roughly four times a day to "entitled users" — flag, port, and coastal states — under an IMO-administered distribution plan, with EMSA running the EU's data centre.⁷ LRIT matters for a structural reason that shapes the competitive landscape: it is government-access-only data. Commercial analytics vendors cannot resell it, so they build on AIS plus commercial satellite feeds, while a government fusion centre holds an extra feed no vendor can offer in its cloud product. I read this as a quiet asymmetry that favors on-premises and sovereign deployment models — the government can fuse LRIT into a vendor's deployed platform, but the vendor can never put it in a software-as-a-service (SaaS) catalogue. If that reading is right, we should expect government-facing MDA vendors to invest disproportionately in deployable, customer-hosted versions of their software; if vendors instead converge on pure cloud delivery, the reading is wrong.

Then come the non-cooperative layers, each catching a different slice of what AIS misses. Satellite radio-frequency (RF) geolocation finds emitters that have turned their transponders off: HawkEye 360 operates a commercial constellation that detects and geolocates RF signals and markets it explicitly to "analyze AIS gaps for ships at sea" for "coast guards, navies, law enforcement, fisheries,"⁸ while France's Unseenlabs runs a similar constellation serving European maritime-intelligence and defence clients.⁹ RF is the classic "tip" — it tells you something is transmitting where no ship is declared, with a rough position. Synthetic Aperture Radar (SAR) then images the hull through clouds and darkness: ICEYE markets its SAR constellation directly for MDA and dark-vessel detection,¹⁰ and EMSA's CleanSeaNet service processes more than 3,000 SAR images a year and delivers them to 34 coastal states "in less than thirty minutes after the acquisition."¹¹ SAR is the classic "confirm." Optical (electro-optical, or EO) imaging closes the loop with high-resolution identification once a target is localized — but it is daylight- and weather-limited, which is why it sits at the end of the chain, not the start. Finally, coastal radar and shore-sensor chains form the sovereign perimeter layer (Thales markets a coastal radar able to "simultaneously detect and track up to 1,000 targets"¹²), and uncrewed surface vehicles — Saildrone markets its sea-going drones explicitly for MDA¹³ — provide persistent on-water presence, blurring the line between a sensor and a platform.

The technical logic of all this is a workflow called tip-and-cue: a cheap, wide, coarse sensor (RF) tips the system that something is there, and a narrow, expensive, precise sensor (SAR, then optical) is cued to confirm and identify it. The whole non-cooperative stack is one long tip-and-cue chain compensating for the unreliability of the cooperative signal at its base.

The layer this chapter actually contests: fusion and analytics. Sitting above all the sensors is the software that turns feeds into decisions. Its functions are now well-defined in public product material: dark-vessel detection (correlating an RF tip, a SAR detection, and an AIS absence — Starboard Maritime Intelligence publishes its method for "satellite dark vessel detection for maritime domain awareness"¹⁴); anomaly and behavior detection (flagging loitering, ship-to-ship transfers, route deviation — Windward says its platform "processes and fuses vast volumes of maritime data — including vessel behavior, trade patterns, ownership structures, and risk indicators"¹⁵); identity and spoofing detection (catching impossible movements and manipulated identities⁵); tip-and-cue orchestration (Skylight documents "Tip and Cue in Action" workflows and announced "the first real-time system for detecting ships in public optical imagery globally"¹⁶); and predictive analytics (destination and intent prediction¹⁷). On that last one a word of caution: "predictive" is heavily marketed, but how much of it goes beyond estimated-arrival-time inference is not publicly verifiable, and a careful reader should treat predictive claims as positioning language unless a vendor ties them to a published, testable capability.

The competitively decisive fact about this layer is that a free, capable version of it already exists, supplied by governments and philanthropy. SeaVision — "a Google Maps based marine vessel visualization tool, developed by the United States Department of Transportation and the Volpe Center" — is provided by the US government to "hundreds of government agencies globally."¹⁸ Skylight, built by the Allen Institute for AI, is offered "at no cost to national and regional agencies fighting maritime crime" and integrates with SeaVision and other government tools.¹⁹ The US Defense Innovation Unit partnered with Global Fishing Watch and the Coast Guard against illegal fishing and ran an open challenge (xView3) to open-source the underlying dark-vessel detection algorithms.²⁰ I read this as the most important shaping force in the market: free, government-furnished and philanthropic fusion layers compress the price floor for the "basic picture" to zero, which pushes commercial vendors up-stack toward things the free layer cannot do — proprietary sensor data, deeper analytics, classified-side fusion, and accredited sovereign deployment. That pressure is what produces the structural barriers below. If the free tiers were withdrawn or stagnated, commercial vendors could compete lower in the stack; as long as they persist and improve, the only defensible commercial ground is the gated territory.

Q2 — the structural barriers to entry. This is the heart of what a software specialist needs to understand before pitching a government vessel-tracking or border-monitoring contract. The barriers below are each built from quoted requirement lines in official solicitations, tenders, and audit reports. Where the public record goes thin, that thinness is itself a finding, and I flag it rather than paper over it. Anything beyond the quoted evidence I label as structural inference.

Barrier 1 — security clearances and accredited facilities. This is the hardest-edged and most literally quotable barrier. Frontex's aerial-surveillance terms of reference require bidders to hold an "EU MS security clearance certificate" showing access to information classified "CONFIDENTIEL UE/EU CONFIDENTIAL or above," and makes that clearance a bid-time condition: the certificates "shall be valid at the moment of submission of offers."²¹ On the NATO side, UK government guidance to industry is blunter still: bidders "may be required to handle and store classified information up to the level of NATO SECRET, and only companies maintaining cleared facilities and appropriate personnel clearances will be able to perform the resulting contract."²² Why this is structural rather than procedural: facility clearances are issued by national authorities to entities established in that country, and personnel clearances chain to nationality and residency. A foreign software vendor cannot buy its way in quickly. The realistic entry routes are (a) years-long establishment of a cleared local subsidiary, (b) teaming under a cleared prime contractor, or (c) acquiring a firm that is already cleared.

Barrier 2 — sovereign hosting and data residency. The Frontex terms of reference contain an unusually candid statement of where the buyer is heading: as Frontex moves "toward processing classified data and personal data," it "may lead Frontex to a need for taking over or building its own infrastructure to fully control processing and storage of the surveillance data," with the contractor obliged to migrate accordingly on request.²³ The EU's cross-sector data-sharing environment (CISE) is itself built as "a decentralized infrastructure" with "no central storage of information."²⁴ For a cloud-native analytics vendor, sovereign or on-premises deployment is not a checkbox — it forks the engineering roadmap toward air-gapped model updates, no telemetry home, and in-country support staff who themselves need clearances under Barrier 1. I read this as the single most product-shaping barrier for software-only entrants, because it converts a clean multi-tenant cost structure into a per-country delivery business. The clearest public signal that this gate is real: Windward, a specialist, acquired a cleared US defense-tech firm (Prominent Edge, announced April 2026, holding "ISO 27001, ISO 9001, and CMMI Dev Level 3 certifications") explicitly "to deepen U.S. defense and national security capabilities."²⁵ A SaaS motion alone does not clear the deployment gate; a specialist bought its way through it.

Barrier 3 — export control and remote-sensing licensing. This binds the data layer and flows downstream to the analytics vendors that depend on it. US commercial satellite operators are licensed by NOAA, and under 2020 rules the most capable SAR operators faced "limited-operations directives" capping the resolution they could sell to non-US-government buyers "at the insistence of the Defense Department" — restrictions NOAA removed in 2023.²⁶ The relaxation moved the line, but the residual structure persists by design: government customers can obtain capabilities that commercial customers cannot, which advantages vendors with established government channels in each jurisdiction. I could not locate a quotable public license condition for European RF-sensing data in this research pass — a thin spot I note below rather than fill with inference.

Barrier 4 — certification and cyber accreditation. The Frontex aerial terms make equipment certification ("according to EASA and competent authorities' prescriptions") a condition of even performing the service.²⁷ On the software side, the equivalent is the delivery organization's credentials — exactly the ISO and maturity-model certifications Windward cited as its rationale for the Prominent Edge acquisition.²⁵ On the US side, cloud software sold into federal missions runs an accreditation path (FedRAMP authorization, Defense Department impact levels, and CMMC for the defense supply base). I could not extract verbatim accreditation lines from a live US MDA solicitation in this pass — the relevant opportunity (a Navy "Data and Analytics for Maritime Domain Awareness" solicitation, live with a January 2027 close date) renders behind dynamic pages that defeat public retrieval.²⁸ That is a named thin spot, not evidence the requirement is absent. The general magnitude — an accreditation path measured in quarters-to-years — is widely understood, but I flag that magnitude as inference rather than quoted fact.

Barrier 5 — integration with national command-and-control. A standalone picture is demo-ware in this market; the software has to plug into the buyer's operational systems. The EU's own auditor documents how hard that is: the European Court of Auditors found that Frontex's near-real-time situational awareness "is hampered by weaknesses in data quality," producing "a patchwork of national situational pictures rather than a single, uniform European situational picture," with member states submitting reports "in different formats" and some updating "only once a week."²⁹ The US analog is just as old: the Government Accountability Office has reported for more than a decade on the Coast Guard's difficulties "in developing and implementing" a common operational picture.³⁰ Fragmentation here is a barrier, not an opening, because each national integration is bespoke, relationship-mediated, and clearance-gated — so integration effort scales linearly with the number of countries served, favoring incumbents who have already paid those per-country costs.

Barrier 6 — framework-contract lock-in. The dominant European contracting vehicle is the multi-year framework contract, and its mechanics are visible in public awards. EMSA has run successive roughly four-year satellite-AIS data contracts — to LuxSpace with ORBCOMM, then to Spire Global ("two framework contracts... with a maximum overall budget of €8.4 million" over "a four-year period"), with exactEarth and Hisdesat holding an earlier one.³¹ Frontex's aerial framework operates by "Specific Contracts with re-opening of competition" among framework holders only — meaning once the framework closes, non-holders are locked out of every task order for its life.²¹ Stated plainly: a four-year framework means one genuine competitive window every four years per agency per service line. Miss it, and the next real entry point is years away. Combined with Barrier 1's requirement that clearances be valid at offer submission, the system rewards firms that maintain standing cleared capacity between competitions — a fixed cost small software entrants cannot easily carry.

Barrier 7 — proof of live capability at bid time. The EMSA satellite-AIS awards followed "a competitive tender process, which included both a full technical and commercial evaluation"³¹ for a service that includes real-time feeds — meaning only an operator with a constellation already on orbit at bid time can credibly tender. The Frontex methodology similarly scores demonstrated capability margins (awarding graduated points for exceeding minimum endurance requirements²¹). And the US Defense Innovation Unit's commercial-solutions process makes live demonstration the selection mechanism itself.²⁰ You must possess the capability, not promise it.

Barrier 8 — the free-tier floor. The least obvious barrier is the one already introduced above: the entry-level of this market is supplied by governments and philanthropy (SeaVision, Skylight, Global Fishing Watch¹⁸¹⁹²⁰). For a new commercial entrant, the addressable problem is therefore not "agencies lack a maritime picture" — it is "agencies need what the free picture cannot do": sovereign deployment, classified-side fusion (including LRIT and national sensors), accredited integration, contractual service-level agreements, and liability someone will stand behind. That compresses the entry wedge to exactly the territory gated by Barriers 1 through 6.

I read these eight barriers as compounding rather than adding. Clearances are per-jurisdiction; sovereign hosting is per-jurisdiction; C2 integration is per-jurisdiction; frameworks are per-agency-per-cycle. A software specialist therefore faces a market that is structurally not one market but dozens of small, slow, gated national ones — while the satellite data layer beneath it and the free layer beside it are global. The adaptive strategies that follow from this, each anchored to a cited event above, are: acquire cleared delivery capacity (Windward buying Prominent Edge), win the data-layer frameworks where owning a constellation is the differentiator (Spire and LuxSpace at EMSA), subcontract as the analytics layer under a prime, or retreat to the commercial twin market (sanctions compliance) where none of Barriers 1 through 6 apply and revenue scales like ordinary enterprise software.

Where the public record goes thin. For a consulting-grade view, the honest gaps matter as much as the findings, because they are exactly what desk research cannot supply. Evaluation weightings inside MDA software tenders are not public — aerial-service tenders publish scoring grids, but analytics procurements run under negotiated procedures whose matrices stay private. Classified annexes on both the EU and US sides keep the operationally interesting requirements (fusion with national intelligence feeds, classified-side performance) entirely out of public view. US solicitation text sits behind portal rendering that defeats public retrieval. Whether and how analytics specialists ride inside prime-led national programs is largely invisible in award notices, which name only the primes. And there is no public dataset on how often MDA framework incumbents are displaced at re-compete — the satellite-AIS succession shows displacement happens at the data layer, but nothing equivalent is public for the analytics layer.

If §N.1 is the technology and the mission, §N.2 is the market: who buys, why they are buying more right now, and through what machinery the money actually moves. The single most decision-relevant finding here is that government MDA is bought through two structurally different procurement models, and how accessible the market is to a new software provider depends entirely on which model a given deal runs through.

The two doors into the market: a shorter-cycle, recompeted services-and-data framework that a software specialist can realistically enter, versus a decadal national-system program effectively gated to sovereign integrators or firms with export-credit financing.

Five buyer segments that share a sensor stack but not a wallet. MDA demand splits into five distinguishable segments. They use the same underlying sensors and data, but they are bought by different institutions through different mechanisms — and a vendor needs to know which one it is selling into, because the procurement rules change completely between them.

Naval and defense MDA is bought by navies, defense ministries, and alliance commands. The notable 2026 pattern is that navies increasingly buy MDA as a renewable commercial subscription rather than as a one-time system acquisition: the US Navy renewed its HawkEye 360 contract "for a fourth consecutive year" under the Quad's Indo-Pacific Partnership for Maritime Domain Awareness, buying commercial RF data and analytics for dark-vessel detection across the Pacific.³² NATO's Baltic Sentry activity, launched 14 January 2025, deployed "frigates, maritime patrol aircraft, naval drones" plus surveillance technology to protect undersea infrastructure,³³ and NATO ran an unmanned-systems demonstration with Saildrone vehicles in the Baltic in June 2025.³⁴ I read the subscription pattern as the structurally important one: a navy treating MDA as an operating expense it can switch is a very different customer from a navy buying a capital system it will own for twenty years.

Coastguard and border surveillance is bought by coast guards and border agencies — Frontex and its national equivalents — through EU agency frameworks and national tenders. Frontex's aerial-surveillance services feed the EUROSUR border system, whose fusion services include "automated vessel tracking and detection capabilities."³⁵ Two illustrative deals: Indonesia's coast guard began a EUR 167 million national maritime-security system with the UK-listed firm SRT Marine in 2025, financed through a UK-export-credit-backed inter-government loan;³⁶ and Australia runs what is described as "the world's largest outsourced civil maritime surveillance operation," patrolling an 8.2-million-square-kilometre EEZ — a contract currently held by Leidos and transitioning to a new provider (Matrea) from 1 January 2028.³⁷ That transition is a clean example of a recompete that actually changes hands.

Fisheries and EEZ enforcement — the fight against illegal, unreported, and unregulated (IUU) fishing — is bought by fisheries-control agencies, and is distinctive for a feature found nowhere else: philanthropically subsidized platforms supplied to developing coastal states at no cost. The regulatory engine here is concrete and dated: Regulation (EU) 2023/2842 makes vessel tracking compulsory across the entire EU fishing fleet, phasing in all under-12-metre vessels by January 2028.³⁸ That mandate mechanically grows the population of tracked vessels. Against that, NOAA runs an IUU-detection pilot using Skylight, and Global Fishing Watch signed an agreement to give "select governments and agencies... no-cost access to vessel tracking data and analysis tools."³⁹ I read the free tier as a structural feature of this segment: it caps the commercially addressable price point for analytics-only offerings, while expanding the base of agencies habituated to satellite-analytics workflows — a double-edged dynamic for any vendor.

Port and critical-infrastructure protection is the least-settled segment, bought by navies, coast guards, and increasingly energy and telecom operators acting on critical-infrastructure mandates. The forcing event is the wave of undersea-cable damage in the Baltic: at least eleven cables damaged in fifteen months,⁴⁰ precipitating NATO's Baltic Sentry, whose stated method is unmistakably MDA-shaped — building "patterns of life" and watching for anomalies "like ships frequently changing direction, loitering or slowing near critical cabling," with reaction targets "within a half an hour or an hour."⁴¹ The revised EU Maritime Security Strategy (approved October 2023) explicitly elevated protection of "gas pipelines, undersea cables, ports, offshore energy facilities and LNG terminals,"⁴² and Denmark fielded Saildrone surveillance drones in its waters from June 2025.⁴³

Commercial maritime-intelligence adjacency is the fifth segment — marine insurers, commodity traders, banks, and shipping companies buying the same analytics as SaaS subscriptions with no tender process at all. Kpler (which acquired MarineTraffic) positions its risk product for "compliance, risk, legal, trade finance, insurance, energy, and trading teams,"⁴⁴ and Windward frames 2026 demand around "geopolitical tensions, shadow fleets, and surging sanctions."⁴⁵ This segment matters to the chapter because the same vendors straddle it and government MDA — the commercial side provides the recurring-revenue base from which several specialists fund their attacks on government tenders.

Why demand is rising — the forcing functions. Rather than lead with a market-size number (the sizing is treated honestly below, and it is soft), the more reliable way to understand this market is through the dated policy and incident drivers pushing money into it. Each is a forcing function with a legal instrument or a counted event behind it.

The first is post-2022 sanctions and shadow-fleet enforcement. By December 2025 the US, UK, EU, Canada, Australia and New Zealand had collectively designated 621 unique oil tankers, and the shadow fleet "kept on growing" through 2025;⁴⁶ the EU alone added 41 vessels to its list on a single day in December 2025.⁴⁷ Around 60% of Russia's seaborne oil exports transit the Baltic and the Danish Straits,⁴⁸ concentrating enforcement demand in a small, heavily-watched sea space. This drives MDA spend directly, because a designation list is only enforceable if a state can identify, track, and attribute vessels that spoof, go dark, and swap flags — which is precisely the dark-vessel-detection product category.

The second is the growth of the dark fleet itself. Estimates from named sources diverge but agree on direction: Windward identified more than 1,900 dark-fleet vessels as of Q3 2025,⁴⁹ the tanker broker Clarksons estimated around 1,500,⁵⁰ and Ukraine's government catalogue listed 1,337 ships as of February 2026.⁴⁹ I treat the spread — roughly 1,300 to 1,900 depending on who is counting and how they define it — as itself a market signal: there is no authoritative registry, and that gap is exactly what commercial maritime-intelligence vendors monetize.

The third is IUU fishing enforcement programs (the NOAA-Skylight pilot and the Global Fishing Watch no-cost access agreement above³⁹), institutionalized further by the Quad's Indo-Pacific partnership, which supplies partner nations with near-real-time dark-vessel detection (SeaVision software plus HawkEye 360 RF data).⁵¹ The fourth is carriage mandates as a regulatory floor — the IMO's global AIS and LRIT requirements plus the EU's 2023 fisheries-tracking expansion³⁸ — every extension of which enlarges both the data exhaust to be fused and the baseline against which "dark" behavior is defined. The fifth is undersea-infrastructure protection after Nord Stream and the Baltic cable incidents, a tightly-dated sequence running from the cable damage through Baltic Sentry to the Danish drone deployment, with the revised EMSA mandate (provisionally agreed May 2025) explicitly citing "Russia's war of aggression against Ukraine, its shadow fleet, suspicious ship-to-ship transfers and the deactivation of shipborne automatic identification systems."⁵²

The sixth driver is the hardest currency of all: named program budgets. These are worth more than any market-size estimate because each is a dated, sourced commitment. EMSA awarded Airbus a EUR 30 million framework for drone surveillance services (two years plus options, operating from 2026)⁵³ and a multi-year contract to Schiebel for its CAMCOPTER drone.⁵⁴ Frontex's aerial-surveillance rounds have run from roughly EUR 84.5 million (2021) to a smaller EUR 14-million-plus-option round (2023), with an earlier EUR 50 million drone framework — though these specific Frontex values come via campaign-organization republication of public award data and should be verified against the official tender database before being quoted as authoritative.⁵⁵ On the US side, the Coast Guard received a roughly $25 billion FY2025 supplemental for acquisitions and modernization, and a separate $170 million line specifically for "improving maritime domain awareness, including in the cyber domain."⁵⁶ India's navy contracted its state-owned champion BEL in July 2025 to upgrade its national MDA network with "AI-enabled software" (value not publicly disclosed),⁵⁷ and Indonesia's EUR 167 million Bakamla system runs as a two-year build plus eight-year support engagement.⁵⁸

The two procurement models — the core finding. Now the decision-relevant structure. Government MDA is bought two fundamentally different ways, and a new software provider's odds differ radically between them.

Model A — surveillance-as-a-service frameworks. EU agencies and several Anglosphere governments buy MDA capability hours and data, not systems. EMSA awards framework contracts for drone flight services (Airbus and Schiebel, both 2025), for SAR satellite imagery (KSAT, a 24-month framework renewable twice⁵⁹), and for pollution and ship-detection services (CLS, renewed since 2014⁶⁰) — with all outputs streaming to EMSA's own data centres and served to member states as integrated services. Frontex charters its aerial surveillance on 24-month auto-renewable frameworks with guaranteed minimum flight hours. Australia outsources civil surveillance wholesale. The defining properties of Model A: two-to-four-year durations, periodic recompete with real incumbent displacement (Leidos to Matrea in Australia; Airbus joining EMSA's drone roster in 2025), price discovery through public tender databases, and — critically — a software or data specialist can enter as the analytics or sensor-data layer without owning aircraft. Incumbency is real here but bounded; it is the realistic door for a specialist.

Model B — national system procurements. Nation-state MDA systems — fusion centres, national networks, coastal sensor chains — are bought as defense capital programs, and they behave completely differently. India routed its national MDA project to its state-owned integrator BEL,⁵⁷ which I read as effectively closed to foreign primes except as technology subcontractors; the channel is the national defense-industrial base. Indonesia shows the alternative pattern for a state without a domestic champion: a foreign mid-cap (SRT Marine) won a decade-shaped engagement, but only because it could arrange export-credit-agency financing⁵⁸ — meaning financing capacity, not software capability, is the gating requirement in this channel. At the EU member-state level, the European Court of Auditors found EUROSUR implementation uneven: of 20 member states consulted, 15 said it "did not provide all necessary information," and some upload data "only once a week."⁶¹ And in the US, the GAO reports the Coast Guard faces "persistent and longstanding challenges managing its planned $40 billion acquisition programs."⁶² Model B's properties: multi-year competitions, sovereign and clearance requirements, national-champion preference where one exists, export-credit financing as the wedge where one does not, and roughly ten-year incumbency once installed. Contract durations here are effectively decadal.

Three cross-cutting observations follow. First, the services-versus-systems divide is the deepest structural fact in the market, and the same agency often buys both ways (EMSA buys drone services and builds its own data centres). For a new software provider, Model A frameworks and data-layer subscriptions are the realistic entry door; Model B typically requires a prime partner or financing partner. Second, EU buyer fragmentation is institutionalized — EMSA (safety and surveillance), Frontex (borders), and EFCA (fisheries) cooperate but do not share data naturally, which is why the cross-sector CISE environment had to be built. A vendor can sell adjacent slices of the same picture to three different EU agencies. Third, the audit offices keep documenting that capability gaps survive even large appropriations (EUROSUR gaps despite years of revisions; Coast Guard shortfalls despite a ~$25 billion supplemental), which I read as evidence of sustained, multi-cycle demand rather than a one-off procurement wave.

Market sizing — handled honestly, because the public numbers do not deserve more. No authoritative, methodology-disclosed market size for "MDA" exists in the public domain. The figures below come from commercial research-mill publishers and are directional only; the variation between them is so wide that the spread is itself the finding. Dataintelo's "maritime domain awareness market" page puts it at $22.4 billion in 2024 growing to $44.6 billion by 2033;⁶³ Grand View Research's "maritime surveillance market" page gives $24.3 billion in 2024 rising to $35.8 billion by 2030;⁶⁴ Mordor Intelligence's narrower "maritime situational awareness systems" page gives just $2.34 billion in 2025;⁶⁵ and two "MDA-AI" figures from Dataintelo and Growth Market Reports land between roughly $1.9 billion and $3.8 billion depending on the publisher.⁶⁶

My read, flagged explicitly as interpretation: the $22–24 billion "MDA / maritime surveillance" cluster (Dataintelo, Grand View) is platforms-inclusive — it mostly counts aircraft, radar, and prime-contractor revenue. The software-and-analytics-addressable slice is an order of magnitude smaller — roughly $2–4 billion globally circa 2025 if the Mordor "systems" figure and the two "MDA-AI" figures bracket it. So the honest sizing statement for a software-specialist lens is: a low-single-digit-billions global software and analytics opportunity, growing at high-single to low-double digits annually, sitting inside a roughly $20-billion-plus platforms-inclusive envelope. None of these numbers should be quoted as authoritative. The named program budgets in the driver section above are the harder currency, and I would lead with those in any serious conversation.

What the terrain implies. Five reads, each with the condition that would prove it wrong. (1) Demand is structural, not cyclical — the three big forcing functions all have dated legal instruments behind them; this is wrong if EMSA and Frontex framework renewals and MDA-relevant award notices slow materially through 2027. (2) The accessible market for software specialists is the services-and-data layer, not the national system — Model A shows repeated entry events for non-primes while Model B is gated by sovereign integrators or financing; a pure-software entrant winning a national MDA prime contract without a financing or integrator partner by 2028 would falsify this. (3) Europe — the Baltic specifically — is the demand epicenter right now, while the Indo-Pacific is the volume play; track whether Baltic-validated vendors announce Indo-Pacific or Gulf national contracts in 2026–2027. (4) Free and subsidized supply is a genuine competitive feature of the enforcement segments, so commercial vendors must differentiate above that floor (fusion depth, sovereign deployment, C2 integration, compliance-grade evidence chains) or sell to segments the free tiers do not reach (navies, sanctions compliance, insurers). (5) Fragmented buyers mean repeatable sales for multi-agency vendors and hostile terrain for anyone whose model assumes a single consolidated buyer.

One note on a question that hangs over this market — whether the Tier-1 defense primes that bundle MDA into larger system sales are crowding out civilian-focused specialists in international tenders. The terrain evidence here is genuinely mixed and partly invisible: prime-led national programs (India's BEL, the turnkey coastal systems) do bundle the analytics layer into a larger scope, but the recompeted service frameworks (Model A) repeatedly admit non-prime specialists. Whether bundling is increasing in international tenders is an open empirical question the public award record does not yet settle, and I would treat it as a hypothesis to test against named tender outcomes rather than an established fact.

The maritime domain awareness "market" is not one market. It is a four-layer value chain, and the buying question is different at each layer. At the bottom sit the data and sensing providers — companies that own a way of seeing ships: satellite radio-frequency geolocation, synthetic-aperture radar, electro-optical imagery, terrestrial and satellite AIS networks. Above them sit the analytics and fusion software companies — the ones who take those raw detection layers and turn them into an intelligence picture: dark-vessel detection, behavioral anomaly, identity-spoofing, risk scoring. Above those sit the systems integrators and command-and-control (C2) builders — the firms that deliver a whole sovereign coastal-surveillance system, shore radar plus cameras plus AIS plus comms wired into a national operations center. And alongside all of them sit the platform providers — the people who sell the aircraft, drones, and uncrewed surface vessels (and the surveillance service flown or sailed on them). A vendor's real competitive identity comes from which layer it lives on, and the most interesting players are the ones that straddle two. The eight contenders profiled below are grouped into a Gravity tier (the scale players and the two prime exemplars), an Attention tier (specialists with real, growing programs), a wider prime class treated at the level of the class, and a set of wildcards described but not pattern-claimed.

Author's read of public material, June 2026. Conceptual, not data-derived. Not investment advice.

"Windward is the leading Maritime AI™ company delivering a unified operational picture through multi-source fusion" — Windward homepage, accessed 2026-06-12 ⁶⁷

Windward is the reference pure-play of maritime-intelligence analytics, and the most striking thing about it in 2026 is how defense-forward its public language has become for a company that built its actual revenue base in commercial risk and compliance. The homepage leads with the tagline "MISSION-GRADE INTELLIGENCE" and claims the product "fuses best-in-class premium and unrestricted data feeds — AIS, dark vessel signals, EO, SAR, RF, and more" ⁶⁷, while its dedicated government surface promises "Predictive intelligence for national security" ⁶⁸. That vocabulary — "mission-grade," "unified operational picture" — is borrowed from the defense common-operating-picture lexicon, not the commercial-shipping one, and I read the shift as deliberate. On the six-axis read, Windward owns no sensors of its own; it is a fusion layer that pulls third-party feeds (RF historically through a HawkEye 360 partnership announced 2019-05-22 ⁶⁹, electro-optical monitoring through a Vantor partnership announced 2026-04-16 ⁷⁰), runs dark-vessel, behavioral-anomaly and spoofing analytics on top, ships as SaaS and API rather than as an integrated C2 system, and sells across the straddle — navy and intelligence on one side, coastguard and border on another, and the commercial insurers, traders and charterers it grew up serving. It was taken private by FTV Capital at roughly $280M in March 2025 ⁷¹, delisting from the London exchange; the published-material footprint is heavy (vendor site, partnership pressers, named-outlet deal coverage). Windward's wedge is fusion breadth and a category-defining "Maritime AI" brand; what it does not have is its own collection.

Kpler

"Maritime intelligence for defense and government agencies" — Kpler government page, accessed 2026-06-12 ⁷²

Kpler is the scale player on the data side, and it got there by acquisition. It is at heart a commercial commodities-and-trade-intelligence company — the homepage sells "Global trade intelligence at your fingertips," "built on leading maritime intelligence from MarineTraffic" ⁷³ — but it has bought its way into owning the widest unclassified maritime dataset in the field and now runs an explicit government motion on top of it. The government surface is unambiguous: it offers users the ability to "Detect, monitor, and assess vessel activity and trade networks in real time to secure borders, enforce maritime law, and counter illicit activity," and draws the distinction that "Unlike AIS-only solutions, Kpler integrates vessel behavior, cargo flows, trade intelligence, and entity datasets" ⁷². On the six-axis read, Kpler is a hybrid — mostly an aggregator, but one that owns the MarineTraffic terrestrial-AIS receiver network (acquired 2023) and acquired Spire's satellite-AIS maritime data business (closed 2025-04-25 at $233.5M ⁷⁴) — giving it AIS at enormous scale, advertised as "300K+ Vessels tracked / day" and "1B+ AIS signals / day" ⁷³. Its analytics lean toward dark-fleet and sanctions behavior, entity-and-ownership networks, and trade-flow patterns; it delivers as a data feed and API plus SaaS, with an Esri Velocity integration for government geospatial workflows ⁷⁵; it is a data layer feeding other people's pictures rather than a C2 vendor; and its customer base is commercial-first but with a fast-growing defense and government practice. My read: Kpler's wedge against an analytics player like Windward is breadth of data — cargo, customs, ownership graphs — rather than depth of analytics. It is selling governments "the unclassified data layer."

"CoastShield is a modular and scalable advanced integrated Coastal Surveillance System (CSS) designed to monitor and detect coastal activities in real-time, from the shore to the open sea, as well as in low altitude airspace" — Thales CoastShield catalogue page, accessed 2026-06-12 ⁷⁶

Thales is the first of two exemplars of how a Tier-1 defense prime shows up in MDA, and it illustrates the productized version of the prime offer: not a one-off bespoke program, but a named, catalogued, modular system. CoastShield integrates Thales's own sensors — the Coast Watcher radar family, electro-optical cameras — together with third-party feeds and AIS into a customer-tailored C2 system, with AI-assisted confirmation of abnormal behavior built in ⁷⁶. (The CoastShield page is bot-gated to direct fetch, so the verbatim pillar above was captured via a search index rather than a live page load; this is noted for audit honesty.) On the six-axis read, Thales owns its sensor portfolio and integrates others' on top; it covers coastal radar plus electro-optical plus AIS and even a low-altitude air picture; analytics is a feature embedded inside the system rather than a product sold on its own; delivery is turnkey on-premise sovereign systems; and the system is the command-and-control center. The customer class is navies, coastguards and border agencies. I read the prime model here as sensor-pull: the radar and camera portfolio is the moat, the C2 software is the lock-in, and analytics rides along as a system feature — the exact inverse of the specialist model, where analytics is the product and sensors are borrowed.

Saab

"Saab maritime traffic management leads the way in providing safe, efficient and easily configurable solutions for marine geomatics, navigation, search & rescue, coastal surveillance, port security and traffic management" — Saab press release, 2017-10-11 ⁷⁷

Saab is the second prime exemplar, and it represents a different path into the same tenders: not defense-sensor pull, but civil maritime-traffic heritage scaled up into coastal surveillance. In its Queensland award release Saab describes a system "powered by Saab's V3000 vessel traffic services system" delivering "marine geomatics, navigation, search & rescue, coastal surveillance, port security and traffic management" ⁷⁷. Its vessel-traffic-services installed base is decades deep and globally significant — the Scheldt Radar Network, which Saab describes as the largest VTS in the world ⁷⁸, and the Shanghai VTS ⁷⁹ among them. On the six-axis read, Saab is an integrator with its own software products (V3000 at the core) and integrated radar networks; it covers shore radar plus AIS plus electro-optical at the system level; its analytics center on traffic management and the safety-and-security picture, with defense-specific MDA living in a separate Saab portfolio; it delivers turnkey on-premise systems with long-term software support; and it is the C2 and VTS center. Its customer class is port and maritime authorities first, coastguards second, defense through other product lines. I read Saab's position as the "civil-authority incumbent" play: decades of port and VTS relationships give it the procurement track record civilian maritime authorities require — a different moat from Thales's defense-sensor pull, but one that lands in the same Layer-3 tenders.

"the global leader in signals intelligence data and analytics" — HawkEye 360 release via PR Newswire, 2025-12-09 ⁸⁰

HawkEye 360 is the data-and-sensing specialist that turned radio-frequency geolocation into the recognized answer to the dark-vessel problem — the problem of a ship that has switched off its AIS transponder to disappear. In its 2025 US Navy renewal release the company explains that "HawkEye 360's constellation detects, characterizes, and geolocates RF signals from ships and other emitters worldwide," helping users "identify potential 'dark vessels,' uncover patterns of activity, and build a consistent, shared operating picture" ⁸⁰. The same release reports a $98.8M firm-fixed-price IDIQ extending US Navy access to its commercial RF data under the Indo-Pacific Partnership for Maritime Domain Awareness. On the six-axis read, HawkEye 360 is one of the few players that owns its own constellation (radio-frequency); it pairs that core RF layer with third-party AIS for association analytics; it does dark-vessel detection by matching RF emissions against AIS reports, plus pattern-of-activity work; it delivers as a data feed and analytics products fed into government tools and operating pictures rather than running its own C2; and its customer class is navy, defense and intelligence, with civilian-enforcement reach through partner programs. I read HawkEye 360 as the clearest proof that governments will buy commercial data layers à la carte — one sensor type at a time — rather than only as part of a prime-integrated system, which is structurally good news for every analytics specialist that fuses such data.

SRT Marine Systems

"The global leader in advanced maritime domain awareness technologies and solutions, from sophisticated national surveillance systems for coast guards to large scale fisheries monitoring" — SRT Marine homepage, accessed 2026-06-12 ⁸¹

SRT Marine Systems is the awkward, decision-relevant data point of this chapter: a UK small-cap specialist winning exactly the kind of sovereign turnkey national-MDA programs the value-chain map says ought to belong to the primes. Its site claims "A new generation of intelligent maritime domain awareness" and positions the firm as the global leader spanning national surveillance systems for coastguards through to large-scale fisheries monitoring ⁸¹, and the claims are backed by disclosed contracts: a US$180M national coastguard MDA system (regulatory news announcement, December 2023 ⁸²) and a US$213M new-MDA-system award (regulatory news announcement, October 2024 ⁸³), each structured as a roughly two-year build followed by multi-year support. On the six-axis read, SRT is a system integrator that owns its own AIS transceiver and product technology and procures radar and cameras into systems; it covers AIS (its own heritage) plus coastal radar plus electro-optical and infrared at the system level; its analytics center on fleet monitoring and a fisheries management system it markets as "Intelligent fisheries management system. Conserve, protect, sustain." ⁸¹; it delivers on-premise sovereign turnkey systems with multi-year support; it is the C2, delivering the national operations-center picture; and its customer class is coastguard, border and fisheries — the civilian-defense core. I read SRT as evidence that the systems-integrator layer is contestable by specialists in mid-tier sovereign markets — the Gulf, Southeast Asia — where the buyer wants a dedicated MDA system rather than a whole defense-platform ecosystem. The standard, opinion-signaled observation alongside it: a small-cap carrying a handful of nine-figure sovereign contracts has revenue-concentration exposure that a prime does not. That is a structural inference from the disclosed contract pattern, not a distress claim.

Saildrone

"Operating in the world's harshest environments, Saildrone delivers persistent, wide-area intelligence and payload effects to enable informed decisions across the spectrum of defense missions" — Saildrone homepage, accessed 2026-06-12 ⁸⁴

Saildrone is the platform-layer entrant that matters for the MDA software conversation precisely because it bundles the collection and the data service. Its current positioning is overtly defense-first — the tagline "Detect. Deter. Dominate." sits above the promise of "persistent, wide-area intelligence and payload effects" across the spectrum of defense missions ⁸⁴ — and its European push is documented in a tight sequence of dated releases: a Copenhagen-based European subsidiary (2025-04-09 ⁸⁵), a $60M financing led by Denmark's Export and Investment Fund "to bring Saildrone technology to Europe" (2025-05-13 ⁸⁶), and first operational testing with the Danish Armed Forces in the Baltic (2025-05-08 ⁸⁷). On the six-axis read, Saildrone owns both its platforms and its sensors — a fleet of uncrewed surface vessels carrying radar, electro-optical, AIS, and acoustic and subsea-mapping payloads; it generates radar, electro-optical and AIS detections from those surface platforms plus seabed and bathymetry data; it runs machine learning on its own collected data for threat detection and infrastructure protection; it delivers an intelligence service (platform-as-a-service) rather than pure SaaS; it feeds national and defense operating pictures; and its customers run from navies and coastguards through science agencies (a NOAA heritage) into NATO-Europe defense. I read Saildrone as selling "MDA-as-a-service" — government buyers get persistent collection without owning platforms, which competes less with software companies like Windward or Kpler and more with the manned-patrol and crewed-aircraft line items of a national budget.

Skylight (Allen Institute for AI)

"Skylight's technology is free to government, regional, and nongovernmental organizations worldwide that work to protect our seas from illegal fishing" — Skylight homepage, accessed 2026-06-12 ⁸⁸

Skylight is the price-floor anomaly of the entire field: philanthropically funded, free to enforcement users, and genuinely deployed at scale. Its site states that "Skylight provides AI-powered, real-time intelligence to stop illegal, unreported, and unregulated (IUU) fishing and protect our oceans," and the strategically decisive sentence is the business model itself — the technology is free to government, regional and nongovernmental organizations worldwide working against illegal fishing ⁸⁸. It remains a program of the Allen Institute for AI (Ai2) ⁸⁹, reported in use across more than 70 countries, and recently extended with "Shippy," a free conversational AI agent for maritime analysts ⁹⁰. On the six-axis read, Skylight is an aggregator drawing on public and third-party data (Sentinel-1 SAR, AIS); it fuses AIS plus SAR with vessel-detection machine learning; its analytics cover dark-vessel and suspicious-activity detection with behavioral models and now a conversational agent; it delivers as free SaaS to qualifying government and NGO users; it is a standalone alerting picture rather than an integrated C2; and its customer class is fisheries, exclusive-economic-zone monitoring and coastguards in developing coastal states, plus NGOs. I read Skylight as capping what commercial vendors can charge for baseline IUU and fisheries analytics in developing coastal states — the monetizable ground for specialists shifts upward, to sovereign deployment, classified integration, and richer fused data that a free Sentinel-1-plus-AIS service cannot match. (Global Fishing Watch, a nonprofit open vessel-tracking platform, sits alongside Skylight as the other free-to-enforcement option in this segment.)

Beyond Thales and Saab, the wider Tier-1 prime class — Airbus, RTX/Raytheon, Israel Aerospace Industries, Leonardo, Indra, Kongsberg, Elbit, HENSOLDT, Terma — participates in MDA through three cited motions rather than through standalone analytics products. The first is surveillance-as-a-service into civilian EU agencies: Airbus won a €30M EMSA framework for Flexrotor drone maritime-surveillance services ⁹¹, and Airbus Defence and Space Airborne Solutions, with Israel Aerospace Industries, signed a follow-on Frontex framework in December 2024 to operate the Maritime Heron over the Mediterranean for four years from 2025 ⁹². The second is turnkey national border and coastal systems: Indra built and exports Spain's SIVE external-surveillance system for the Guardia Civil, with the Tarragona deployment among its awards ⁹³ and exports including Latvia and Romania ⁹⁴, while Raytheon designed and equipped the Philippines' civilian-led National Coast Watch Center under US DTRA funding ⁹⁵. The third is sensor supply into others' systems — the radar and electro-optical portfolios of Terma, HENSOLDT and Thales feed both prime-led and specialist-led national systems; this last motion is a structural inference rather than an award-specific claim, and is labeled as such. (Worth naming separately: MDA Space of Canada — no relation to the MDA acronym — runs the Government of Canada's Dark Vessel Detection program for Fisheries and Oceans Canada, fusing RADARSAT-2 SAR with space-based RF to support IUU enforcement for partner states including Ecuador and the Galapagos ⁹⁶.)

A set of vendors round out the field and are noted descriptively, with no pattern claims attached. Starboard Maritime Intelligence (New Zealand) offers maritime-intelligence analytics with public-sector users in the Pacific. SynMax (US) does satellite-data analytics including maritime intelligence. Vake (Norway) is a satellite-based vessel-detection startup. Unseenlabs (France) operates its own RF-detection nanosatellite constellation for maritime surveillance — the European counterpart niche to HawkEye 360 — and ICEYE (Finland) operates a SAR constellation whose imagery serves maritime detection among other missions. Anduril (autonomy and uncrewed surface vessels) and Palantir (a horizontal government data platform that can host maritime workflows) both appeared on the seed list with a verify tag for specific MDA deployments; named-outlet citations specific to MDA software contracts were not located in this pass, so per the sourcing rules neither carries any claim here beyond this verification note.

Seven strategic moves are in motion across the maritime domain awareness field in 2026. They are not all the same kind of move. Some are repositioning bets by a single company; others are structural shifts that several players are riding at once. Each is anchored to a cited public event, read in plain operator voice, and stated as a falsifiable expectation that the next twelve to twenty-four months will either confirm or break. Where a move's success depends on something we cannot yet see in public sources, the prediction is grounded only on signals anyone can verify later: a press release with a named buyer, a tender award notice, an earnings disclosure, a public product-page change.

Observation. FTV Capital completed the take-private of Windward on 2025-03-18, delisting the company from the London Stock Exchange's AIM market and re-establishing it as a privately held U.S. company at a price of roughly £216M (about $270M).⁹⁷ Windward had built its revenue base in commercial risk and compliance, selling to insurers, traders, and charterers. Its public framing since the deal is conspicuously defense-forward: the homepage now leads with "MISSION-GRADE INTELLIGENCE" and calls itself "the leading Maritime AI™ company delivering a unified operational picture through multi-source fusion,"⁹⁸ and the dedicated government page promises "Predictive intelligence for national security."⁹⁹ The company has been quietly assembling sensor reach through partnerships rather than buying satellites: a radio-frequency tracking partnership with HawkEye 360 back in 2019, and an electro-optical monitoring partnership with Vantor announced 2026-04-16.¹⁰⁰

Participants. Windward, with FTV Capital as the owner underwriting the pivot.

Conditional outcome. If the government pivot is working, then within twelve to eighteen months named-outlet journalism should report U.S. federal MDA contract awards that name Windward as the winner. If by the end of 2027 Windward's publicly visible reference customers are still dominated by commercial compliance use cases, the pivot has stalled and the take-private bought freedom without traction.

Evidence. FTV Capital completion release and PR Newswire, 2025-03-18;⁹⁷ Windward homepage and defense page, accessed 2026-06-12;⁹⁸⁹⁹ ExecutiveBiz on the Vantor partnership, 2026-04-16.¹⁰⁰

Observation. Kpler, a commercial trade-intelligence company, has bought its way into being the scale player of the maritime data layer. It acquired both MarineTraffic and FleetMon in 2023 (announced 2023-02-15, closed in March 2023; terms not disclosed),¹⁰¹ and closed the acquisition of Spire's maritime data business, including the equity of exactEarth, on 2025-04-25 for about $233.5M.¹⁰² It now runs an explicit government motion: a Defense Intelligence product, an Esri Velocity geospatial integration, and a U.S. government symposium, with trade press describing the defense unit as its highest-performing business.¹⁰³ Kpler's wedge against analytics-led players is breadth of data, not depth of analytics. Its government page states the difference plainly: "Unlike AIS-only solutions, Kpler integrates vessel behavior, cargo flows, trade intelligence, and entity datasets."¹⁰⁴

Participants. Kpler, with Insight Partners and Five Arrows (Rothschild & Co) as the growth investors behind the platform.¹⁰⁵

Conditional outcome. If the UK Competition and Markets Authority clearance of the Spire deal holds and the defense unit keeps its internal priority, Kpler becomes the unclassified-data pricing benchmark every analytics specialist, Windward included, has to position against. The CMA closed its inquiry on 2025-07-25, finding no relevant merger situation and revoking its initial enforcement order, so the clean-clearance branch is already realized;¹⁰⁶ the remaining falsifier is a visible loss of the defense unit's priority, observable if Kpler's government surfaces or symposia presence quietly recede over the next renewal cycle.

Evidence. Kpler acquisition announcements, 2023-02-15;¹⁰¹ Spire completion release and SpaceNews, 2025-04-25;¹⁰² Kpler government and Defense Intelligence pages plus ExecutiveBiz interview, accessed 2026-06-12;¹⁰³¹⁰⁴ GOV.UK CMA case decision, 2025-07-25.¹⁰⁶

Observation. HawkEye 360 renewed a $98.8M firm-fixed-price IDIQ vehicle with the U.S. Navy on 2025-12-09, extending Navy access to its commercial radio-frequency geolocation data under the Indo-Pacific Partnership for Maritime Domain Awareness (IPMDA).¹⁰⁷ In the same release the company calls itself "the global leader in signals intelligence data and analytics" and explains that its "constellation detects, characterizes, and geolocates RF signals from ships and other emitters worldwide" to help identify "dark vessels."¹⁰⁷ The company then went public, IPO-ing on the New York Stock Exchange as "HAWK" in May 2026, raising $416M at roughly a $2.4B valuation.¹⁰⁸

Participants. HawkEye 360, with the U.S. Navy and allied IPMDA partners as the anchor customers.

Conditional outcome. RF geolocation is the single most defensible sensing niche in MDA because it directly answers the dark-vessel problem that AIS-based systems cannot: a ship that switches off its transponder still emits radio signals. If RF stays a standing line item in allied MDA budgets, HawkEye 360 should book continued renewals and additional allied-government frameworks, each visible as a named award. A visible loss of the Navy vehicle at its next recompete would falsify the read.

Evidence. PR Newswire on the Navy IDIQ renewal, 2025-12-09;¹⁰⁷ Yahoo Finance and SEC Form 424B4 on the IPO, May 2026.¹⁰⁸

Observation. SRT Marine Systems, a UK small-cap listed on London's AIM market, is winning the kind of sovereign turnkey national-MDA programs the value-chain map says belong to defense primes. It disclosed a $180M national coast guard MDA system contract (RNS, December 2023) and a $213M new MDA system award (RNS, October 2024).¹⁰⁹ Its full-year results to 30 June 2025 reported revenue of £78M, up 426% year on year, with MDA systems contributing £68.5M, and a validated sales pipeline of £1.8B.¹¹⁰ Its site positions the company as "The global leader in advanced maritime domain awareness technologies and solutions, from sophisticated national surveillance systems for coast guards to large scale fisheries monitoring."¹¹¹

Participants. SRT Marine Systems, selling to coastguard and fisheries buyers in mid-tier sovereign markets.

Conditional outcome. SRT's wedge is "MDA-only, faster and cheaper than a defense prime, with a fisheries and coastguard mission focus" — it sells to civilian-defense buyers who do not need weapon-system integration. If SRT lands further nine-figure sovereign awards, publicly disclosed through its regulatory news (RNS) announcements, the assumption that only primes win the turnkey national-system layer breaks for mid-tier states. Contract slippage disclosed in those same RNS announcements would be the public falsifier.

Evidence. SRT RNS contract disclosures, December 2023 and October 2024;¹⁰⁹ SRT FY25 final results RNS, 2025-10-22, and MoneyWeek coverage;¹¹⁰ SRT homepage, accessed 2026-06-12.¹¹¹

Observation. Skylight, a program of the Allen Institute for AI (Ai2), is free to government, regional, and non-governmental enforcement users and is genuinely deployed in more than 70 countries.¹¹² Its site states that "Skylight's technology is free to government, regional, and nongovernmental organizations worldwide that work to protect our seas from illegal fishing."¹¹² Ai2 pledged $3M to accelerate AI against illegal fishing, and in 2026 launched "Shippy," a free conversational AI agent for maritime analysts on the platform.¹¹³ Global Fishing Watch, a nonprofit funded through a $60M five-year commitment via TED's Audacious Project, occupies the same free-to-enforcement ground.¹¹⁴

Participants. Skylight (Ai2) and Global Fishing Watch, with commercial specialists as the parties whose pricing they constrain.

Conditional outcome. A free, philanthropically funded service caps what any commercial vendor can charge for baseline illegal-fishing and fisheries analytics in a developing coastal state — that vendor is now bidding against zero. If the read is right, commercial specialists respond by moving up-stack toward sovereign deployment, classified fusion, and command-and-control integration, which a free Sentinel-1-plus-AIS service cannot match. That response is observable in their public product pages over the next twelve to twenty-four months.

Evidence. Skylight homepage and Ai2 program page, accessed 2026-06-12;¹¹² GeekWire on Shippy and the Ai2 $3M pledge, 2026;¹¹³ Global Fishing Watch Audacious Project release, 2023-06-08.¹¹⁴

Observation. Saildrone, a U.S. uncrewed surface vehicle company with a NOAA science heritage, is repositioning overtly around defense, with the tagline "Detect. Deter. Dominate."¹¹⁵ It opened a Copenhagen-based European subsidiary (2025-04-09), closed a $60M financing led by Denmark's Export and Investment Fund (EIFO) explicitly to bring maritime autonomy to Europe (2025-05-13), and began operational testing with the Danish Armed Forces in the Baltic in May 2025.¹¹⁶ Lockheed Martin then invested $50M (announced 2025-10-29) to integrate its missile launcher onto the Saildrone Surveyor for the U.S. Navy.¹¹⁷

Participants. Saildrone, with EIFO, the Danish Armed Forces, and Lockheed Martin as the strategic backers.

Conditional outcome. A sovereign-adjacent fund underwriting a foreign platform vendor is the tell: it signals European demand for collection capacity that European primes are not filling quickly enough. Saildrone sells "MDA-as-a-service" — governments get persistent collection without owning the platforms — which competes less with Windward and Kpler software and more with the manned-patrol and uncrewed-aircraft line items of national budgets. If additional NATO-country deployments are announced through 2026, the beachhead read holds; an absence of new sovereign deployments by mid-2027 falsifies it.

Evidence. Saildrone homepage, accessed 2026-06-12;¹¹⁵ Saildrone European-expansion and EIFO-round releases, 2025-04 and 2025-05, and Danish Armed Forces testing release;¹¹⁶ Saildrone and Lockheed Martin release, 2025-10-29.¹¹⁷

Observation. The Tier-1 defense primes are meeting specialists head-on by accepting the contracting shape that civilian European agencies prefer: surveillance bought as a service, not platforms sold outright. Airbus won a €30M EMSA framework for uncrewed-aircraft maritime surveillance services (2025-12-18), and Airbus with Israel Aerospace Industries signed a follow-on Frontex framework (December 2024) to operate the Maritime Heron over the Mediterranean for four years from 2025.¹¹⁸ These awards sit in the same procurement universe where specialist operators win parallel lots — Tekever with CLS, and Nordic Unmanned, both holding EMSA remotely piloted aircraft contracts.¹¹⁹ At the turnkey-system layer, Indra builds and exports Spain's SIVE coastal surveillance system for the Guardia Civil, and Raytheon designed the Philippines' National Coast Watch Center under U.S. funding — both civilian-led missions delivered by primes.¹²⁰

Participants. The prime class (Airbus, IAI, Indra, Raytheon, and peers), competing directly against specialist UAS operators inside civilian-agency tenders.

Conditional outcome. If primes continue winning civilian-agency surveillance-service frameworks alongside specialists, the "primes only sell platforms, specialists own services" boundary keeps dissolving. The public scoreboard is the stream of future EMSA, Frontex, and EFCA framework award notices published through EU tender channels (TED). One caveat, stated honestly: the public award record shows primes bundling sensors, command-and-control, and integration — it does not show MDA software packaged with weapon systems, so that narrower bundling mechanism remains an open question rather than a confirmed play (see §N.6 and the contenders analysis for the full treatment).

Evidence. Context.ph on the Airbus EMSA award, 2025-12-18, and Airbus newsroom plus Naval News on the Frontex Heron follow-on, December 2024;¹¹⁸ CLS and MarineLink on the Tekever/CLS and Nordic Unmanned EMSA contracts, accessed 2026-06-12;¹¹⁹ Indra company news on SIVE and Raytheon release on the Philippines center, 2015-05-18.¹²⁰

This is the money story of the maritime domain awareness front: who is capitalized, at what scale, by whom, and which players carry a documented distress signal. Every figure below traces to a vendor-controlled surface, an SEC or exchange filing, an official government portal, or named-outlet journalism; where a figure is not public, "not disclosed" is recorded as the finding rather than estimated. Aggregator totals of the Crunchbase or Tracxn variety are labeled directional only. Most important for this front: none of the defense primes discloses an MDA-specific revenue number, so for every prime, MDA revenue is reported as "not disclosed" — a finding, not a gap to be filled with a guess.

The table below captures the most recent capital event, total disclosed funding, and any M&A activity for each profiled and adjacent specialist. It is deliberately specialist-heavy: the primes appear in the segment-financials discussion that follows, because they cannot be placed on the same funding-round basis.

The defense primes that compete in MDA fund it from multi-billion-euro defense segments with record order books, but none of them breaks out an MDA-specific revenue line. At the segment level, Thales reported FY2025 group sales of €22,136M with a Defence segment order book of €41.6B, equal to about 3.4 years of sales; Saab reported FY2025 sales of SEK 79.1B with record order bookings of SEK 168.5B and "particularly strong development in Surveillance"; and Kongsberg reported its KSAT satellite-services unit at NOK 2.4B in revenue for FY2025, the closest disclosed proxy to a maritime-monitoring line at any prime, while completing the N3X satellite constellation for Norway's maritime domain awareness.¹³³ For every one of these companies, MDA-specific revenue is not disclosed, and this report does not estimate it. The honest capitalization fact is an asymmetry of visibility, not of capability: primes fund MDA out of opaque defense segments, while specialists fund it from discrete, disclosed war chests. Nothing in the prime financials confirms or refutes whether MDA is bundled with weapon systems — that question cannot be answered from financial disclosures and is left to the tender evidence in §N.6.

There are, however, citable demand signals that primes and public-market players have put on the record. NATO selected Planet Labs in a 2025-06-12 release for "persistent space-based surveillance" with "Maritime Domain Awareness (MDA) functions" named as a contracted capability of a commercial EO constellation, and ICEYE framed its June 2026 mega-round as "a new era of sovereign intelligence from space" while disclosing 2025 revenue above €250M with backlog over €1.5B.¹³⁴ These are the demand quotes the front can stand on: a Tier-1 alliance customer naming MDA as a contracted function, and the largest SAR constellation owner framing its raise around sovereign demand.

Between February 2023 and June 2026 the MDA specialist landscape produced one roll-up platform (Kpler, absorbing MarineTraffic, FleetMon, and Spire Maritime including exactEarth), two private-equity take-privates (ORBCOMM in 2021, Windward in 2025), one incumbent absorption (Vake into KSAT), one distressed divestiture (Spire), one IPO (HawkEye 360), and one decacorn-scale round (ICEYE). I read this as a two-track capitalization structure forming. On one track, the sensor owners — ICEYE, HawkEye 360, Unseenlabs, Saildrone, Planet — are raising or listing at scale on sovereign-defense demand, with government-anchored balance sheets behind them. On the other, the AIS data-aggregation and analytics layer is consolidating into either commodity-intelligence platforms (Kpler) or private-equity ownership (Windward under FTV, ORBCOMM under GI Partners) — the standalone-public-smallcap path for maritime analytics has effectively closed, with SRT Marine the remaining listed pure-play and its model is systems integration, not data.

Only one player on this front carries a casualty label, and it is footnoted to specific cited events. Spire Global was the distressed seller of the cycle. The casualty framing rests on a chain of disclosed events: its audit committee concluded on 2024-08-23 that its FY2022–FY2023 and Q1-2024 financials had to be restated over revenue recognition on "Space as a Service" contracts; it received an NYSE noncompliance notice on 2024-08-21 for the late filing of its Q2-2024 quarterly report; it sold its highest-profile data franchise, the maritime business built on the $161M exactEarth acquisition, to Kpler with proceeds used to retire all outstanding debt; and that sale itself stalled into litigation with the buyer before settling pre-close.¹³⁵ I read this as the clearest MDA-market casualty of the cycle: a constellation owner that bought its way into maritime data leadership in 2021 and exited it under balance-sheet pressure in 2025. (Spire also announced an executive transition alongside the restatement; as a single-source signal, that is recorded here as a neutral fact only and carries none of the casualty framing.)

Every other consolidation event on this front is descriptive, not a casualty. Windward's AIM chapter ended in a take-private, but the economics do not support a distress label: FTV's £216M price was roughly double the £126.5M IPO market capitalization, and its last public financials showed first-half 2024 revenue up 37% year on year with management stating the company was financed to profitability.¹²³ The accurate pattern statement is that a growing maritime-AI vendor of about $33M revenue found more value private, with a U.S.-growth-focused owner, than as a small-cap on AIM. Vake was a seed-stage company absorbed by KSAT with no distress evidence disclosed — consolidation of an AI capability into a ground-segment incumbent. exactEarth is a twice-traded asset (Spire in 2021, Kpler in 2025), evidence of churn in the AIS data layer rather than operating failure. And ahead of HawkEye 360's IPO, a search for layoff or down-round evidence returned no named-outlet citation; the company priced its IPO at the top of its range, so no struggle framing is warranted. No down-rounds were found among the profiled specialists in named-outlet coverage during the research window; the observable 2024–2026 capital events were all up-trajectory or new-capital events.

Falsifiable prediction. If the two-track read is right, then within four to six quarters we should observe at least one further analytics-layer specialist still independent today — Starboard, SynMax, or a peer — announcing either acquisition by a platform or prime, or a strategic round led by a defense-industrial investor, observable through a press release with a named acquirer or lead. We should also continue to see sovereign-flavored capital events on the sensor side, observable through funding announcements with named leads. If instead an independent analytics specialist completes a standalone IPO on a major exchange in that window, the consolidation read weakens.

Not investment advice. See front-matter disclosure.

Three patterns shape who is winning and losing in Maritime Domain Awareness right now. Each rests on public evidence already gathered in the war-chest and playing-ground research; each is stated as my own read of that evidence; and each carries a falsifiable prediction the next year or two will confirm or break. After the three patterns, a short prose subsection answers the question the client actually asked — the one about defense primes versus civilian-focused specialists — and answers it honestly, including the part the public record cannot yet settle.

A note on what "winning" means here. MDA is not one market. It is a stack of four layers — the satellites and sensors that collect the raw picture, the software that fuses and interprets it, the systems integrators that build sovereign command centers, and the platforms (drones, uncrewed surface vessels) that carry sensors into the field. A vendor can be winning at one layer and irrelevant at another. The patterns below are read across the whole stack, because the most interesting thing happening in MDA is the way capital, customers, and competition are sorting themselves differently at each layer.

>

Observation. Between early 2023 and mid-2026, the money flowing into MDA split cleanly into two tracks. On the sensor-owning track, capital arrived at scale and on sovereign-flavored terms: ICEYE closed a round exceeding €1 billion (€450M primary, ~$520M) at a valuation above €10 billion in June 2026, led by General Atlantic, with Finnish state-linked investors Solidium, Tesi, Varma and Ilmarinen plus the Qatar Investment Authority alongside ¹³⁶; HawkEye 360 went public on the NYSE in May 2026, raising $416M at roughly a $2.4 billion valuation ¹³⁷; Unseenlabs raised €85M in February 2024 to bring its total to €120M ¹³⁸; and Saildrone took a $60M round led by Denmark's state Export and Investment Fund in May 2025, followed by a $50M strategic investment from Lockheed Martin in October 2025 ¹³⁹. On the analytics-and-aggregation track, the story was consolidation, not fresh growth capital: Kpler rolled up MarineTraffic and FleetMon (2023) and then Spire's maritime business including exactEarth (closed April 2025, ~$233.5M) ¹⁴⁰; Windward was taken private off the London exchange by FTV Capital at £216M (~$270M) in March 2025 ¹⁴¹; ORBCOMM had already gone private via GI Partners in 2021 ¹⁴²; and Norway's Vake was absorbed by Kongsberg Satellite Services in March 2024 ¹⁴³.

>

My read. I read this as a two-track capitalization structure hardening into permanence. The sensor owners are being funded the way national infrastructure is funded — by sovereign wealth funds, state development banks, defense primes, and growth investors writing checks against multi-year sovereign demand. The analytics and data-aggregation layer, by contrast, is being tidied up: the independent path of a standalone, publicly-traded maritime-analytics small-cap (Windward on London's AIM, exactEarth on Toronto's exchange, Spire's maritime unit inside a SPAC) has effectively closed, with those businesses ending up inside larger commodity-intelligence platforms or private-equity portfolios. The one notable exception, SRT Marine Systems, is still publicly listed — but it sells turnkey national systems, not data or analytics, so it does not contradict the read. Money is treating the satellites as scarce strategic assets and the software-on-top as something to be owned by whoever already owns the data.

>

Conditional prediction. If the two-track read is right, then within four to six quarters we should see at least one more independent analytics specialist (of the kind still standing alone — a Starboard or SynMax-class company, or a peer) announce either acquisition by a larger platform or prime, or a strategic funding round led by a defense-industrial investor, observable in a press release that names the acquirer or lead; and we should see continued sovereign-flavored capital events on the sensor side, observable in funding announcements with named state-linked or prime leads. If instead an independent analytics specialist completes a standalone listing on a major exchange in that window, the consolidation read weakens and the two tracks are not as fixed as they look today.

>

Sources. ^{136 137 138 139 140 141 142 143}

Sensor owners (ICEYE, HawkEye 360, Unseenlabs, Saildrone) capitalized at scale on sovereign demand; the analytics/data layer (Windward, Kpler-absorbed assets, ORBCOMM, Vake) consolidated rather than raising fresh growth capital. Author's read of public funding events, June 2026. Not investment advice.

>

Observation. The government MDA software market is not one market — it is dozens of small, slow, separately-gated national markets, and the gating is documented in the procurement record. Personnel security clearances are required at the moment of bid: Frontex's 2025 aerial-surveillance terms of reference require that tenderers' staff hold valid EU CONFIDENTIAL-or-above clearance certificates "at the moment of submission of offers," not as a post-award formality ¹⁴⁴; UK government guidance to industry on NATO procurement states plainly that "only companies maintaining cleared facilities and appropriate personnel clearances will be able to perform the resulting contract" ¹⁴⁵. Data must often live in-country: the same Frontex document anticipates that the agency "may lead Frontex to a need for taking over or building its own infrastructure to fully control processing and storage of the surveillance data," with the contractor required to migrate accordingly ¹⁴⁶. And the contracts themselves lock for years: Europe's dominant vehicle is the multi-year framework contract — EMSA's satellite-AIS data frameworks run roughly four years each, awarded successively to LuxSpace/ORBCOMM, then exactEarth/Hisdesat, then Spire (two contracts, up to €8.4M, four-year term) ¹⁴⁷ — and Frontex frameworks re-open competition only among the firms already holding the framework, locking everyone else out for the framework's life ¹⁴⁸. The clearest public proof that these gates bind: Windward, a software company, acquired the US defense-tech firm Prominent Edge in April 2026 — a firm whose stated value was its "ISO 27001, ISO 9001, and CMMI Dev Level 3 certifications" — explicitly "to deepen U.S. defense and national security capabilities" ¹⁴⁹.

>

My read. I read this as the single most important structural fact for any software company eyeing government MDA: there is no "the market" to enter. There are roughly forty front doors, each with its own lock, and the locks are national by design — clearances are issued by national authorities to nationally-established entities, hosting is sovereign, command-and-system integration is bespoke per country, and the framework contracts open for competition perhaps once every four years per agency. A foreign software vendor cannot buy its way through quickly; it has to either stand up a cleared local subsidiary over years, team under a cleared prime, or buy a cleared firm outright. Windward buying Prominent Edge is the textbook example of that third route — a maritime-AI company paying for a credentialed delivery organization because the SaaS motion alone does not clear the gate. The barriers do not add up; they multiply, because each one is per-jurisdiction.

>

Conditional prediction. If this read holds, then over the next four to six quarters we should see further public evidence of specialists buying or building cleared, in-country delivery capacity — additional acquisitions of credentialed local firms, or announced local subsidiaries with cleared staff, visible in press releases or filings — rather than specialists winning national defense MDA contracts on a pure cross-border SaaS model. If instead a software-only vendor with no cleared local presence is publicly named as the prime winner of a national-security MDA contract (in an official award notice or named-outlet coverage), the "forty front doors" barrier is more porous than the procurement record suggests, and the read needs revising.

>

Sources. ^{144 145 146 147 148 149}

Each barrier — personnel clearance at bid, sovereign data hosting, bespoke command-system integration, and four-year framework lock-in — is issued and enforced per jurisdiction, so they compound rather than add. Windward's April 2026 acquisition of cleared US firm Prominent Edge is the proof point. Author's read of public procurement evidence, June 2026.

>

Observation. The basic maritime picture is now available for free, supplied by governments and philanthropy. SeaVision, built by the US Department of Transportation's Volpe Center, is provided free to "hundreds of government agencies globally" ¹⁵⁰. Skylight, run by the Allen Institute for AI, is free to "government, regional, and nongovernmental organizations worldwide that work to protect our seas from illegal fishing," is reported in use in 70-plus countries, and in 2026 added a free conversational AI agent called Shippy ¹⁵¹. Global Fishing Watch publishes free enforcement-grade vessel data, backed by a five-year $60M philanthropic commitment, and partnered with the US Defense Innovation Unit and Coast Guard on the open-source xView3 dark-vessel detection challenge ¹⁵². Meanwhile the commercial vendors that sell into this space lead with language that points away from the basic picture: Windward markets "mission-grade intelligence" and a "unified operational picture" through multi-source fusion ¹⁵³; Kpler tells governments that "unlike AIS-only solutions," it integrates "vessel behavior, cargo flows, trade intelligence, and entity datasets" ¹⁵⁴.

>

My read. I read this as a free floor under the category that quietly resets where money can be made. When a developing coastal state can get a credible illegal-fishing picture from Skylight or SeaVision at no cost, no commercial vendor can charge for that baseline picture anymore. So the commercial vendors are pushed up-stack — toward the things the free tier structurally cannot do: sovereign and on-premises deployment, fusion with classified national feeds the free services have no access to, accredited integration into national command systems, and contractual service-level and liability guarantees a philanthropic project will never sign. That is exactly why Windward's and Kpler's messaging reaches for "mission-grade" and "beyond AIS-only" — they are not competing for the basic picture, they are competing for the part of the problem the free floor cannot reach. The free tier is not a charity footnote to this market; it is a force that shapes commercial positioning.

>

Conditional prediction. If this read is right, then over the next twelve to twenty-four months the public product pages and messaging of the commercial analytics specialists should keep moving up-stack — more emphasis on sovereign/on-prem deployment, classified-side fusion, accreditation, and C2 integration, and less emphasis on baseline dark-vessel detection as a headline differentiator — observable directly on their websites and in their press releases. If instead a commercial vendor publicly builds its differentiation around free-tier-equivalent baseline detection and wins on that basis (named in a press release or named-outlet win), the free floor is exerting less pressure than the read claims.

>

Sources. ^{150 151 152 153 154}

Free, government-furnished and philanthropic services set a price floor on the basic maritime picture; commercial vendors respond by moving their differentiation up-stack into the territory the free tier cannot reach. Author's read of public product surfaces, June 2026.

The most pointed question in this space is whether the big defense primes — Saab, Thales, Airbus, Raytheon and their peers — crowd out the focused civilian specialists, or whether the two are increasingly fighting over the same tenders. The honest answer is layered, and the layers matter, because each is supported by a different kind of evidence and one of them is not supported by the public record at all.

First, the overlap is real, and it runs in both directions. The cleanest evidence sits in Europe's civilian agencies. The EU's maritime-safety agency, EMSA, awarded Airbus a €30M framework for drone-based maritime surveillance services in the same procurement universe where it also awarded parallel contracts to specialist drone operators like Tekever (with CLS) and Nordic Unmanned ¹⁵⁵. The EU border agency, Frontex, selected an Airbus-and-IAI team to fly maritime surveillance drones over the Mediterranean for four years ¹⁵⁶. And primes are deeply embedded in civilian turnkey systems — Indra built and exports Spain's national coastal-surveillance system, and Raytheon designed the Philippines' National Coast Watch Center ¹⁵⁷. So primes do not stay in some pure "lethal weapons" lane; they compete hard for civilian coastguard and border work. The overlap also runs the other way: SRT Marine Systems, a UK small-cap specialist, has won sovereign national-MDA system contracts disclosed at $180M and $213M — exactly the turnkey national-system layer the conventional wisdom assigns to primes ¹⁵⁸.

Second — and this is the part to be careful about — the specific mechanism in the client's framing, primes bundling MDA software with high-value lethal weapon systems, is not evidenced in any award document reviewed in this research. What the public record actually shows primes bundling is sensors, command-and-control, and integration into turnkey coastal-surveillance systems (Thales's CoastShield is the catalogue example), and what they sell to civilian agencies is surveillance delivered as a service ¹⁵⁹. No reviewed tender or award shows MDA analytics packaged with a weapons platform. That does not prove it never happens — defense-tender documentation at that level is largely non-public, and a targeted sweep of government audit-office reports (the US GAO, the UK NAO, the European Court of Auditors) would be the logical way to test it further. But on the evidence available, the weapons-bundling mechanism should be reported as an open question, not as an established market fact.

Third, where civilian-focused specialists genuinely face less prime competition is the analytics-and-fusion software layer specifically. Across every analytics deployment surfaced in this research — Windward's partnerships, Kpler's government build-out, Skylight's free deployments, the DIU/Global Fishing Watch/Coast Guard collaboration, Canada's MDA Space dark-vessel program — the named winners are specialists, nonprofits, or state-funded programs, not Tier-1 primes selling standalone analytics software ¹⁶⁰. The primes' MDA-adjacent offerings are predominantly sensor-plus-command-system turnkey programs, a different layer of the stack. But "less prime competition" at this layer does not mean "easier market," because this is exactly the layer with the free-floor problem from Pattern Claim 3. The specialist who escapes the primes here runs straight into Skylight and SeaVision at the low end. The honest synthesis for anyone weighing this market: overlap with primes is real and growing in services and turnkey tenders; the analytics software layer is genuinely more specialist territory but is price-disrupted from below; and the weapons-bundling story specifically is unproven on public evidence.

Eight things to watch over the next twelve to twenty-four months. Each is a publicly-observable proxy — something that will show up in an earnings call, a named-outlet story, an analyst report, a public product page, a regulatory filing, or a funding announcement — so that anyone tracking this market can check whether the patterns above are holding or breaking, without needing access to any private information.

1. Windward's US federal footprint. The thesis behind Windward's take-private and its Prominent Edge acquisition is US government growth. The proxy: named-outlet coverage or official award notices naming Windward as a winner of US federal MDA work. If those appear within twelve to eighteen months, the gov pivot is working; if Windward's public references through end-2027 stay dominated by commercial compliance use cases, the pivot stalled ¹⁶¹.

1. Kpler's ownership and defense priority. Kpler is the consolidator of the unclassified data layer, and its minority owners have been reported (in talks only) as considering a stake sale at a reported enterprise value above €3 billion. The proxies: whether any UK competition-authority review of its acquired maritime assets resolves cleanly, and whether its defense unit retains its reported status as a top-performing business line, both observable in regulatory pages and trade-press interviews. A clean resolution plus sustained defense priority makes Kpler the pricing benchmark every analytics specialist must position against ¹⁶².

1. HawkEye 360's next US Navy recompete. RF geolocation is the most defensible niche on the sensor side because it directly answers the dark-vessel problem that AIS cannot. HawkEye 360 holds a US Navy contract for RF data under the Indo-Pacific maritime-awareness partnership. The proxy: continued renewals and additional allied-government frameworks (confirming RF as a standing budget line), versus a visible loss of the Navy vehicle at recompete (which would falsify) — both observable in press releases and government award notices ¹⁶³.

1. SRT Marine's next sovereign award (RNS). SRT is the live counter-example to "only primes win turnkey national systems." Because it is publicly listed in London, its contract wins and any slippage are disclosed through regulatory news (RNS) announcements. The proxy: further nine-figure sovereign MDA system awards disclosed via RNS confirm specialists can hold the turnkey layer in mid-tier states; disclosed contract slippage or delay is the public falsifier ¹⁶⁴.

1. EU agency awards on the public tender record (TED). EMSA and Frontex publish their framework awards through the EU's official tenders database (TED) and agency announcements. The proxy: the running scoreboard of who wins the next maritime-surveillance service frameworks — primes, specialists, or mixed lots — is the single best public read on whether prime-vs-specialist overlap is widening or narrowing in the civilian European market ¹⁶⁵.

1. Consolidation of an analytics-layer specialist. Pattern Claim 1 predicts a still-independent analytics specialist (a Starboard- or SynMax-class company, named here at the class level with no acquisition assertion) will be acquired by a platform or prime, or take a strategically-led round, within four to six quarters. The proxy: a press release naming the acquirer or the defense-industrial lead investor. A standalone major-exchange listing by such a company instead would weaken the consolidation read ¹⁶⁶.

1. ICEYE's deployment of its €1B round. ICEYE raised more than €1 billion in June 2026 against company-stated 2025 revenue above €250M and a contracted backlog above €1.5 billion. The proxy: how that capital is deployed — additional sovereign SAR contracts, constellation expansion, or downstream analytics moves — observable in press releases and government award notices, and a direct test of whether sensor-side capital keeps arriving on sovereign terms as Pattern Claim 1 predicts ¹⁶⁷.

1. Saildrone's Baltic-to-Indo-Pacific follow-through. Saildrone established a European beachhead with a Copenhagen subsidiary, a state-fund-led round, and Danish Armed Forces testing in the Baltic, with stated intent to roll out to multiple NATO countries. The proxy: additional NATO-country deployments announced through 2026 and into 2027 confirm European demand for collection capacity that European primes are not filling quickly; an absence of new deployments by mid-2027 falsifies the beachhead thesis ¹⁶⁸.