The Reagent Is the Business: Why Diagnostic Instrument Placements Are a Cost, Not Revenue

The number that should stop a reader

Roche Diagnostics closed 2025 with CHF 13,847 million of sales, up 2 percent at constant exchange rates. Its operating profit fell from CHF 1,946 million to CHF 1,070 million, and the divisional margin went from 13.6 percent to 7.7 percent. In the same year, cost of sales rose 10 percent at constant exchange rates against 2 percent sales growth. Read those two sentences together and the shape of the in vitro diagnostics business becomes visible: revenue was held roughly flat while the cost of earning it rose five times faster.

That is not a procurement accident. It is what happens to a business whose real product is a reagent, sold into an installed base of instruments the vendor has usually given away or placed at a discount. The instrument is a distribution channel. The reagent is the annuity. When the annuity stops growing, the channel turns from an asset into a standing cost, because every placed analyser carries service obligation, consumable logistics and field engineering forever.

The global diagnostic tests market is forecast to grow from USD 118.60 billion in 2025 to USD 197.85 billion in 2033 at 6.60 percent, and that aggregate hides the whole argument. Whether the rate is delivered depends less on how many instruments are placed than on what flows through each one.

The split inside Roche is the whole market in miniature

Roche’s own disclosure separates four businesses, and the dispersion between them is not a rounding difference. Core Lab, the central-lab chemistry and immunoassay franchise, booked CHF 7,614 million at 0 percent constant-currency growth. Molecular Diagnostics did CHF 2,527 million at plus 4 percent. Near Patient Care fell 3 percent to CHF 1,983 million. Pathology grew 14 percent to CHF 1,723 million.

Core Lab is the largest and flattest. It is also the purest expression of the classic model: enormous installed base, deeply embedded workflow, and underlying test volume that in mature markets is now growing more slowly than price per test is falling. When a franchise of that size posts zero, the group has to earn its growth somewhere else.

Pathology at plus 14 percent is where it earned it. That is the digitisation and advanced-staining frontier, where a new instrument generation genuinely changes what a laboratory can bill for rather than simply replacing an ageing analyser on the same bench. The diagnostic tests report treats this as the central contrast: replacement demand grows the installed base but not the annuity; capability demand grows both.

Twenty-six thousand euros per box

The clearest quantification of the placement model comes from Siemens Healthineers, which disclosed roughly EUR 26,250 of revenue per SPOTFIRE point-of-care instrument against an order of magnitude more for a central-lab analyser. That is the entire economics in one ratio. A placed box in a central laboratory is a multi-million-euro platform decision with a decade-long reagent tail. A placed box at the point of care is a rounding error on the capital budget and a much smaller annuity.

Which is why unit growth at the point of care can be real and still not move revenue. Siemens placed 6,400 SPOTFIRE units in the year, up 110 percent, while Roche’s Near Patient Care business declined 3 percent over the same period. Both statements are true and neither contradicts the other: instrument counts and reagent pull-through are two different series, and in a deflationary reimbursement environment the first can rise while the second stalls.

bioMerieux shows the same mechanism from the other direction. Its BIOFIRE non-respiratory panel franchise reached EUR 598 million, up 10 percent, on roughly 1,800 additional instrument placements. Here the placements converted, because each one opened a panel menu that a laboratory had no alternative route to. The distinction is not the instrument. It is whether the assay on it is contestable.

The margin spread says the same thing

Across the peer set, the diagnostics operators with the highest margins are not the largest. Danaher’s diagnostics segment reported USD 9,941 million at 26.7 percent. Thermo Fisher’s diagnostics business, USD 4,676 million, ran at 26.9 percent. QIAGEN posted USD 2,090 million at 29.5 percent with 5 percent constant-currency growth. Abbott’s diagnostics line, at USD 8,937 million, saw the fourth quarter fall 3.6 percent on an organic basis and 0.2 percent excluding COVID testing. bioMerieux grew 6.2 percent to EUR 4,070 million.

QIAGEN is the interesting case because its revenue base is a fraction of Abbott’s and its margin is higher. It sells consumables-led molecular workflows with less capital equipment attached. Abbott sells a great deal of capital equipment, and its growth in the quarter that mattered was negative once the pandemic-era testing tail is stripped out. Scale in this market does not protect margin if the scale was bought through placements.

The same logic explains why Danaher’s diagnostics adjusted profit slipped from USD 2,916 million to USD 2,846 million on slightly lower revenue, and why the group has been pruning platforms rather than adding them. QIAGEN’s withdrawal from NeuMoDx and its Dialunox wind-down are the same decision seen from a smaller balance sheet: a platform that cannot generate enough reagent pull-through to justify its service burden is not an asset at any price.

What buyers are actually paying for

Agilent’s USD 950 million acquisition of Biocare Medical, a business with more than USD 90 million of revenue, was framed by the buyer around non-instrument revenue mix. That is a precise statement of strategy and it is worth reading literally. The premium was paid for the annuity, not the hardware. The same reading applies to every transaction in this market: the multiple attaches to the recurring consumable, and the instrument is the toll the buyer accepts to reach it.

For anyone underwriting growth here, that reframes the diligence question. The question is not how many systems were placed. It is what proportion of the placed base is running at full menu, in a jurisdiction where the assay is reimbursed, on a platform whose consumables are sole-source. Three placements that satisfy those conditions outperform thirty that do not, and the accounting will not distinguish between them until the miss shows up two years later as a margin decline.

Where the growth has to come from

At 6.60 percent for the category, and 0 percent in the largest single franchise, the arithmetic only works if volume and menu expansion do the work that price cannot. That means panels that a laboratory cannot assemble itself, tests that change a treatment decision rather than confirm one, and geography where the installed base is still being built rather than replaced.

The regulatory and payment side of that equation is not a footnote to the commercial story; it is the mechanism that decides which assays qualify for the annuity at all. That is the subject of the companion piece on how price schedules, not analytical validity, set growth, and it applies with equal force to the sequencing side of the market, where the sequencer stopped being the product for exactly the same reason.

The one-line version: in diagnostics, the instrument is a promise and the reagent is the invoice. Companies get paid for keeping the second one growing, and the market does not reward the first on its own.