Press Release
The Same Camera Rents for USD 23,700 a Year and Sells for USD 128,000 a Position — 78.3 Per Cent of the World's Enforcement Positions Carry 34.2 Per Cent of Its Value
Pheonix Research has published Global Road Safety & Enforcement Systems Market Report, Size, Share and Forecast 2026–2033: The Camera Was Never the Product, which reconstructs the global installed base of automated enforcement at 435,641 positions and values the market those positions generate at USD 5,179.2 million in 2025, reaching USD 10,069.9 million by 2033 at a compound annual rate of 8.67 per cent — an increment of USD 4,890.7 million, or 1.94 times the base year. The full report is available at https://www.pheonixresearch.com/automotive-transportation/road-safety-enforcement-systems/market-report/global-road-safety-enforcement-systems-market/.
The report’s central finding is that the unit of sale is the enforcement position, and that a position is priced by what the contract carries rather than by what the camera does. The same roadside function rents for roughly USD 23,700 a year in a 50,000-person municipality and for USD 76,769 a position-year in New York City — a 3.24 multiple — and 78.3 per cent of the world’s enforcement positions carry 34.2 per cent of the market’s value.
A position is priced by what the contract carries, not by what the camera does
Automated road safety enforcement is bought, sold, reported and forecast as a camera market. It is not priced like one. Across the three disclosed transactions that set this category’s brackets, the same physical function — a roadside position that measures a vehicle, assembles an evidence package and issues a notice to a registered keeper — changes hands at a spread of 3.24 times. Very little of that spread is hardware.
| Transaction | Structure | Value per position |
|---|---|---|
| New York City — five-year contract effective 1 January 2026, with a five-year extension option | Managed service. Supplier finances, installs, maintains and operates the estate. Fines excluded. | USD 998.0 million total contract value = USD 199.6 million a year across roughly 2,600 speed and red-light positions — USD 76,769 per position-year |
| NYCDOT red-light expansion — 300 positions installed in Q3 and Q4 2025 | Capital sale and integration under a change order, on legislatively authorised expansion | USD 38.4 million in the period — USD 23.9 million installation services, USD 14.5 million product — USD 128,000 per position, one-time |
| Belleville, Ontario — council resolution 496-2023 | Four cameras rented for five years, including housing, deployment, processing and the city’s share of administration | CAD 649,214.16 over five years = CAD 32,460.71 per camera-year, roughly USD 23,700 |
Read the first and third rows together. A metropolitan authority pays 3.24 times what a 50,000-person municipality pays for a position doing the same job, and the difference is not optics. It is who owns the asset, who carries the maintenance obligation, who indemnifies the evidentiary process, and who is exposed if a change of council cancels the programme. The wider the obligation, the higher the annual value — and the longer the contract has to be to finance it. That is why New York’s term runs to ten years with the option and Belleville’s runs to five.
The camera is the cheapest, least differentiated and least valuable component of what is being bought. What a road authority buys is an enforceable outcome: a measurement that survives legal challenge, an evidence package that survives appeal, a notice that reaches a registered keeper, and a payment or a court date at the end of it.
435,641 positions, and a value split that does not follow them
Pheonix Research reconstructs a global installed base of 435,641 enforcement positions in 2025 and values the market at USD 5,179.2 million, reaching USD 10,069.9 million by 2033 at 8.67 per cent CAGR — an increment of USD 4,890.7 million, or 1.94 times the base.
| Tier | Positions, 2025 | Value per position-year | 2025 (USD m) | Share of value | CAGR 2025–33 | 2033 (USD m) |
|---|---|---|---|---|---|---|
| Tier 1 — managed-service markets Western Europe, North America, Australia and New Zealand, Japan, Korea, Singapore, Taiwan |
94,606 | USD 36,000 | 3,405.8 | 65.8% | 6.6% | 5,679.1 |
| Tier 2 — system-sales markets China, India, Russia, Brazil, the Gulf, Southeast Asia, Africa, rest of Latin America |
341,035 | USD 5,200 | 1,773.4 | 34.2% | 12.0% | 4,390.8 |
| Total | 435,641 | USD 11,889 blended | 5,179.2 | 100% | 8.67% | 10,069.9 |
Tier 1 holds 21.7 per cent of the world’s enforcement positions and 65.8 per cent of its value. Tier 2 holds 78.3 per cent of the positions and 34.2 per cent of the value. Tier 2 supplies 53.5 per cent of the eight-year increment and never exceeds 43.6 per cent of the market by value. Blended value per position is USD 11,889; in Tier 1 it is 3.03 times that.
That is the strategic problem this report exists to answer. The fastest-growing half of the market is the half where a position is worth least, and it grows because cities buy capital projects. The half where a position is worth most grows at 6.6 per cent, and it grows because incumbents sell additional violation categories onto estates already installed — bus-lane, bike-lane, school-bus stop-arm, seatbelt and mobile-phone detection. The marginal cost of a new use case is a fraction of the cost of the position.
The denominator is an estimate, and the estimate is soft
France publishes more about its automated enforcement programme than almost any other country. It appears twice in authoritative sources: at 4,029 devices in the national enforcement agency’s own 2025 count — 3,423 fixed, 396 deployable and 210 mobile — and at 3,745 in the commercial database the industry quotes. That is a 284-device gap, 7.6 per cent, in the single best-documented programme in Europe. Neither number is wrong. They measure different things, on different dates, with different definitions of what counts as installed.
The published counts are inconsistent in instructive ways elsewhere too. Italy reports 11,238 devices, which its own press describes as 17 per cent of the European estate. Russia counts 18,421 and Brazil 17,614. The seven largest European programmes sum to 35,385 devices. Outside Tier 1, position counts are modelled from enforcement activity and should be read as an order of magnitude with a wide band, not a census. Anyone quoting a global camera count to three significant figures is quoting a survey of a survey.
The revenue base is the behaviour the product exists to eliminate
This is the structural contradiction, and it is not a forecast risk — it is a measured outcome.
New York City’s average daily red-light violations per camera fell from more than thirty in 1994 to seven in 2021. Its speed programme recorded a 73 per cent reduction in speeding at fixed camera locations between 2019 and 2021, with one-third fewer pedestrian injuries. A national review cited reductions in the proportion of speeding vehicles of 14 to 65 per cent, with cuts in serious and fatal crashes of up to 44 per cent. Those are precisely the outcomes a road authority is buying, and they are precisely the outcomes that shrink the violation base a per-notice contract is paid from.
The industry’s decade-long migration to fixed service fees does not remove the risk; it transfers it. Under a fixed fee the authority is now paying for a service whose measured output is falling, and it will eventually ask why. The commercial durability of a managed-service contract depends on the buyer accepting that the service is the deterrent, not the revenue — an argument that is harder to make in jurisdictions where the programme is expected to be self-funding.
The attrition is large even before that. In England and Wales in 2024, 843,849 notices were cancelled out of 3,028,022 speed-limit offences dealt with — 27.9 per cent — most often because the driver could not be identified. Every cancelled notice has already consumed the full cost of capture, processing and service. That is the economic case for the automated quality gate now entering the evidence chain: a notice suppressed before issue saves the whole downstream cost, and an algorithm that decides which violations proceed is making an enforcement decision subject to the same appeal and disclosure obligations as a human officer’s.
And the market can be switched off by a vote. US communities operating automated red-light enforcement fell from more than 500 a decade ago to 338 by March 2024, even as speed camera programmes rose from 211 to 222 — school-zone speed enforcement carries public support that intersection enforcement does not. Italy holds Europe’s largest estate at 11,238 devices and recorded 3,159 road deaths in 2022, up 9.9 per cent, with municipal fine revenue up almost 24 per cent against 2019, concentrated in small municipalities, and fourteen years without the decree that would regulate where and how devices may be installed. Device count and safety outcome are separable, and a market built on device count is exposed when the outcome fails to arrive.
Where the margin actually sits
The largest company in the category is not an enforcement company. Verra Mobility reported fiscal 2025 revenue of USD 979.1 million, up 11 per cent. Its Government Solutions segment — speed, red-light, school-bus stop-arm and bus-lane programmes — ran a 24 per cent segment margin in the fourth quarter, down from 34 per cent a year earlier on implementation and readiness costs for the New York expansion. Its Commercial Services segment, which manages tolls and violations for rental and fleet operators, ran 64 per cent on USD 108.1 million of quarterly revenue. The commercial side of the same vehicle-identification rail is more than twice as profitable as the statutory side, because it carries no statutory risk and no capital.
The specialists show the same gradient from the other direction. Sensys Gatso’s 2024 sales were SEK 631 million, up 1 per cent, but TRaaS revenue reached SEK 370 million — 58.6 per cent of sales now recurring — with more than SEK 1,000 million of backlog against that revenue, 1.58 times annual sales. Jenoptik’s Smart Mobility Solutions segment grew 8.5 per cent to EUR 129.7 million at a 13.6 per cent EBITDA margin, on an order backlog of EUR 63.0 million — under six months of visibility. The difference between those two profiles is not product quality. It is contract duration.
The investment conclusion is uncomfortable for anyone modelling this as a growth sector. Verra’s 2026 guidance implies capital expenditure of roughly USD 125 million, about 12.2 per cent of guided revenue, and in a managed-service contract it is the supplier that funds the estate. Growth in Tier 1 is therefore constrained by balance sheet, not by demand. The vendors that compound are the ones with contracted duration long enough to depreciate what they build.
No vendor holds ten per cent
The largest single vendor, by enforcement-attributable revenue, holds approximately 9.5 per cent of the market. Nothing close sits behind it. Three mechanisms cap share, and all three are regulatory rather than competitive.
Type approval is national. A measurement device must be certified as legally accurate by the metrology authority of the jurisdiction that will prosecute on its output, and there is no mutual recognition. Entering a country is not entering a market with a product; it is beginning a certification programme that typically runs two to four years and yields no revenue until it completes.
The statute is the specification. Every jurisdiction defines its own violation, its own evidential standard, its own notice period and its own appeal route. A platform built for red-light enforcement under New York’s programme is not the platform for average-speed enforcement under a European point-to-point regime, because the two are proving different propositions to different tribunals.
Procurement is relational and slow. Contracts run five to ten years with extension options, and renewal is the default. Changing supplier mid-programme means re-certifying the estate, re-proving the evidence chain and absorbing political risk no programme manager wants to own.
The result is a market fragmented globally and concentrated locally: two to four qualified bidders in any country, and often only one with an installed estate. It is also why vendors compete on evidentiary grounds rather than technical ones — whether the evidence package survives first-instance appeal without manual intervention, whether notice throughput can absorb a step change in volume, and whether the vendor will indemnify the authority against a successful challenge to the measurement. Price enters late and matters less than the buyer’s exposure.
The leading indicator is a legislative calendar, not a fatality rate
The World Health Organization counted 1.19 million road traffic deaths in its most recent global status report, noted a slight fall, described road injury as the leading killer of people aged 5 to 29, and set a target of at least halving deaths and injuries by 2030. That is the strongest possible public-health case and it has been broadly stable for a decade. It has not produced a corresponding procurement curve.
New York’s largest-ever red-light expansion followed Governor Hochul signing S9504A/A.3120A in October 2024, not a fatality trend, and 300 positions were installed within two quarters. New South Wales added 2,700 new roadside sites for mobile speed camera deployment while holding total enforcement hours constant, and committed to trialling average-speed cameras for light vehicles. The 2021 US infrastructure law allowed states to spend up to 10 per cent of federal highway safety funds on non-infrastructure measures including automated enforcement, having previously restricted that money to school zones. None of those is a demand signal in any conventional sense. All of them moved the market.
A jurisdiction with authorisation and no budget waits. A jurisdiction with budget and no authorisation cannot buy. Both are common, and neither appears in a demand forecast built on road trauma.
What would have to be true
Because the position count is held flat in the model, every point of the 8.67 per cent is earned as value per position. That makes the sensitivity unusually clean.
- If the Tier 1 estate only tracked inflation at 2 per cent, 2033 is USD 8,381.2 million — 6.20 per cent.
- If Tier 2 grew at Tier 1’s rate rather than its own, 2033 is USD 8,636.2 million — 6.60 per cent.
- The base case requires Tier 2 to compound at 12.0 per cent on municipal capital programmes, and Tier 1 to raise revenue per position through category extension rather than by adding positions.
The second of those is doing most of the work. Roughly 53 per cent of the increment depends on city-level command-and-control build-out in markets where the supplier captures a hardware line item and no recurring position — which is a real investment, made by someone else.
Risks
- The self-defeating revenue base. Success reduces the violation volume the funding is built on; managed-service conversion transfers rather than removes the exposure.
- Political substitution, in both directions. 500-plus US red-light programmes became 338. Italy’s estate grew while its deaths rose 9.9 per cent. A large estate does not protect against repeal, and may invite it.
- Contract concentration. New York’s USD 199.6 million a year is 3.9 per cent of the entire global market’s annual value from one city.
- Data protection and evidentiary challenge. Shortened retention, prohibited secondary use, or a finding that automated decision-making is insufficiently explainable invalidates notices in bulk rather than individually — a binary revenue risk, not a gradual one.
- Substitution from the vehicle. Intelligent speed assistance is now required on new EU vehicles. It does not cover the existing parc and cannot issue a notice, but it compresses the addressable violation base over a ten-year horizon, which is why the industry has moved towards violations vehicle systems cannot self-enforce.
Overall market risk is assessed as high, with substitution risk high and geopolitical exposure moderate. The 8.67 per cent is not a measure of safety. It is a measure of how much statutory authorisation and municipal capital is expected to be available.
What this report counts, and what it refuses to
The category is unusually easy to inflate. Almost every number published about it either counts general urban surveillance as enforcement, or counts the money raised by fines as market revenue. Both are wrong, and both are common.
This report counts the technology and service layer: enforcement camera hardware and housings, roadside detection and measurement subsystems, evidence and violation-management software, managed enforcement services, installation and integration, and maintenance. It does not count fine and penalty revenue — a fine is a transfer from a driver to a public authority, not a payment to a supplier — nor police and adjudicator labour, general-purpose surveillance CCTV, in-vehicle ADAS, or road infrastructure. A camera on a gantry that monitors congestion is not an enforcement position unless it produces a notice.
The scope decision determines the denominator. A market defined by the vehicle parc grows with car ownership. A market defined by the enforced estate grows with statutory authorisation — council resolutions, state laws and ministerial decrees — and those two things move independently.
Method
Two layers, combined rather than blended, so that each figure can be attributed to the layer that produced it. The secondary layer, at approximately 70 per cent of analytical weight, comprises official device and programme counts, statutory and legislative records, audited company disclosures and published implementation-cost studies; device counts were cross-checked between official and commercial databases wherever both exist, and the discrepancies are reported rather than smoothed. The primary layer, at approximately 30 per cent, comprises structured interviews with road-authority programme managers and procurement leads, open-ended interviews with supplier commercial and bid teams, a structured survey of buying criteria, and conference and site work in the category including Intertraffic Amsterdam in March 2026.
The primary layer is what establishes that a position is priced on obligation rather than function, and that contract length is set by the financing structure rather than the asset life. No third-party market-research estimate was used in, or reconciled against, any figure in this report.
Companion analysis
Two articles examine the parts of this in detail:
- The Global Enforcement Estate Is 435,641 Positions. None of Them Was Counted.
- The Same Camera Rents for USD 23,700 a Year and Sells for USD 128,000 a Position
About this report
Global Road Safety & Enforcement Systems Market Report, Size, Share and Forecast 2026–2033: The Camera Was Never the Product is available at https://www.pheonixresearch.com/automotive-transportation/road-safety-enforcement-systems/market-report/global-road-safety-enforcement-systems-market/, with competitive landscape, value chain, investment activity, technology and innovation, market risk and regulatory landscape chapters, plus the full methodology and scope definitions summarised above. The two companion articles are linked in Companion analysis.
About Pheonix Research
Pheonix Research is a market intelligence platform that sizes markets bottom-up from physical volume and prices them at the point where a buyer actually pays, then translates the result into role-specific decisions. Its research is organised around value chains rather than industry labels: the same market looks different to procurement, to strategy and to the executive, and the output is built to end in an action and a timeframe rather than a chart. Pheonix publishes market reports, bespoke intelligence engagements, the role-based VALUE platform and curated signal bundles. For further information, visit https://www.pheonixresearch.com/.
