---
mi_id: MIMR-node_105c505
type: market_report
canonical_url: "https://www.pheonixresearch.com/semiconductor-electronics/test-measurement-inspection/market-report/global-inline-metrology-market/"
md_url: "https://www.pheonixresearch.com/semiconductor-electronics/test-measurement-inspection/market-report/global-inline-metrology-market.md"

schema:
  "@type": Dataset
  "@id": MIMR-node_105c505
  url: "https://www.pheonixresearch.com/semiconductor-electronics/test-measurement-inspection/market-report/global-inline-metrology-market/"
  title: "Global Inline Metrology Market Report, Size, Share and Forecast 2026–2033: Machine Vision Overtakes the CMM in 2028"
  description: "Inline metrology market USD 8,650m (2025) to USD 14,158m (2033) at 6.35%. Machine vision passes the CMM in 2028 and 56.3% of growth is additional systems, not price."
  datePublished: "2024-01-13T13:04:35+00:00"
  dateModified: "2026-11-18T13:40:31+00:00"
  keywords:
    - inline metrology market
    - machine vision dimensional measurement
    - coordinate measuring machine market
    - inline CMM
    - optical metrology forecast 2033
  isPartOf:
    - id: MISG-node_c6d5bf2
      type: CollectionPage
      url: "https://www.pheonixresearch.com/semiconductor-electronics/test-measurement-inspection/"
      name: "Test, Measurement & Inspection"
    - id: MIIN-node_89e33a6
      type: CollectionPage
      url: "https://www.pheonixresearch.com/semiconductor-electronics/"
      name: "Semiconductor & Electronics"
    - id: MIWS-root
      type: WebSite
      url: "https://www.pheonixresearch.com/"
      name: Pheonix Research
  mentions:
    - id: MIEN-node_6ee92fe4
      type: Organization
      url: "https://hexagon.com"
      name: Hexagon AB
      sameAs:
        - "https://www.wikidata.org/wiki/Q820970"
    - id: MIEN-node_c2d5c881
      type: Organization
      url: "https://www.cognex.com"
      name: Cognex Corporation
      sameAs:
        - "https://www.wikidata.org/wiki/Q2981857"
    - id: MIMR-node_d086f66
      type: Dataset
      url: "https://www.pheonixresearch.com/semiconductor-electronics/test-measurement-inspection/market-report/global-industrial-radiography-digital-x-ray-inspection-market/"
      name: "Global Industrial Radiography & Digital X-ray Inspection Market Report, Size, Share and Forecast 2026–2033: Film Was Priced Out, Not Designed Out"
    - id: MIMR-node_4528b79
      type: Dataset
      url: "https://www.pheonixresearch.com/semiconductor-electronics/test-measurement-inspection/market-report/global-non-destructive-testing-equipment-market/"
      name: "Global Non-Destructive Testing Equipment Market Report, Size, Share and Forecast 2026–2033: Sixty-Two Per Cent of the Growth Is Not More Inspection"
    - id: MIMR-node_46152e1
      type: Dataset
      url: "https://www.pheonixresearch.com/semiconductor-electronics/test-measurement-inspection/market-report/global-ultrasonic-ndt-equipment-market/"
      name: "Global Ultrasonic NDT Equipment Market Report, Size, Share and Forecast 2026–2033: Thirty-Four Per Cent of the Market, Fifty-Seven Per Cent of the Growth"
  about:
    - "@type": Thing
      name: Market Intelligence
      mi_id: "MIMR-node_105c505#intelligence"
    - "https://www.wikidata.org/wiki/Q1901028"
  base_year: 2025
  forecast_year: 2033
  value_base_year: 8650
  value_forecast_year: 14158
  value_cagr: 6.35
  value_currency: USD
  value_unit_scale: Million

acf:
  base_year: 2025
  forecast_year: 2033
  value_base_year: 8650
  value_forecast_year: 14158
  value_cagr: 6.35
  value_currency: USD
  value_unit_scale: Million
  coverage_type: Global
  coverage_name: Global Inline Metrology
  competitive_intensity_level: high
  market_structure_type: moderately_consolidated
  tier1_player_count: 8
  investment_trend_direction: rising
  capital_intensity_level: medium
  recent_mna_activity: yes
  overall_market_risk_level: moderate
  geopolitical_exposure_level: high
  substitution_risk_level: high
  regulatory_complexity_level: moderate
  innovation_intensity_level: high
  technology_maturity_stage: growth
  patent_activity_level: high
  supply_chain_complexity_level: moderate
  distribution_structure: distributor_led
  primary_operational_model: hybrid
  executive_summary: "<p>The global inline metrology market is <strong>USD 8,650 million in 2025</strong> and reaches <strong>USD 14,158 million in 2033</strong>, a compound annual growth rate of <strong>6.35 per cent</strong>. The headline hides the most interesting thing in the market, which is that its largest technology is about to lose first place to the second largest.</p><p><strong>Machine vision passes the coordinate measuring machine in 2028.</strong> Inline CMM and multisensor cells are 27.0 per cent of 2025 value at USD 2,335 million, growing 4.60 per cent. Machine vision dimensional systems are 24.5 per cent at USD 2,119 million, growing 8.70 per cent. Two growth rates and two bases put the crossover in the middle of 2028. By 2033 machine vision holds 29.2 per cent of the market and 36.5 per cent of the increment; the CMM holds 23.6 per cent and 18.3 per cent. The second-largest technology adds <strong>USD 2,012 million</strong>; the largest adds <strong>USD 1,011 million</strong>.</p><p>The reason is where measurement is being done. The CMM is a measurement-room instrument: granite, air bearings, temperature control, a part brought to it in a batch and fixtured. Machine vision is an inline instrument: no fixture, no batch, every part, and a result in the takt time of the line. The commercial argument is not accuracy — a good CMM is more accurate. It is that 100 per cent inspection became affordable as the sensor cost fell, and a sampled inspection cannot be closed-loop.</p><p><strong>The second finding runs against the rest of this cluster.</strong> We found price and mix to be 60.1 per cent of ultrasonic growth and 62.1 per cent of non-destructive testing growth, because those are replacement markets where an instrument displaces one already owned against a code obligation. Here it is the reverse: <strong>56.3 per cent of the USD 5,506 million increment is additional systems</strong> — 96,500 placed in 2025 against 131,060 in 2033 at 3.90 per cent — and realised spend per system contributes only 43.7 per cent at 2.36 per cent. Inline metrology is not replacing what was already bought. It is being installed where nothing was installed before.</p><p>That produces a different risk profile. Realisation growth of 2.36 per cent is the lowest in the cluster, because machine vision hardware rides a semiconductor cost curve that goes down rather than up. A supplier shifting its mix toward vision buys unit growth at the price of price growth. The trade only works if the software and service attached to those units carries a margin the hardware does not — which is exactly why every CMM incumbent in this market has spent a decade acquiring software.</p><p>The geography is unambiguous. <strong>Asia-Pacific is 38.5 per cent of value and takes 49.4 per cent of the increment</strong>; half the world’s growth in this market lands in one region, because that is where the line is being built and where the inspection is being specified alongside it. Europe grows 4.55 per cent on a dense installed base. Automotive and EV is the largest end market at 26.0 per cent, and it is also the most cyclical buyer in the cluster — a fact that matters more here than in any code-mandated inspection market, because nothing in this report is bought because a regulation says it must be.</p>"
  toc: "<ol><li><a href=\"#section-overview\">What this report measures</a></li><li><a href=\"#section-size\">Market size, forecast and the technology split</a></li><li><a href=\"#section-vc\">Value chain</a></li><li><a href=\"#section-ti\">Technology: why vision overtakes the CMM</a></li><li><a href=\"#section-cl\">Competitive landscape</a></li><li><a href=\"#section-rl\">Regulatory landscape: acceptance is the fence</a></li><li><a href=\"#section-ia\">Investment activity</a></li><li><a href=\"#section-mr\">Market risk and scenarios</a></li></ol>"
  competitive_landscape_content: "<h2 id=\"section-cl\">5. Competitive landscape</h2><p>This is a moderately consolidated market with an unusual shape: the incumbents are measurement companies that are buying vision capability, and the challengers are vision companies that are buying measurement credibility. Both directions are rational and both are expensive.</p><p><strong>Our estimate: the five largest suppliers account for roughly 45 to 50 per cent of system value.</strong> We label it our own estimate because the denominator cannot be observed directly — a meaningful share of inline gauging is bundled into machine tools, assembly cells and integrators' projects and never appears as a metrology line item.</p><ul><li><strong>Hexagon AB</strong> — the broadest position in the market, spanning tactile CMMs, laser scanning and trackers, portable measurement and the software layer that ties them together. Its strategic logic is explicit: own the measurement and the software that interprets it, so that the hardware choice follows.</li><li><strong>Carl Zeiss</strong> — the deepest position in high-accuracy tactile metrology and industrial computed tomography, with a strong optical portfolio. Concentrated in the applications where accuracy, not throughput, is the buying criterion.</li><li><strong>Mitutoyo</strong> — the volume leader in tactile CMMs, particularly strong in Asia, and the supplier most exposed to the shift this report describes.</li><li><strong>Cognex</strong> — the purest machine vision position, concentrated in the vision and identification hardware that is taking share. Not a metrology company by heritage, which is precisely why it is winning in the fastest-growing technology.</li><li><strong>Keyence</strong>, <strong>Nikon Metrology</strong>, <strong>Renishaw</strong>, <strong>Faro</strong>, <strong>LMI Technologies</strong> and <strong>Jenoptik</strong> — specialist positions across vision sensors, encoders and probes, portable scanning and 3D scanning, each defending an application where its specification matters more than its scale.</li></ul><h3>How the competition actually resolves</h3><p><strong>On the software.</strong> A modern CMM and a modern vision system are both capable. What a buyer cannot easily change is the measurement programme, the nominal CAD model, the reporting format and the operator's familiarity — which is why the installed software base is the strongest thing any incumbent in this market owns. Every large acquisition in this field has been, at bottom, a software acquisition.</p><p><strong>On the multisensor platform.</strong> The incumbents' answer to vision is not to compete with it but to absorb it: a cell that carries a tactile probe and an optical sensor on the same bridge, so that a feature can be measured by whichever technique suits it. This is technically the right answer and commercially a defensive one — it protects the installed base while conceding that vision wins the simple measurements.</p><p><strong>On cost, at the low end.</strong> Vision sensors are semiconductor products with a semiconductor cost curve. The low end of this market is being commoditised faster than any other segment we cover in this cluster, and it is where the 2.36 per cent realisation number comes from.</p><p><strong>On acceptance.</strong> The decisive arena, and the one covered in the regulatory chapter: a technology that cannot demonstrate measurement uncertainty to a recognised standard cannot be used for a conformance decision, however good its image.</p>"
  value_chain_content: "<h2 id=\"section-vc\">3. Value chain</h2><p>Four layers, and the value is not where the physics is.</p><p><strong>The sensing layer.</strong> Image sensors and optics for vision; encoders and scales for motion; probe and stylus assemblies for tactile; laser sources and interferometer optics for the precision end. Almost none of this is proprietary to a metrology company. Image sensors come from the same merchant suppliers that serve every camera market, which is the root cause of the cost curve and of the 2.36 per cent realisation growth.</p><p><strong>The platform layer.</strong> Granite and ceramic structures, air bearings, motorised stages, thermal compensation, and the mechanical engineering that makes a CMM a CMM. This is where the traditional incumbents are genuinely excellent and where the barriers to entry remain real — but it is also where capital intensity sits highest and where growth is slowest.</p><p><strong>The software layer.</strong> Measurement programming, nominal comparison, reporting, and increasingly the closed loop back to the machine tool that cut the part. This is the layer with the highest gross margin, the stickiest customer relationship and the lowest capital requirement. A buyer who has programmed a thousand part numbers in one software environment does not change it to save ten per cent on hardware. Every large acquisition in this market has been, in substance, the purchase of a software installed base.</p><p><strong>The integration layer.</strong> Cells, fixturing, robot-mounted sensors, line integration and commissioning. Low margin percentage, high absolute value, and the layer through which most mid-market volume actually reaches the end user.</p><h3>Who buys, and how that changes the price</h3><p>Three buyers. The <strong>quality function</strong> in a mature plant buys a CMM to qualify a process, has a capital budget, compares specifications and measures payback in years. The <strong>manufacturing engineer</strong> buys inline gauging to control a process, measures payback in scrap and rework avoided, and cares about the takt time available rather than about the best achievable uncertainty. The <strong>corporate metrology standard owner</strong> — common in automotive and aerospace — does not buy a system at all; they approve which systems may be bought, and their approval becomes the constraint that determines what the other two can specify.</p><p>That third buyer is the one the whole competitive structure turns on. Acceptance is granted centrally, centrally granted acceptance favours incumbency, and incumbency is currently held by tactile measurement. Machine vision is winning on volume because it is entering applications where no standard was previously imposed — new lines, new products, new plants with no legacy programme to match. It is not winning by displacing approved systems, and any forecast that assumes it will is assuming a regulatory change this report does not yet observe.</p>"
  investment_activity_content: "<h2 id=\"section-ia\">7. Investment activity</h2><p>Capital in this market has one purpose: to buy the software and the installed base that make hardware replaceable and the customer not.</p><p>The logic follows from the numbers. Realisation growth of 2.36 per cent is the lowest in this cluster because the sensing layer rides a declining semiconductor cost curve. A supplier that sells only the sensing layer is selling a deflating product. A supplier that sells the measurement programme, the nominal model comparison, the reporting and the closed loop back to the machine tool is selling something the customer cannot re-shop on price without rewriting a thousand programmes — and that is what every large transaction in this field has actually bought.</p><h3>Why the incumbents keep buying</h3><p>The pattern is consistent across the field. <strong>Hexagon</strong> has assembled a portfolio spanning tactile metrology, laser scanning, portable measurement and the software that connects them, and its strategic logic is explicitly software-led: own the interpretation layer so the hardware decision follows. <strong>Carl Zeiss</strong> has concentrated on the accuracy end and on computed tomography, where the capability barrier is strongest and the price pressure weakest. <strong>Mitutoyo</strong> has defended the volume end of tactile measurement, which is the position most exposed to the shift this report describes.</p><p>The direction of travel is one-way. There is no plausible strategy in this market that consists of making a better CMM and selling it at a higher price. There are only two: bundle a multisensor platform to protect an installed base, or own the software layer the base depends on. Most large suppliers are attempting both.</p><h3>What is not being bought</h3><p>Mechanical capacity. This is a low capital-intensity business at the platform layer — the granite, the bearings and the assembly are expensive per unit and cheap relative to revenue, in the way instrument businesses generally are. The scarce inputs are measurement software engineers, calibration and traceability infrastructure, and application knowledge in specific verticals. None of those can be bought quickly, and all of them are what an acquirer is actually paying for.</p><h3>Where capital goes next</h3><p>Two places. <strong>Closed-loop software</strong>, which feeds a measurement result back to the machine tool that cut the part: this is the one application in which inline metrology stops being an inspection cost and becomes a yield instrument, and it is where the highest multiples are being paid. And <strong>computed tomography</strong>, which is the only remaining capability barrier in dimensional measurement — internal features no optical or contact method can reach. Neither is cheap, and both are defensible in a way that another vision sensor is not.</p>"
  technology_innovation_content: "<h2 id=\"section-ti\">4. Technology: why vision overtakes the CMM</h2><p>The coordinate measuring machine is the most precise industrial measuring instrument ever built at scale. A granite bridge on air bearings, driven by a scanning probe that touches the surface, sitting in a temperature-controlled room at twenty degrees. It resolves features to fractions of a micron.</p><p>It is also, structurally, in the wrong place. A CMM requires the part to be brought to it, fixtured, and measured in a batch, and the result arrives after the process that made the part has moved on. That is fine for qualification and for first-article inspection. It is useless for closed-loop control of a machining cell, where the correction has to happen before the next part is cut.</p><h3>What changed</h3><p>Three things, none of them a breakthrough in precision.</p><p><strong>Sensor cost.</strong> Machine vision is built from image sensors, optics and compute — all semiconductor products on a cost curve that declines every year. A dimensional vision system that cost six figures a decade ago is now a mid-five-figure installation. That is what turned 100 per cent inspection from an aspiration into a purchase order.</p><p><strong>Structured light and laser profiling.</strong> A single camera gives a 2D measurement. Structured light or a laser line adds the third dimension, which is what makes vision useful for dimensional metrology rather than only for presence and defect detection. The mathematics is not new; the cost of the projector and the compute to process the pattern is.</p><p><strong>Enough accuracy, relative to the tolerance.</strong> This is the point most often missed. Vision does not need to match a CMM. It needs to be adequate against the tolerance being inspected — and as tolerances in automotive body assembly, battery module dimensions and electronics packaging settled in the tens of microns rather than the single microns, vision crossed the threshold for a large fraction of the applications that matter by volume.</p><h3>The techniques, and where each sits</h3><ul><li><strong>Tactile CMM</strong> — the reference standard. Grows 4.60 per cent and holds the top of the accuracy market.</li><li><strong>Multisensor cells</strong> — tactile and optical on one platform. The incumbents' bridge, and the reason the CMM line here grows at all.</li><li><strong>Optical and laser gauging</strong> — in-process measurement of a single dimension at line speed. Grows 6.40 per cent and is the least discussed high-volume category in the market.</li><li><strong>Laser scanning and trackers</strong> — large-volume measurement: aircraft assembly, shipbuilding, large castings. Grows 5.15 per cent.</li><li><strong>White-light and interferometric</strong> — surface topography and roughness at nanometre scale. Grows 9.10 per cent from a small base, driven by optics, semiconductor and precision bearing applications.</li><li><strong>Surface and form instruments</strong> — roughness, roundness, contour. The most mature category at 4.10 per cent.</li></ul><h3>Where the constraint now sits</h3><p>In acceptance, not capability. An optical system's measurement uncertainty depends on the surface — its finish, its colour, its translucency, its angle to the sensor — in a way a contact probe's does not. That makes the uncertainty budget harder to construct and harder to defend, and until it is defended the system cannot be used for a conformance decision against a drawing.</p><p>Industrial computed tomography is the wildcard on the other side of that line: it measures internal features no optical or tactile method can reach, which makes it the only genuine capability barrier left in dimensional metrology. We do not size it here; it belongs to our industrial radiography report, and adding the two would double-count it.</p>"
  market_risk_content: "<h2 id=\"section-mr\">8. Market risk and scenarios</h2><h3>1. Nothing here is mandatory — and that is the defining risk</h3><p>In our non-destructive testing and ultrasonic reports, demand is anchored to inspection codes that specify the method and the interval. Nothing equivalent exists in inline metrology. ISO 10360 tells you how to verify a CMM; it does not tell you to buy one. The purchase is a capital decision justified by scrap rate, throughput or a new product launch, and every one of those is cyclical.</p><p>The exposure is concentrated. Automotive and EV is 26.0 per cent of value and electronics and semiconductors 19.5 per cent: <strong>45.5 per cent of this market is tied directly to two capital expenditure cycles</strong>. A pause in either removes volume the model cannot recover from price. This is the single largest difference between this report and its cluster siblings, and it is why our conservative case is a capital-cycle case rather than a pricing case.</p><h3>2. Realisation is already the lowest in the cluster, and can go lower</h3><p>Realised spend per system grows 2.36 per cent — well below the 3.38 per cent of non-destructive testing and the 3.59 per cent of ultrasonic. The cause is structural: machine vision is built from merchant image sensors and compute on a declining cost curve. At the low end of this market, realisation is plausibly flat or negative in nominal terms, with growth carried entirely by units. A supplier shifting mix toward vision is buying unit growth with price give-up, and the only offset is software and service margin.</p><h3>3. Acceptance is the fence, and it cuts both ways</h3><p>An optical system cannot be used for a conformance decision against a drawing until its measurement uncertainty is characterised and defended — a harder problem than for contact metrology, because the uncertainty depends on the surface, its finish and its geometry relative to the sensor. The standards for optical 3D systems are less mature than the ISO 10360 series for tactile machines.</p><p>The important nuance is that <em>this has not slowed vision down</em>. Most of vision’s growth is in process-control applications that were never subject to a conformance standard in the first place — new lines, new products, plants with no legacy to match. The fence confines vision to the biggest part of the market by volume and excludes it from the part where price is highest. If the optical standards mature, vision’s addressable market expands into conformance work and the optimistic scenario is in play. If they stall, the crossover in 2028 still happens on schedule and the ceiling stays where it is.</p><h3>4. The incumbents are not standing still</h3><p>The multisensor platform is the tactile incumbents’ answer, and it is a good one: a cell carrying both a contact probe and an optical sensor concedes the simple measurements to vision while keeping the customer. If that bundling is executed well, the CMM line in this report could grow faster than 4.60 per cent and the crossover could slip past 2028. Our base case assumes bundling slows the CMM’s decline but does not stop it.</p><h3>5. EV capex timing</h3><p>Battery module and motor inspection is the largest genuinely new demand pool in this market and the reason the volume story holds. Battery lines are dimensionally demanding in a way that did not exist at scale a decade ago, and they are specified with inline inspection from the start. That demand is real and it is not yet fully built. It is also tied to a vehicle transition whose pace differs sharply by region, and a slower EV capital cycle is the most direct route to the conservative case.</p><h3>6. Trade and regional concentration</h3><p>We classify geopolitical exposure as high: Asia-Pacific is 38.5 per cent of value and 49.4 per cent of the increment, sensing components cross borders, and a large share of assembly and integration sits in a small number of countries. A tariff or export-control change in this chain does not merely raise cost; it moves where lines are built and therefore where the inspection is bought.</p><h3>Scenarios</h3><table><thead><tr><th>Scenario</th><th>2033 (USD m)</th><th>CAGR</th><th>Driver</th></tr></thead><tbody><tr><td>Conservative</td><td>12,600</td><td>4.81%</td><td>Automotive and electronics capital spending pauses; vision deflation outruns volume</td></tr><tr><td>Base</td><td>14,158</td><td>6.35%</td><td>Machine vision holds 8.70%; EV battery and motor inspection compounds on schedule</td></tr><tr><td>Optimistic</td><td>15,700</td><td>7.74%</td><td>Optical systems win conformance acceptance, lifting both placement and specification</td></tr></tbody></table>"
  regulatory_landscape_content: "<h2 id=\"section-rl\">6. Regulatory landscape: acceptance is the fence</h2><p>Metrology is regulated less by what you may measure than by what your measurement may be used to certify. That distinction is the whole of the regulatory story in this market.</p><h3>The acceptance standards</h3><ul><li><strong>ISO 10360</strong> — the acceptance and reverification tests for coordinate measuring machines. It defines how a CMM's performance is demonstrated and re-demonstrated, and it is the technical foundation on which every tactile machine in this report is sold. It has no equivalent maturity for optical systems.</li><li><strong>ISO 14253-1</strong> — decision rules for proving conformance or nonconformance with specification, which force the uncertainty of the measurement into the accept-or-reject decision. This is the standard that makes measurement uncertainty consequential rather than academic: a system whose uncertainty you cannot state cannot be used to declare a part good.</li><li><strong>VDI/VDE 2634</strong> — the German-led series covering optical 3D measuring systems, spanning imaging systems and surface scanning. It is the closest thing the optical world has to ISO 10360, and its narrower adoption is precisely the gap that confines vision to process control.</li><li><strong>ISO 17025</strong> — competence of calibration and testing laboratories, which governs who may certify a traceable result.</li><li><strong>IATF 16949 and measurement systems analysis</strong> — the automotive quality regime, which requires that a gauge be shown capable before it is used to control a characteristic. In practice MSA, and gauge repeatability and reproducibility in particular, is the gate that decides which measurement technology an automotive plant may deploy on a safety- or fit-critical feature.</li></ul><h3>Why the fence has not slowed vision down</h3><p>This is the most important point in the chapter, and it is easy to get backwards.</p><p>Machine vision cannot currently be used for the conformance decisions that ISO 14253-1 and automotive MSA govern. On a strict reading, that should cap its growth. It has not, because <strong>most of the growth in this market is not in conformance work at all.</strong> It is in process control on new lines, in new products and in plants where no legacy measurement programme exists to be matched. A battery module line being specified today imposes no acceptance requirement inherited from a tactile programme; it specifies inline inspection because inline inspection is what the line needs.</p><p>So the fence protects the highest-value slice of the market and excludes vision from it, while vision grows around the fence into the larger slice by volume. The scenario that changes is not a technology scenario; it is a standards scenario. Mature optical acceptance standards would open conformance work to vision and justify the optimistic case. Their absence caps the price ceiling without capping the units.</p><h3>What to watch</h3><p>Three signals. First, any broadening of VDI/VDE 2634 or a new ISO work item giving optical systems a defensible route to acceptance testing — the single most valuable regulatory development this market could see. Second, revision of automotive measurement systems analysis guidance to admit optical systems for controlled characteristics, which would move the largest single end market. Third, traceability infrastructure: a supplier who can deliver a documented uncertainty budget for an optical measurement of a real surface, not a calibration artefact, has effectively written its own acceptance argument.</p>"
  faq_question_1: "What is the size of the global inline metrology market?"
  faq_answer_1: "USD 8,650 million in 2025, rising to USD 14,158 million in 2033 at a compound annual growth rate of 6.35 per cent. Systems placed grow from 96,500 to 131,060 at 3.90 per cent, and realised spend per system rises from USD 89,637 to USD 108,012 at 2.36 per cent."
  faq_question_2: "When does machine vision overtake the coordinate measuring machine?"
  faq_answer_2: "During 2028. Inline CMM and multisensor cells are USD 2,335 million in 2025 growing 4.60 per cent; machine vision dimensional systems are USD 2,119 million growing 8.70 per cent. Those two rates cross in the middle of 2028, and by 2033 machine vision holds 29.2 per cent of value against the CMM share of 23.6 per cent."
  faq_question_3: "Why is this market volume-led when the rest of its cluster is price-led?"
  faq_answer_3: "Because inline metrology is not a replacement market. 56.3 per cent of the USD 5,506 million increment is additional systems rather than price and mix, against 60.1 and 62.1 per cent price and mix in our ultrasonic and non-destructive testing reports. Those markets replace instruments already owned against a code obligation; inline metrology is being installed where nothing was installed before, which is why realised-spend growth here is only 2.36 per cent."
  faq_question_4: "Does this report include industrial computed tomography?"
  faq_answer_4: "No. Industrial computed tomography measures internal geometry non-destructively and sits at the boundary between metrology and radiographic inspection. It is sized in our industrial radiography and digital X-ray inspection report, and adding the two would double-count it. This report and that one are siblings in the same cluster, not increments of each other."

global_schema:
  organization:
    "@id": MIWS-root#organization
    "@type": Organization
    url: "https://www.pheonixresearch.com/"
    name: Pheonix Research
  website:
    "@id": MIWS-root
    "@type": WebSite
    url: "https://www.pheonixresearch.com/"
    name: Pheonix Research
  api:
    "@id": MIWS-root#api
    "@type": WebAPI
    url: "https://www.pheonixresearch.com/"

graph:
  node_id: MIMR-node_105c505
  graph_node_endpoint: "https://graph.pheonixresearch.com/api/v1/query/live/node/MIMR-node_105c505"
  graph_snapshot: "https://graph.pheonixresearch.com/api/v1/read/live/graph"
  graph_exclusions: "https://graph.pheonixresearch.com/api/v1/read/live/exclusions"
  graph_meta: "https://graph.pheonixresearch.com/api/v1/read/live/meta"
  graph_bundle: "https://graph.pheonixresearch.com/api/v1/read/live/bundle"

discovery:
  discovery_json: "https://www.pheonixresearch.com/.well-known/pheonix-discovery.json"
  llms_txt: "https://www.pheonixresearch.com/llms.txt"
  sitemap: "https://www.pheonixresearch.com/sitemap.xml"
---
# Global Inline Metrology Market Report, Size, Share and Forecast 2026–2033: Machine Vision Overtakes the CMM in 2028

- **Market Size (Base Year):** USD 8650 Million
- **Forecast Value:** USD 14158 Million
- **CAGR:** 6.35%
- **Forecast Period:** 2026 - 2033
- **Coverage:** Global - Global Inline Metrology

## Executive Summary

The global inline metrology market is USD 8,650 million in 2025 and reaches USD 14,158 million in 2033, a compound annual growth rate of 6.35 per cent. The headline hides the most interesting thing in the market, which is that its largest technology is about to lose first place to the second largest.Machine vision passes the coordinate measuring machine in 2028. Inline CMM and multisensor cells are 27.0 per cent of 2025 value at USD 2,335 million, growing 4.60 per cent. Machine vision dimensional systems are 24.5 per cent at USD 2,119 million, growing 8.70 per cent. Two growth rates and two bases put the crossover in the middle of 2028. By 2033 machine vision holds 29.2 per cent of the market and 36.5 per cent of the increment; the CMM holds 23.6 per cent and 18.3 per cent. The second-largest technology adds USD 2,012 million; the largest adds USD 1,011 million.The reason is where measurement is being done. The CMM is a measurement-room instrument: granite, air bearings, temperature control, a part brought to it in a batch and fixtured. Machine vision is an inline instrument: no fixture, no batch, every part, and a result in the takt time of the line. The commercial argument is not accuracy — a good CMM is more accurate. It is that 100 per cent inspection became affordable as the sensor cost fell, and a sampled inspection cannot be closed-loop.The second finding runs against the rest of this cluster. We found price and mix to be 60.1 per cent of ultrasonic growth and 62.1 per cent of non-destructive testing growth, because those are replacement markets where an instrument displaces one already owned against a code obligation. Here it is the reverse: 56.3 per cent of the USD 5,506 million increment is additional systems — 96,500 placed in 2025 against 131,060 in 2033 at 3.90 per cent — and realised spend per system contributes only 43.7 per cent at 2.36 per cent. Inline metrology is not replacing what was already bought. It is being installed where nothing was installed before.That produces a different risk profile. Realisation growth of 2.36 per cent is the lowest in the cluster, because machine vision hardware rides a semiconductor cost curve that goes down rather than up. A supplier shifting its mix toward vision buys unit growth at the price of price growth. The trade only works if the software and service attached to those units carries a margin the hardware does not — which is exactly why every CMM incumbent in this market has spent a decade acquiring software.The geography is unambiguous. Asia-Pacific is 38.5 per cent of value and takes 49.4 per cent of the increment; half the world’s growth in this market lands in one region, because that is where the line is being built and where the inspection is being specified alongside it. Europe grows 4.55 per cent on a dense installed base. Automotive and EV is the largest end market at 26.0 per cent, and it is also the most cyclical buyer in the cluster — a fact that matters more here than in any code-mandated inspection market, because nothing in this report is bought because a regulation says it must be.

## Market Report

## 1. What this report measures

Inline metrology is measurement performed where the part is made, rather than in a room set aside for it. It covers the systems that establish a dimensional or surface result at or beside a production line — inline coordinate measuring machines and multisensor cells, machine vision dimensional systems, laser scanners and trackers, optical and laser gauging, surface and form instruments, and white-light and interferometric systems.

This report replaces a legacy page on the same subject that carried unfilled template values and a 2021–2028 horizon. Everything below is rebuilt: a new base year, a new model, and a forecast to 2033.

**Scope boundary, and one overlap to declare.** Industrial computed tomography measures internal geometry non-destructively and sits at the boundary between metrology and radiographic inspection. We do *not* size CT here. It is counted in our industrial radiography and digital X-ray inspection report, and adding the two would double-count it. That report and this one are siblings in the same cluster, not increments of each other.

### The finding, and why it is the opposite of its siblings

**Machine vision passes the coordinate measuring machine in 2028.**

Inline CMM and multisensor cells are the largest technology in this market at USD 2,335 million and 27.0 per cent of 2025 value, growing 4.60 per cent. Machine vision dimensional systems are second at USD 2,119 million and 24.5 per cent, growing 8.70 per cent. Those two rates put machine vision ahead during 2028, and by 2033 it holds 29.2 per cent of the market against the CMM’s 23.6 per cent. Machine vision adds **USD 2,012 million** over the window; the CMM adds **USD 1,011 million**. The second-largest technology adds twice what the largest does.

The second finding is where this report departs from the rest of its cluster, and it is worth being explicit about because it would be easy to assume otherwise.

**Here, volume does most of the work.** The market adds USD 5,506 million by 2033, and **USD 3,098 million of it — 56.3 per cent — is additional systems**: 96,500 placed in 2025 against 131,060 in 2033, growth of 3.90 per cent a year. Realisation contributes USD 2,408 million, or 43.7 per cent, rising from USD 89,637 to USD 108,012 per system at 2.36 per cent.

Our ultrasonic and non-destructive testing reports in this cluster found the opposite — 60.1 and 62.1 per cent price and mix respectively, because those are markets where the instrument is replacing one already owned against a code obligation. Inline metrology is not replacing something already bought. It is being installed where nothing was installed before: at the end of a machining cell that previously sampled one part in fifty, on a battery line that did not exist a decade ago, in an electronics plant where 100 per cent inspection became affordable as the sensor cost fell. That is a genuine volume story, and it is the reason this market’s risk profile is different from its siblings in almost every respect.

The consequence for suppliers is uncomfortable. Realisation growth of 2.36 per cent is the lowest in the cluster, because machine vision hardware rides a semiconductor cost curve that goes down. A supplier whose mix shifts toward vision is buying growth in units at the price of growth in price. That trade is worth making only if the software and service attached to those units carries a margin the hardware does not — which is precisely why every CMM incumbent in this market has spent a decade buying software companies.

## 2. Market size, forecast and the technology split

- Market value: USD 8,650 million in 2025 to USD 14,158 million in 2033 at 6.35 per cent.

- Systems placed: 96,500 to 131,060 at 3.90 per cent.

- Realised spend per system: USD 89,637 to USD 108,012 at 2.36 per cent.

Source: Pheonix Research model. Systems placed multiplied by realised spend per system, 2025–2033.

### Technology split

 | Technology |  | 2025 (USD m) |  | Share |  | 2033 (USD m) |  | Share |  | CAGR |  | Increment |  | Share of increment |

 | Machine vision dimensional (2D and 3D) |  | 2,119 |  | 24.5% |  | 4,131 |  | 29.2% |  | 8.70% |  | 2,012 |  | 36.5% |

 | Inline CMM and multisensor cells |  | 2,335 |  | 27.0% |  | 3,346 |  | 23.6% |  | 4.60% |  | 1,011 |  | 18.3% |

 | Optical and laser gauging (in-process) |  | 1,341 |  | 15.5% |  | 2,203 |  | 15.6% |  | 6.40% |  | 862 |  | 15.6% |

 | Laser scanning and trackers |  | 1,384 |  | 16.0% |  | 2,068 |  | 14.6% |  | 5.15% |  | 684 |  | 12.4% |

 | White-light and interferometric |  | 606 |  | 7.0% |  | 1,217 |  | 8.6% |  | 9.10% |  | 611 |  | 11.1% |

 | Surface and form (roughness, roundness, contour) |  | 865 |  | 10.0% |  | 1,193 |  | 8.4% |  | 4.10% |  | 328 |  | 6.0% |

 | **Total** |  | **8,650** |  | **100.0%** |  | **14,158** |  | **100.0%** |  | **6.35%** |  | **5,508** |  | **99.9%** |

Increment components sum to USD 5,508 million against a total of USD 5,506 million; the difference is rounding.

Source: Pheonix Research model. Machine vision adds USD 2,012 million over the window; inline CMM adds USD 1,011 million.

### Where the USD 5,506 million comes from

 | Component |  | USD m |  | Share of increment |

 | 2025 market value |  | 8,650 |  | — |

 | Volume effect: 34,560 additional systems × USD 89,637 |  | +3,098 |  | 56.3% |

 | Price and mix effect |  | +2,408 |  | 43.7% |

 | **2033 market value** |  | **14,158** |  | **100.0%** |

### End markets

 | End market |  | 2025 (USD m) |  | Share |

 | Automotive, commercial vehicle and EV |  | 2,249 |  | 26.0% |

 | Electronics and semiconductors |  | 1,687 |  | 19.5% |

 | Industrial machinery and metal forming |  | 1,384 |  | 16.0% |

 | Aerospace |  | 1,341 |  | 15.5% |

 | Medical devices |  | 735 |  | 8.5% |

 | Energy and power |  | 649 |  | 7.5% |

 | Additive manufacturing, optics, defence and other |  | 606 |  | 7.0% |

### Regional structure

 | Region |  | 2025 (USD m) |  | Share |  | 2033 (USD m) |  | Share |  | CAGR |  | Share of increment |

 | Asia-Pacific |  | 3,330 |  | 38.5% |  | 6,050 |  | 42.7% |  | 7.75% |  | 49.4% |

 | North America |  | 2,119 |  | 24.5% |  | 3,228 |  | 22.8% |  | 5.40% |  | 20.1% |

 | Europe |  | 2,163 |  | 25.0% |  | 3,088 |  | 21.8% |  | 4.55% |  | 16.8% |

 | Rest of world |  | 1,038 |  | 12.0% |  | 1,794 |  | 12.7% |  | 7.08% |  | 13.7% |

 | **Total** |  | **8,650** |  | **100.0%** |  | **14,160** |  | **100.0%** |  | **6.35%** |  | **100.0%** |

Asia-Pacific is 38.5 per cent of value and takes **49.4 per cent of the increment** — half the world’s growth in this market lands in one region. That is where the line is being built and where the inspection is being specified with it. Europe is the slowest at 4.55 per cent for the same reason Europe is slow in every report in this cluster: the installed base is already dense.

### Scenarios

 | Scenario |  | 2033 (USD m) |  | CAGR |  | What has to be true |

 | Conservative |  | 12,600 |  | 4.81% |  | Automotive and electronics capital spending pauses; vision hardware deflation outruns volume growth |

 | Base |  | 14,158 |  | 6.35% |  | Machine vision holds 8.70 per cent; EV battery and motor inspection demand compounds on schedule |

 | Optimistic |  | 15,700 |  | 7.74% |  | Optical systems win conformance acceptance in regulated applications, lifting both unit placement and specification |

## Table of Contents

What this report measures
Market size, forecast and the technology split
Value chain
Technology: why vision overtakes the CMM
Competitive landscape
Regulatory landscape: acceptance is the fence
Investment activity
Market risk and scenarios

## Competitive Landscape

5. Competitive landscape
This is a moderately consolidated market with an unusual shape: the incumbents are measurement companies that are buying vision capability, and the challengers are vision companies that are buying measurement credibility. Both directions are rational and both are expensive.
Our estimate: the five largest suppliers account for roughly 45 to 50 per cent of system value. We label it our own estimate because the denominator cannot be observed directly — a meaningful share of inline gauging is bundled into machine tools, assembly cells and integrators’ projects and never appears as a metrology line item.

Hexagon AB — the broadest position in the market, spanning tactile CMMs, laser scanning and trackers, portable measurement and the software layer that ties them together. Its strategic logic is explicit: own the measurement and the software that interprets it, so that the hardware choice follows.
Carl Zeiss — the deepest position in high-accuracy tactile metrology and industrial computed tomography, with a strong optical portfolio. Concentrated in the applications where accuracy, not throughput, is the buying criterion.
Mitutoyo — the volume leader in tactile CMMs, particularly strong in Asia, and the supplier most exposed to the shift this report describes.
Cognex — the purest machine vision position, concentrated in the vision and identification hardware that is taking share. Not a metrology company by heritage, which is precisely why it is winning in the fastest-growing technology.
Keyence, Nikon Metrology, Renishaw, Faro, LMI Technologies and Jenoptik — specialist positions across vision sensors, encoders and probes, portable scanning and 3D scanning, each defending an application where its specification matters more than its scale.

How the competition actually resolves
On the software. A modern CMM and a modern vision system are both capable. What a buyer cannot easily change is the measurement programme, the nominal CAD model, the reporting format and the operator’s familiarity — which is why the installed software base is the strongest thing any incumbent in this market owns. Every large acquisition in this field has been, at bottom, a software acquisition.
On the multisensor platform. The incumbents’ answer to vision is not to compete with it but to absorb it: a cell that carries a tactile probe and an optical sensor on the same bridge, so that a feature can be measured by whichever technique suits it. This is technically the right answer and commercially a defensive one — it protects the installed base while conceding that vision wins the simple measurements.
On cost, at the low end. Vision sensors are semiconductor products with a semiconductor cost curve. The low end of this market is being commoditised faster than any other segment we cover in this cluster, and it is where the 2.36 per cent realisation number comes from.
On acceptance. The decisive arena, and the one covered in the regulatory chapter: a technology that cannot demonstrate measurement uncertainty to a recognised standard cannot be used for a conformance decision, however good its image.

- **Structure:** moderately_consolidated
- **Tier 1 Players:** 8
- **Intensity:** high

## Value Chain

3. Value chain
Four layers, and the value is not where the physics is.
The sensing layer. Image sensors and optics for vision; encoders and scales for motion; probe and stylus assemblies for tactile; laser sources and interferometer optics for the precision end. Almost none of this is proprietary to a metrology company. Image sensors come from the same merchant suppliers that serve every camera market, which is the root cause of the cost curve and of the 2.36 per cent realisation growth.
The platform layer. Granite and ceramic structures, air bearings, motorised stages, thermal compensation, and the mechanical engineering that makes a CMM a CMM. This is where the traditional incumbents are genuinely excellent and where the barriers to entry remain real — but it is also where capital intensity sits highest and where growth is slowest.
The software layer. Measurement programming, nominal comparison, reporting, and increasingly the closed loop back to the machine tool that cut the part. This is the layer with the highest gross margin, the stickiest customer relationship and the lowest capital requirement. A buyer who has programmed a thousand part numbers in one software environment does not change it to save ten per cent on hardware. Every large acquisition in this market has been, in substance, the purchase of a software installed base.
The integration layer. Cells, fixturing, robot-mounted sensors, line integration and commissioning. Low margin percentage, high absolute value, and the layer through which most mid-market volume actually reaches the end user.
Who buys, and how that changes the price
Three buyers. The quality function in a mature plant buys a CMM to qualify a process, has a capital budget, compares specifications and measures payback in years. The manufacturing engineer buys inline gauging to control a process, measures payback in scrap and rework avoided, and cares about the takt time available rather than about the best achievable uncertainty. The corporate metrology standard owner — common in automotive and aerospace — does not buy a system at all; they approve which systems may be bought, and their approval becomes the constraint that determines what the other two can specify.
That third buyer is the one the whole competitive structure turns on. Acceptance is granted centrally, centrally granted acceptance favours incumbency, and incumbency is currently held by tactile measurement. Machine vision is winning on volume because it is entering applications where no standard was previously imposed — new lines, new products, new plants with no legacy programme to match. It is not winning by displacing approved systems, and any forecast that assumes it will is assuming a regulatory change this report does not yet observe.

- **Model:** hybrid
- **Distribution:** distributor_led
- **Supply Complexity:** moderate

## Investment Activity

7. Investment activity
Capital in this market has one purpose: to buy the software and the installed base that make hardware replaceable and the customer not.
The logic follows from the numbers. Realisation growth of 2.36 per cent is the lowest in this cluster because the sensing layer rides a declining semiconductor cost curve. A supplier that sells only the sensing layer is selling a deflating product. A supplier that sells the measurement programme, the nominal model comparison, the reporting and the closed loop back to the machine tool is selling something the customer cannot re-shop on price without rewriting a thousand programmes — and that is what every large transaction in this field has actually bought.
Why the incumbents keep buying
The pattern is consistent across the field. Hexagon has assembled a portfolio spanning tactile metrology, laser scanning, portable measurement and the software that connects them, and its strategic logic is explicitly software-led: own the interpretation layer so the hardware decision follows. Carl Zeiss has concentrated on the accuracy end and on computed tomography, where the capability barrier is strongest and the price pressure weakest. Mitutoyo has defended the volume end of tactile measurement, which is the position most exposed to the shift this report describes.
The direction of travel is one-way. There is no plausible strategy in this market that consists of making a better CMM and selling it at a higher price. There are only two: bundle a multisensor platform to protect an installed base, or own the software layer the base depends on. Most large suppliers are attempting both.
What is not being bought
Mechanical capacity. This is a low capital-intensity business at the platform layer — the granite, the bearings and the assembly are expensive per unit and cheap relative to revenue, in the way instrument businesses generally are. The scarce inputs are measurement software engineers, calibration and traceability infrastructure, and application knowledge in specific verticals. None of those can be bought quickly, and all of them are what an acquirer is actually paying for.
Where capital goes next
Two places. Closed-loop software, which feeds a measurement result back to the machine tool that cut the part: this is the one application in which inline metrology stops being an inspection cost and becomes a yield instrument, and it is where the highest multiples are being paid. And computed tomography, which is the only remaining capability barrier in dimensional measurement — internal features no optical or contact method can reach. Neither is cheap, and both are defensible in a way that another vision sensor is not.

- **Trend:** rising
- **Capital Intensity:** medium
- **Recent M&A:** yes

## Technology & Innovation

4. Technology: why vision overtakes the CMM
The coordinate measuring machine is the most precise industrial measuring instrument ever built at scale. A granite bridge on air bearings, driven by a scanning probe that touches the surface, sitting in a temperature-controlled room at twenty degrees. It resolves features to fractions of a micron.
It is also, structurally, in the wrong place. A CMM requires the part to be brought to it, fixtured, and measured in a batch, and the result arrives after the process that made the part has moved on. That is fine for qualification and for first-article inspection. It is useless for closed-loop control of a machining cell, where the correction has to happen before the next part is cut.
What changed
Three things, none of them a breakthrough in precision.
Sensor cost. Machine vision is built from image sensors, optics and compute — all semiconductor products on a cost curve that declines every year. A dimensional vision system that cost six figures a decade ago is now a mid-five-figure installation. That is what turned 100 per cent inspection from an aspiration into a purchase order.
Structured light and laser profiling. A single camera gives a 2D measurement. Structured light or a laser line adds the third dimension, which is what makes vision useful for dimensional metrology rather than only for presence and defect detection. The mathematics is not new; the cost of the projector and the compute to process the pattern is.
Enough accuracy, relative to the tolerance. This is the point most often missed. Vision does not need to match a CMM. It needs to be adequate against the tolerance being inspected — and as tolerances in automotive body assembly, battery module dimensions and electronics packaging settled in the tens of microns rather than the single microns, vision crossed the threshold for a large fraction of the applications that matter by volume.
The techniques, and where each sits

Tactile CMM — the reference standard. Grows 4.60 per cent and holds the top of the accuracy market.
Multisensor cells — tactile and optical on one platform. The incumbents’ bridge, and the reason the CMM line here grows at all.
Optical and laser gauging — in-process measurement of a single dimension at line speed. Grows 6.40 per cent and is the least discussed high-volume category in the market.
Laser scanning and trackers — large-volume measurement: aircraft assembly, shipbuilding, large castings. Grows 5.15 per cent.
White-light and interferometric — surface topography and roughness at nanometre scale. Grows 9.10 per cent from a small base, driven by optics, semiconductor and precision bearing applications.
Surface and form instruments — roughness, roundness, contour. The most mature category at 4.10 per cent.

Where the constraint now sits
In acceptance, not capability. An optical system’s measurement uncertainty depends on the surface — its finish, its colour, its translucency, its angle to the sensor — in a way a contact probe’s does not. That makes the uncertainty budget harder to construct and harder to defend, and until it is defended the system cannot be used for a conformance decision against a drawing.
Industrial computed tomography is the wildcard on the other side of that line: it measures internal features no optical or tactile method can reach, which makes it the only genuine capability barrier left in dimensional metrology. We do not size it here; it belongs to our industrial radiography report, and adding the two would double-count it.

- **Innovation:** high
- **Patent Activity:** high
- **Maturity:** growth

## Market Risk

8. Market risk and scenarios
1. Nothing here is mandatory — and that is the defining risk
In our non-destructive testing and ultrasonic reports, demand is anchored to inspection codes that specify the method and the interval. Nothing equivalent exists in inline metrology. ISO 10360 tells you how to verify a CMM; it does not tell you to buy one. The purchase is a capital decision justified by scrap rate, throughput or a new product launch, and every one of those is cyclical.
The exposure is concentrated. Automotive and EV is 26.0 per cent of value and electronics and semiconductors 19.5 per cent: 45.5 per cent of this market is tied directly to two capital expenditure cycles. A pause in either removes volume the model cannot recover from price. This is the single largest difference between this report and its cluster siblings, and it is why our conservative case is a capital-cycle case rather than a pricing case.
2. Realisation is already the lowest in the cluster, and can go lower
Realised spend per system grows 2.36 per cent — well below the 3.38 per cent of non-destructive testing and the 3.59 per cent of ultrasonic. The cause is structural: machine vision is built from merchant image sensors and compute on a declining cost curve. At the low end of this market, realisation is plausibly flat or negative in nominal terms, with growth carried entirely by units. A supplier shifting mix toward vision is buying unit growth with price give-up, and the only offset is software and service margin.
3. Acceptance is the fence, and it cuts both ways
An optical system cannot be used for a conformance decision against a drawing until its measurement uncertainty is characterised and defended — a harder problem than for contact metrology, because the uncertainty depends on the surface, its finish and its geometry relative to the sensor. The standards for optical 3D systems are less mature than the ISO 10360 series for tactile machines.
The important nuance is that this has not slowed vision down. Most of vision’s growth is in process-control applications that were never subject to a conformance standard in the first place — new lines, new products, plants with no legacy to match. The fence confines vision to the biggest part of the market by volume and excludes it from the part where price is highest. If the optical standards mature, vision’s addressable market expands into conformance work and the optimistic scenario is in play. If they stall, the crossover in 2028 still happens on schedule and the ceiling stays where it is.
4. The incumbents are not standing still
The multisensor platform is the tactile incumbents’ answer, and it is a good one: a cell carrying both a contact probe and an optical sensor concedes the simple measurements to vision while keeping the customer. If that bundling is executed well, the CMM line in this report could grow faster than 4.60 per cent and the crossover could slip past 2028. Our base case assumes bundling slows the CMM’s decline but does not stop it.
5. EV capex timing
Battery module and motor inspection is the largest genuinely new demand pool in this market and the reason the volume story holds. Battery lines are dimensionally demanding in a way that did not exist at scale a decade ago, and they are specified with inline inspection from the start. That demand is real and it is not yet fully built. It is also tied to a vehicle transition whose pace differs sharply by region, and a slower EV capital cycle is the most direct route to the conservative case.
6. Trade and regional concentration
We classify geopolitical exposure as high: Asia-Pacific is 38.5 per cent of value and 49.4 per cent of the increment, sensing components cross borders, and a large share of assembly and integration sits in a small number of countries. A tariff or export-control change in this chain does not merely raise cost; it moves where lines are built and therefore where the inspection is bought.
Scenarios



Scenario
2033 (USD m)
CAGR
Driver




Conservative
12,600
4.81%
Automotive and electronics capital spending pauses; vision deflation outruns volume


Base
14,158
6.35%
Machine vision holds 8.70%; EV battery and motor inspection compounds on schedule


Optimistic
15,700
7.74%
Optical systems win conformance acceptance, lifting both placement and specification

- **Overall Risk:** moderate
- **Geopolitical Exposure:** high
- **Substitution Risk:** high

## Regulatory Landscape

6. Regulatory landscape: acceptance is the fence
Metrology is regulated less by what you may measure than by what your measurement may be used to certify. That distinction is the whole of the regulatory story in this market.
The acceptance standards

ISO 10360 — the acceptance and reverification tests for coordinate measuring machines. It defines how a CMM’s performance is demonstrated and re-demonstrated, and it is the technical foundation on which every tactile machine in this report is sold. It has no equivalent maturity for optical systems.
ISO 14253-1 — decision rules for proving conformance or nonconformance with specification, which force the uncertainty of the measurement into the accept-or-reject decision. This is the standard that makes measurement uncertainty consequential rather than academic: a system whose uncertainty you cannot state cannot be used to declare a part good.
VDI/VDE 2634 — the German-led series covering optical 3D measuring systems, spanning imaging systems and surface scanning. It is the closest thing the optical world has to ISO 10360, and its narrower adoption is precisely the gap that confines vision to process control.
ISO 17025 — competence of calibration and testing laboratories, which governs who may certify a traceable result.
IATF 16949 and measurement systems analysis — the automotive quality regime, which requires that a gauge be shown capable before it is used to control a characteristic. In practice MSA, and gauge repeatability and reproducibility in particular, is the gate that decides which measurement technology an automotive plant may deploy on a safety- or fit-critical feature.

Why the fence has not slowed vision down
This is the most important point in the chapter, and it is easy to get backwards.
Machine vision cannot currently be used for the conformance decisions that ISO 14253-1 and automotive MSA govern. On a strict reading, that should cap its growth. It has not, because most of the growth in this market is not in conformance work at all. It is in process control on new lines, in new products and in plants where no legacy measurement programme exists to be matched. A battery module line being specified today imposes no acceptance requirement inherited from a tactile programme; it specifies inline inspection because inline inspection is what the line needs.
So the fence protects the highest-value slice of the market and excludes vision from it, while vision grows around the fence into the larger slice by volume. The scenario that changes is not a technology scenario; it is a standards scenario. Mature optical acceptance standards would open conformance work to vision and justify the optimistic case. Their absence caps the price ceiling without capping the units.
What to watch
Three signals. First, any broadening of VDI/VDE 2634 or a new ISO work item giving optical systems a defensible route to acceptance testing — the single most valuable regulatory development this market could see. Second, revision of automotive measurement systems analysis guidance to admit optical systems for controlled characteristics, which would move the largest single end market. Third, traceability infrastructure: a supplier who can deliver a documented uncertainty budget for an optical measurement of a real surface, not a calibration artefact, has effectively written its own acceptance argument.

- **Complexity:** moderate

## Frequently Asked Questions

**Q: What is the size of the global inline metrology market?**

USD 8,650 million in 2025, rising to USD 14,158 million in 2033 at a compound annual growth rate of 6.35 per cent. Systems placed grow from 96,500 to 131,060 at 3.90 per cent, and realised spend per system rises from USD 89,637 to USD 108,012 at 2.36 per cent.

**Q: When does machine vision overtake the coordinate measuring machine?**

During 2028. Inline CMM and multisensor cells are USD 2,335 million in 2025 growing 4.60 per cent; machine vision dimensional systems are USD 2,119 million growing 8.70 per cent. Those two rates cross in the middle of 2028, and by 2033 machine vision holds 29.2 per cent of value against the CMM share of 23.6 per cent.

**Q: Why is this market volume-led when the rest of its cluster is price-led?**

Because inline metrology is not a replacement market. 56.3 per cent of the USD 5,506 million increment is additional systems rather than price and mix, against 60.1 and 62.1 per cent price and mix in our ultrasonic and non-destructive testing reports. Those markets replace instruments already owned against a code obligation; inline metrology is being installed where nothing was installed before, which is why realised-spend growth here is only 2.36 per cent.

**Q: Does this report include industrial computed tomography?**

No. Industrial computed tomography measures internal geometry non-destructively and sits at the boundary between metrology and radiographic inspection. It is sized in our industrial radiography and digital X-ray inspection report, and adding the two would double-count it. This report and that one are siblings in the same cluster, not increments of each other.
