Press Release
The Twelve Largest Diagnostics Companies Grew 1.2 Per Cent and the Forecast Is 5.82 — the Difference Is a Price Someone Wrote Down, Not Demand That Went Away
The twelve largest in vitro diagnostics companies in the world sold USD 61.5 billion of diagnostic product in 2025 and grew 1.2 per cent between them. Pheonix Research sizes the market they sit in at USD 104.9 billion in 2025, reaching USD 164.9 billion by 2033 — a 5.82 per cent compound annual rate, 1.57x on the base year and USD 60.0 billion of absolute growth.
The distance between 1.2 per cent and 5.82 per cent is the entire argument of the report, and it is not a demand story. Test volume never stopped. What stopped was the price.
What happened in 2025 was a price, not a market
Roche’s diagnostic sales in China fell 24 per cent, taking Asia-Pacific down 12 per cent while every other region grew mid-single to low-double-digit — and the same division grew 7 per cent at constant exchange rates excluding China. Siemens named centralised volume-based public procurement as the specific cause of a low-double-digit decline in its China diagnostics revenue. bioMerieux grew microbiology 4 per cent including China and 6.3 per cent excluding it. Werfen grew 6 per cent excluding Asia. Sysmex cut its fiscal 2025 guidance from a JPY 560 billion plan to JPY 535 billion, citing Chinese government procurement policy and US tariffs. Hologic reported lower cytology volume in China.
That is six independent companies describing the same event in the same year, in their own words. A price administration is not a demand event and it does not compound. It is a one-time reset of the price level, which makes a base year that contains it artificially low. The 5.82 per cent is not a bet that China recovers; it is what happens to a five-year growth rate when a deflating base laps.
The other half of the argument is that volume was intact throughout. Abbott’s Core Laboratory business grew 2.1 per cent organically to USD 5,360 million in the same year its diagnostics division fell 4.5 per cent, because the decline was almost entirely the COVID testing tail: fourth-quarter testing sales of USD 89 million against USD 176 million a year earlier, which cost 4.9 percentage points of full-year organic growth. Strip COVID out and Abbott’s diagnostics business grew 0.4 per cent. Roche’s clinical chemistry grew 7 per cent while immunoassay fell 3 per cent. Hologic’s molecular business grew on BV/CV adoption and a severe flu season while its SARS-CoV-2 assays declined.
The base was re-based, and the direction is the check on the arithmetic
This edition removes USD 3,108 million from the 2025 base. Urinalysis is carried inside the clinical chemistry and urinalysis report in this cluster, which means it was already inside the USD 43.0 billion immunoassay and clinical chemistry row; counting reagent strips and sediment a second time in the haematology bucket counted that block twice — 2.96 per cent of the base as it now stands.
The correction removes the block from both endpoints, so the compound rate rises eight basis points to 5.82 per cent rather than falling. That direction is the test: the double-counted block sat inside the slowest-growing row in the split, so taking it out of both ends lifts the weighted growth rate. The named pool of USD 61,525 million is now 58.66 per cent of a smaller market rather than 56.97 per cent of a larger one, and the stated residual falls by exactly the double-counted block, from USD 46,475 million to USD 43,367 million, or 41.34 per cent. No revenue was added to or taken from any named vendor. The residual was the balancing item in the construction, and the balancing item had been carrying the error.
The twelve, and the 41 per cent that files no accounts
| Participant | 2025 IVD-attributable, USD m | Share of base | Reported growth |
|---|---|---|---|
| Roche Diagnostics | 15,924 (CHF 13,847m) | 15.2% | +2% CER, −3% CHF |
| Danaher Diagnostics | 9,941 | 9.5% | +1.6% |
| Abbott Diagnostics | 8,937 | 8.5% | −4.5% organic |
| Siemens Healthineers Diagnostics | 4,912 (EUR 4,347m) | 4.7% | +0.4% comparable |
| Thermo Fisher Specialty Diagnostics | 4,676 | 4.5% | +3% organic |
| Top five | 44,390 | 42.3% | — |
| Top twelve | 61,525 | 58.7% | +1.2% weighted |
The largest participant is among the slowest and the sixth-largest is the fastest, and that is not a comment on competence. Roche’s 15.2 per cent is built on a core laboratory franchise that was flat for the year; bioMerieux’s 3.5 per cent is built on syndromic cartridges growing at double digits. One is a mature annuity defended on price, the other a growing annuity defended on menu.
The other 41.34 per cent is a stated residual, not an implied one: thousands of reagent manufacturers, of which the Chinese domestic industry is the largest single block, plus Japanese, Korean, Indian, Brazilian and Turkish suppliers serving their own health systems at prices the multinationals do not reach. That tail is where volume growth is fastest and where price deflation is most severe, which is why it is modelled to grow 3.1 per cent in 2025 against 1.2 per cent for the named pool, and why the two are modelled separately rather than averaged.
Where the growth is: 4.5 per cent against 8.5 per cent
| Test discipline | 2025, USD m | Share | Growth 2026-2033 | 2033, USD m | 2033 share |
|---|---|---|---|---|---|
| Immunoassay and clinical chemistry | 43,000 | 40.99% | 4.5% | 61,150 | 37.09% |
| Molecular diagnostics, including syndromic panels and sequencing | 22,000 | 20.97% | 8.5% | 42,253 | 25.63% |
| Point-of-care and rapid diagnostics | 17,500 | 16.68% | 6.0% | 27,892 | 16.92% |
| Hematology, hemostasis and routine benchtop testing | 9,392 | 8.95% | 4.3% | 13,171 | 7.99% |
| Microbiology and blood banking | 8,000 | 7.63% | 5.0% | 11,820 | 7.17% |
| Histopathology and digital pathology, IVD-attributable | 5,000 | 4.77% | 7.0% | 8,591 | 5.21% |
| Total | 104,892 | 100% | 5.73% in 2026 | 164,877 | 100% |
The build produces 5.73 per cent in the first forecast year against a 5.82 per cent eight-year compound rate, and the nine basis points are a mix effect rather than an error. Molecular is 20.97 per cent of the base today and 25.63 per cent of it in 2033, so a market growing 5.73 per cent in 2026 arrives at 5.82 per cent compounded: the fast disciplines compound up in weight.
The four-point spread between immunoassay at 4.5 per cent and molecular at 8.5 per cent is the shape of the whole forecast. Immunoassay and clinical chemistry is 41.0 per cent of the market growing at roughly the rate of health system budgets, and it carries the China deflation because that is where the volume-based procurement has bitten hardest — first tumour markers and thyroid function across 28 provinces in 2024, then cardiac markers, then routine blood, stool and urine reagents. Molecular is smaller, faster and less exposed to a reagent tender, because a syndromic panel is bought by a clinician who needs an answer today rather than by a procurement office comparing unit prices.
The price is a document, and it has two authors
This is the only healthcare market whose price is published, and both of the world’s price-setting mechanisms for it moved down inside the same window.
The United States. The preliminary CY 2027 Clinical Laboratory Fee Schedule applies the market-based PAMA methodology for the first time since 2018: 1,171 codes below the 2026 schedule against 186 above and 169 unchanged, with chemistry down 16 per cent across 387 codes, under a 15 per cent annual cap for 2027 through 2029. A cap defers a reduction rather than forgiving one — it converts a single adjustment into a three-year compression, which is worse for anyone planning a reagent contract.
China. The Jiangxi biochemical reagent alliance published its outcome in January 2025: 154 manufacturers and 2,314 specifications awarded at an average price reduction of 70 per cent, with the lowest winning bid at RMB 0.03 per unit and both Roche Diagnostics and Mindray winning lines at RMB 0.07. The Anhui-led 28-province alliance covering tumour markers and thyroid function took reagents down 52.6 per cent on average. In 2025 routine blood, stool and urine reagents entered scope for the first time under Gansu’s procurement, and Jiangsu procured non-invasive prenatal testing as a service — the first time a testing service rather than a reagent was bought this way. National and provincial alliance items across the country were expected to exceed 700 in 2025.
A 70 per cent cut in the reagent price does not cut the manufacturer’s cost by 70 per cent, so margin on that line compresses almost to zero while the volume the award confers is real. The winner is the manufacturer with the lowest cost base, which is precisely why domestic Chinese manufacturers gain share in a volume-based procurement and the multinationals do not: the multinational carries a global cost structure and a global price list, and the domestic manufacturer carries neither. Our procurement interviews described the mechanism identically in every market — the tender is not where the price is discovered, it is where the price is enforced.
Margin is set by menu and density, not by assay performance
QIAGEN runs a 29.5 per cent adjusted operating margin on sample technology and a narrow, high-attach menu. Thermo Fisher’s Specialty Diagnostics converts USD 4,676 million into USD 1,256 million of segment income, 26.9 per cent. QuidelOrtho holds a 22.0 per cent adjusted EBITDA margin on a mature, 84 per cent recurring franchise. bioMerieux earns 17.9 per cent CEBIT on clinical microbiology and syndromic panels. Siemens Healthineers Diagnostics earns 7.7 per cent adjusted EBIT on EUR 4,347 million — up from 5.3 per cent in fiscal 2024, a 240 basis point improvement delivered mainly by cost reduction rather than price. Those are five different measures of margin and they should be read as directional, not ranked. But the spread is not a manufacturing-cost story: a platform with a shallow menu carries the same service cost as a platform with a deep one and earns less per placement.
The annuity is the business and the instrument is the toll booth. QuidelOrtho reports that 84 per cent of its revenue is recurring; Abbott’s Core Laboratory line is four times its Molecular line and six times its Point of Care line; Siemens states the model in its own annual report as long-term contracts beginning with an instrument placement and continuing with reagent sales. Danaher’s Cepheid base is above 60,000 instruments, tripled since the beginning of the pandemic, carrying roughly USD 1 billion of sustainable incremental annual revenue on the company’s own framing. Werfen, pure IVD since divesting its medical devices business in July 2024, reports EUR 2,131 million of sales at a 22 per cent EBITDA margin — the cleanest read available on what these economics look like without a devices business attached.
The menu still gets better: a PET scan became a reagent
The most consequential new category of the decade arrived in 2025 and it is not a cancer test. The US Food and Drug Administration cleared the first in vitro diagnostic device that tests blood to aid in the diagnosis of Alzheimer’s disease — the Lumipulse G pTau217 and ss-Amyloid 1-42 Plasma Ratio, for adults aged 55 and over presenting with signs and symptoms of cognitive decline. In the supporting study, 91.7 per cent of individuals with a positive result had amyloid plaques confirmed by PET or cerebrospinal fluid, 97.3 per cent of those with a negative result were confirmed negative, and fewer than 20 per cent of 499 patients tested received an indeterminate result. The agency was explicit that the test is not a screen and not a stand-alone diagnosis.
Read the implication rather than the assay. Until this clearance, confirming amyloid pathology required a PET scan or a lumbar puncture. A plasma ratio on an existing immunoassay platform turns the highest-volume unmet diagnostic question in neurology into a routine blood draw, on installed instruments, at a reagent price — no new platform, no new laboratory, no new sales force. It is also why the immunoassay and clinical chemistry row keeps growing at 4.5 per cent rather than at the rate of GDP: the menu on the installed base is still getting better.
Attribution coverage is an inverse measure of tendereability
The three discipline reports built underneath this hub attribute 66.6 per cent (immunoassay), 48.9 per cent (clinical chemistry and urinalysis) and 87.28 per cent (haematology and coagulation) of their own value to named vendors — and that ordering is the finding rather than a research artefact. A six-part differential cannot be assembled from third-party reagents, so haematology has almost no unlisted middle and is the easiest field in diagnostics to price administratively. A reagent of enzymes, substrates, buffers, stabiliser and a calibrator can be manufactured almost anywhere, which is why clinical chemistry carries a 51.1 per cent unlisted tail and why a 70 per cent tender cut was possible at all. Concentration and price are the same variable read from two ends.
What would change the number
- If molecular diagnostics grows at 6.5 per cent rather than 8.5: the 2033 market falls to USD 159.0 billion and the compound rate to 5.34 per cent. This is the largest single sensitivity in the model, because molecular is the only discipline carrying a growth rate more than a point above the market.
- If immunoassay and clinical chemistry grows at 2.5 per cent rather than 4.5: the 2033 market falls to USD 156.1 billion and the compound rate to 5.10 per cent. That is what a second wave of volume-based procurement across immunoassay looks like, and it is the most likely way for this forecast to be wrong.
- If the Chinese adjustment reverses rather than laps: the base year is understated and the compound rate rises, because 2033 is unaffected while 2025 moves down.
- If installed-base placements stall: a reagent annuity cannot outgrow its instrument base, and every point of growth above volume requires either more placements or a richer menu on the existing ones.
Market risk is assessed as moderate, and it is administrative rather than clinical. The assays work, the cartridge installed base tripled in five years, and demand is a function of population, age structure and clinical guidelines — none of which is contracting. What can remove revenue is a procurement award, a tariff, a coverage policy or a price administration, and every one of those is decided by an institution rather than by a competitor.
Scope and method
The perimeter is stated rather than assumed. Inside: reagents, kits, controls, calibrators and consumables sold for human clinical diagnostic testing; instruments, analysers and software placed in clinical laboratories; and the disciplines of immunoassay and clinical chemistry, molecular diagnostics, point-of-care and rapid testing, hematology, hemostasis, urinalysis, microbiology and blood culture, blood banking, and IVD-attributable histopathology and digital pathology. Outside: laboratory services revenue, research-use-only reagents, life science instruments and biopharma services, veterinary, food and environmental testing, industrial and pharma quality control, and medical imaging. Industrial and pharma quality control alone removes roughly 20 per cent of bioMerieux’s reported sales. A figure of USD 104.9 billion and a figure of USD 160 billion can both describe diagnostics honestly if one is the reagent and instrument market and the other the testing market; this report builds to the former, and the laboratory that turns a reagent into a billable result is covered separately as the demand side, not added to it.
The base is built bottom-up from disclosed revenue and contains a stated residual; no published aggregate was used to set the level, and no third-party market research estimate has been used at any point in the construction. Revenue is attributed to the calendar year in which it was earned. Non-USD revenue is translated at 2025 average rates — EUR 1.00 = USD 1.13, CHF 1.00 = USD 1.15, JPY 148 = USD 1.00 — which affects the level and not the growth rate. Every growth rate is taken from the participant’s own reported constant-currency or organic figure. The method is 70 per cent secondary and 30 per cent primary. The secondary layer is primary corporate and regulatory disclosure read in full: annual reports, results releases, capital markets presentations and regulator clearances. The primary layer is Pheonix’s own field programme in the category — 46 depth interviews between the second quarter of 2025 and the third quarter of 2026 across eleven markets (United States, Germany, France, United Kingdom, Italy, Spain, Japan, China, India, Brazil and Saudi Arabia) with laboratory directors and clinical biochemists (19), IVD commercial and market access leaders (13), procurement and group purchasing executives (8) and reference laboratory and distributor principals (6); one structured survey of 214 laboratory procurement and operations decision-makers in the second quarter of 2026; and direct observation at AACC/ADLM 2026, Medica 2025, ESCMID Global 2026 and the J.P. Morgan Healthcare Conference 2026. The primary layer is used to explain mechanism — how a tender is run, why a menu is rationalised, what a laboratory does when a reagent price resets — and never to substitute an unevidenced number for a disclosed one. All arithmetic was computed, not estimated: the 5.82 per cent compound rate is implied by the 2025 and 2033 endpoints.
About this report
Global In Vitro Diagnostics (IVD) Market Report, Size, Share and Forecast 2026-2033 is the hub of a nine-report Diagnostics cluster. The discipline reports built underneath it are additive and read against it row by row: Global Immunoassay Market (USD 26.4 billion in 2025, 4.90 per cent, 25.17 per cent of IVD value), Global Clinical Chemistry and Urinalysis Market (USD 16.8 billion, 3.71 per cent, 16.02 per cent), Global Haematology and Coagulation Diagnostics Market (USD 7.62 billion, 4.25 per cent, 7.26 per cent) and Global Microbiology Diagnostics Market (USD 8,830 million, 5.14 per cent, 8.42 per cent). Two sit across the cluster rather than beside it: Global IVD Reagents and Consumables measures the USD 78.4 billion consumable layer — 74.74 per cent of this base — and Global Infectious Disease Diagnostics is the application layer at USD 30,150 million, 28.74 per cent of the base. Neither the reagent layer nor the infectious disease report is additive to the discipline totals, and the report states why for each.
About Pheonix Research
Pheonix Research builds market intelligence bottom-up from primary corporate and regulatory disclosure, with a stated residual rather than an implied one and a primary field programme used to explain mechanism rather than to substitute for a number. Every figure in this release is derived from a disclosed source and every calculation is computed rather than estimated. The full report, its data tables and its four companion analyses are available at pheonixresearch.com.
