Installed base
Instruments, automation, expanded laboratories and trained operators represent potential capacity. They become recurring activity only when they meet samples, appropriate assays, staff, funding or reimbursement, and decisions that justify their use.

Effective adoption
The proportion of active instruments, tests per instrument and per site, frequency of use, menu mix, invalid results, uptime, renewals and recurring revenue distinguish an installation from a productive platform.

Analytical scenarios
Disciplined expansion, selective normalisation and stranded capacity describe operational configurations that can be recognised through indicators; they are neither probabilistic forecasts nor estimates of market share.

Decision window
Decentralised diagnostics creates value when the result arrives while treatment, isolation, admission, discharge or another action can still be changed and the distributed workflow remains governed.

Portfolio discipline
Quality, change control, post-market activities, interoperability and service determine which models succeed; connectivity and contractual form alone do not demonstrate adoption or value for money [5–10].

The post-pandemic period does not correspond to the emergency peak and does not offer a single curve that applies to all diagnostic segments. The capacity built across instruments, laboratories, expertise and supply chains remains an option until it is converted into routine use. The decisive transition occurs when the installed base supports relevant menus, volumes, quality, updates and decisions that justify its continued operation.

This Insight proposes a way of interpreting developments, not a forecast. It starts with the productivity of installations, distinguishes three operational scenarios, examines when decentralised diagnostics genuinely changes the pathway, and observes how regulation, post-market activities, service-based models and interoperability filter growth. The aim is to separate potential capacity, effective adoption and unabsorbed capacity without turning an aggregate snapshot into a universal CAGR.

Between 2020 and 2022, purchases of molecular diagnostics responded to exceptional conditions: urgent demand, extraordinary funding, temporary procedures and a tolerance for redundancy that is difficult to sustain in normal times. Those three years therefore remain a misleading benchmark, capable of encouraging both excessive expectations and overly pessimistic interpretations.

The decline in COVID-19 volumes left a tangible asset base: instruments, automation, expanded laboratories, trained operators and qualified suppliers. This remains potential capacity until it meets samples, appropriate assays, staff, funding or reimbursement, and decisions that justify its use. It is at that point that the infrastructure starts generating recurring economic activity again.

The European IVD Market Statistics Report 2025 aggregates 2023 and 2022 sales in the EU27, EFTA countries and the United Kingdom. The estimated total value for 2023 is EUR 13.3 billion, down 20.2% from 2022; the report attributes the contraction mainly to the decline in COVID-19 testing. After growth of 29.1% in 2020 and 41.2% in 2021, the market had already recorded declines of 19.3% in 2022 and 20.2% in 2023, while remaining above the EUR 11.1 billion recorded in 2019 [1].

These data nevertheless require caution. The estimates come from an industry association, extend only to 2023 and cover the entire IVD market, not just molecular diagnostics. They document the decline in the emergency component; they do not describe 2026 in real time and do not support a single CAGR across different technologies, geographies or use cases.

During the emergency, the main need was to increase throughput. In routine activity, by contrast, each installation must demonstrate how much value it produces. An instrument can be inactive and offer no resilience; it can operate at full capacity on a single assay and be efficient, but vulnerable to concentration; it can have a broad menu and remain underused because the site cannot support the necessary workflows.

For this reason, the industrial metric no longer stops at units delivered. It considers the proportion of active instruments within the installed base, tests per instrument and per site, frequency of use, menu mix, invalid results and repeat tests, uptime, renewals and recurring revenue. The time needed to transfer an assay across the network must also be measured, provided that change control, analytical verification, training and monitoring support the transition.

The menu must not simply be extensive. It must address questions in which the molecular result changes a decision. In 2025, WHO consolidated recommendations for the diagnosis of tuberculosis infection, disease and resistance into a single framework; the handbook accompanies the recommendations with algorithms and implementation guidance [2,3]. For priority bacterial pathogens, a WHO landscape analysis from the same year maps available tools and those in development, and identifies gaps in detection, identification and antimicrobial resistance [4]. These are examples of structural demand because they connect the test to a pathway, not because they reuse installed capacity.

The speed of updates cannot be separated from quality. MIQE 2.0 requires transparency on the sample, assay design and optimisation, controls, analytical performance, analysis and reporting [5]. It is a scientific guideline, not a regulatory authorisation; it does, however, make visible the information that must be documented when a method is adapted, transferred or compared across sites.

The following are analytical scenarios, not probabilistic forecasts or estimates of market share. They serve to identify different operational configurations through observable indicators.

In the disciplined expansion scenario, validated menus for frequent or high-impact needs, sustainable procurement or reimbursement, available staff, integrated workflows and a reliable supply chain align. Under these conditions, the installed base can shorten time-to-deployment, and consumables or recurring services can grow without depending on new hardware sales.

In the selective normalisation scenario, some areas absorb the inherited capacity, while other installations remain little used or are consolidated. The aggregate result may appear stable while concealing substantial differences between segments, regions, sites and platforms.

In the most fragile scenario, poorly differentiated menus, insufficient volumes, price pressure and maintenance burdens turn the installed base into stranded capacity: instruments are present, but lack viable service economics. Possible responses include menu rationalisation, site consolidation, OEM agreements or withdrawal from unsustainable geographies; none of these options creates value automatically.

The three scenarios can be distinguished by observing site-level utilisation, non-emergency revenue, active installations, renewals, margins on consumables, update or certification times, churn and evidence of impact on the pathway. A market average does not replace these metrics.

The pandemic made it clear that some testing can move beyond the central laboratory, without making distributed testing a rule that applies to every assay. The distinguishing factor is the decision window: the result must arrive while treatment, isolation, admission, discharge or another action can still be changed. If the decision can wait for the central batch, the peripheral node risks adding costs and governance without a proportionate benefit.

The comparison must therefore be made between complete pathways, not between a compact instrument and a central analyser. Collection, transport, operator, quality, maintenance, controls, connectivity, invalid results, referral and time to decision are all part of the same assessment. The performance of a single portfolio or a single commercial territory is not enough to demonstrate a universal advance or retreat of decentralised diagnostics.

A hybrid network often offers the design that is easiest to govern: it brings tests with high time-sensitive value closer to the decision, while keeping volume, complexity, confirmation and supervision at the centre. There is, however, no boundary that applies everywhere; it depends on demand, expertise, risk, intended purpose and escalation capability. Growth becomes defensible when that boundary is shaped by the pathway and measured against predefined outcomes.

The idea that the post-COVID period has made regulation permanently fast is misleading. In Europe, the IVDR has applied since 26 May 2022 and has introduced a more demanding framework for classification, performance evidence, quality systems, surveillance and notified body involvement, with staggered transitional periods for certain devices [6].

On 16 December 2025, the European Commission proposed a targeted revision to simplify and reduce certain burdens under the MDR and IVDR. As of August 2026, it remains a proposal subject to the ordinary legislative procedure: it does not replace current law and cannot be treated as deregulation that has already been secured [7].

On 28 May 2026, the first four EUDAMED modules became mandatory: actor registration, UDI/device registration, notified bodies and certificates, and market surveillance [8]. This step increases traceability and data discipline, but does not mean that all EUDAMED modules are operational or that market surveillance and the manufacturer's post-market surveillance are the same thing.

Post-market activities remain a continuous process within the quality system. MDCG 2025-10 guidance clarifies that the manufacturer must proactively collect and analyse data on quality, performance and safety, updating technical documentation, performance evaluation and risk management throughout the device's lifetime [9]. The industrial implication is that each new assay also competes for documentation, monitoring and updating resources.

The same discipline applies to OEM partnerships. The contract must make the intended purpose, the role of economic operators, dossiers, change control, surveillance and management of modifications explicit. It can allocate activities and contractual responsibilities; it does not rewrite the legal role assigned by the applicable framework.

A service-based model can reduce the initial outlay and align the supplier with availability, consumption and support. The same arrangement can, however, shift costs onto minimum commitments, dependence on consumables and terms that are difficult to compare. Diagnostics-as-a-Service therefore describes a commercial model; its value for money still needs to be demonstrated.

To assess it, the unit of payment, minimum volumes, demand risk, data ownership and access, service levels, response times, management of invalid results and exit conditions must be made explicit. Pay-per-test is viable if volumes support the network and the supplier can absorb variability. With intermittent demand, an availability fee may better reflect reserve capacity, provided that it is tied to uptime, recovery times and verifiable criteria.

OEM arrangements can, in turn, accelerate access by combining hardware, reagents, software and distribution from different parties, but they bring dependencies on critical suppliers and interconnected changes. Their advantage is measured by the ability to govern quality, continuity and responsibility better than the alternative, not by simply reducing the assets owned.

Connecting the device is only the first step. Making the result usable within the healthcare ecosystem requires identity, terminologies, assay version, an audit trail, offline operation, cybersecurity, legal bases, access rights, data minimisation and integration with LIS and EHR. A proprietary dashboard can give the supplier greater visibility without reducing the workload of the laboratory or clinician.

The European Health Data Space entered into force on 26 March 2025, but application is gradual. From March 2029, the first priority categories for primary use, such as patient summaries and electronic prescriptions or dispensations, are expected to become operational; from March 2031, the exchange of medical images, laboratory results and discharge reports is envisaged [10]. The horizon is strategic, not immediate: designing open architectures today is prudent; claiming that European interoperability has already been achieved would not be.

Data creates value if it reduces an error, shortens the time to a decision, supports surveillance or makes the service measurable. Without a defined purpose, accumulating data increases costs, the risk surface and responsibilities.

Within this framework, 2030 is an observation horizon, not the deadline for a forecast. The signals to follow are concrete: how much non-COVID menus absorb the installed base; how many sites remain active after the first year; whether utilisation grows without increasing invalid results and staff workload; whether decentralised diagnostics changes a pathway; and whether updates, uptime and support remain sustainable.

Beyond volume, the composition of growth also matters. An increase in recurring revenue can indicate adoption, or contractual or technological dependence. A broader portfolio can reflect innovation and, at the same time, become burdensome to maintain. Interpreting these indicators also requires consideration of quality, customer concentration, supplier dependence, geographies, reimbursement and the ability to replace a critical component.

For management and industrial investors, the decisive test is to understand how much value still depends on a new emergency shock. A robust model retains preparedness as an option and bases its economics on routine activity. It can remain sustainable without predicting the next pandemic.

The pandemic brought forward investment and familiarity with molecular testing, without suspending the ordinary economics of diagnostics. Today, the issue is not simply to add more instruments, but to make existing ones productive through appropriate menus, documented quality, governed updates, usable data and a reliable service.

Growth can continue without replicating the peak, but not along a single trajectory. The strongest interpretation observes which platforms genuinely turn the inherited infrastructure into recurring use and which remain unabsorbed capacity. Distinguishing them does not require guessing a CAGR: it requires metrics that separate adoption, reserve capacity and unabsorbed capacity.