Thirteen Point Four Three Per Cent a Reaction

The only observable price in the market

10x Genomics published the cleanest measurement in genomics and probably does not regard it as one.

Consumable reactions rose from 357,100 to 424,000 — up 18.73 percent. Consumables revenue rose from USD 493.4 million to USD 507.2 million — up 2.79 percent. Divide the second by the first and realised revenue per reaction fell from USD 1,381.78 to USD 1,196.19, a decline of 13.43 percent.

That is the only place in this market where a price is observable rather than inferred. Every other participant discloses revenue, or units, or occasionally both in a form that cannot be divided — one price assumed across three different products, or a test count that includes products sold at different rates. 10x discloses a unit count and the revenue those units produced, in the same segment, for two consecutive years. It is a gift to anyone trying to size the market, and it is unflattering.

The NGS library preparation and sample preparation market report is built on that measurement, because library preparation is priced per sample — the only layer of the sequencing stack a laboratory must buy again before it can use anything it already owns.

What a 13.43 percent decline does to a forecast

The category is sized at USD 3,340.0 million in 2025 rising to USD 5,471.1 million in 2033 at 6.36 percent. The rate is not a judgement about genomics. It is a bridge between two variables: volume growing at roughly 12 percent and price declining at roughly 5 percent, which multiply to 1.064, or 6.4 percent.

The entire model therefore rests on one assumption, and the report states it plainly: that the rate of price decline narrows from the 13.43 percent measured at 10x toward 5 percent. Everything else is arithmetic.

Which makes the downside testable rather than rhetorical. Hold volume growth at 12 percent and hold price decline at the rate actually measured in 2025. The product is 1.12 multiplied by 0.8657, which is 0.9696. The market contracts 3.04 percent a year. That calculation sits in the risk chapter, not because it will happen but because a forecast that cannot reach its own downside is not a forecast, it is a hope with a table attached.

Why the decline narrows, and where it does not

Price falls hardest where the buyer can substitute freely and least where the reagent is written into something the buyer cannot change. That is the whole mechanism, and it produces two businesses sold through one catalogue.

Commodity chemistry — enzymes, adapters, index primers, beads — is priced toward marginal cost. The largest supplier of library preparation enzymes in the world is New England Biolabs, which is private, publishes nothing, and does not appear in any competitive table as a result. Content — capture panels, validated workflows, automation-locked reagent sets — is priced on the answer it produces rather than on the molecule.

Twist Bioscience demonstrates the split better than any argument can. NGS revenue grew 23.06 percent to USD 208.1 million while gross margin rose from 42.6 percent to 50.7 percent — up 810 basis points — in the same year that commodity suppliers were cutting price. A capture panel has to be redesigned as the knowledge base changes, which means it is a subscription to content rather than a purchase of chemistry.

QIAGEN’s fourth quarter contains the whole split in one sentence

The sharpest single disclosure in the file is not a number. It is a clause. QIAGEN reported that its Digital Insights business grew at a double-digit constant-currency rate while, in the same product group and the same quarter, consumables sales of universal NGS panels declined, reflecting funding constraints.

Software grew and reagents shrank, inside one segment, at one company, in one three-month window. If a reader wanted a single observation to test the claim that content holds price and commodity does not, that is it. It also explains something otherwise puzzling: why a company with a strong portfolio can report weak consumables growth without anything being wrong with its products.

The installed base decides 2030, and it is shrinking

Consumables revenue is a function of instruments already placed, which makes the instrument line the leading indicator for this market rather than a coincident one. On that measure the news is not good.

Illumina’s sequencing instrument revenue was USD 465 million against USD 484 million the prior year, a 4 percent decline, and 1 percent even excluding China. 10x’s instrument revenue fell from USD 92.7 million to USD 56.8 million, a decline of 38.76 percent.

The mix shift that follows is visible in Illumina’s own disclosure: consumables are now 74 percent of revenue against 68 percent in 2023, while instruments fell from 16 percent to 11 percent. A company earning a rising share of its revenue from consumables while placing fewer instruments is harvesting an installed base it is renewing more slowly. The library preparation market in 2030 is being decided now, by boxes placed this year, and what has been decided so far is a smaller cohort than the one that preceded it.

The base, and the internal check that it is coherent

Twelve attributed pools reconcile to USD 2,790.9 million of named revenue, which is 83.56 percent of the base, plus a stated residual of USD 549.1 million — 16.44 percent — for the unlisted and regional supply: MGI, BGI and the Chinese reagent makers, and the private Western enzyme and bead houses.

The check that the perimeter is coherent is a ratio rather than a list. USD 3,340.0 million against an NGS market of USD 15.80 billion is 21.14 percent, which is roughly the share of sequencing spend that should be per-sample consumables. A prep market sized at half the sequencing market would be a category error, and one sized at 5 percent would be missing the point that every sample needs a library.

Two rows carry most of the judgement and both are labelled. 10x at 85 percent of consumables is USD 431.1 million, 12.91 percent of the base, and the report states what removing 10x entirely does to the level and the rate — a base of USD 2,908.9 million and a category rate of 6.27 percent, a nine basis-point move. New England Biolabs at USD 400.0 million is the least comfortable number in the report: private, undisclosed, and probably the largest single library-prep enzyme supplier in the world. It sits in the residual rather than being presented as reported revenue.

What would falsify it

Three observables. Price decline has to narrow, which means the mix has to shift toward content faster than commodity volume grows — visible in gross margin rather than revenue. Instrument placements have to stabilise, because a shrinking installed base cannot support a growing consumables line beyond the life of the boxes already sold. And research funding has to stop constraining award counts, since a new project is where a prep workflow gets specified, and the sequencing market report shows what happens to the layer above when it does not.

None of those is a statement about the technology, which is improving faster than the market is growing. That gap is the story: a category can become technically better every year and commercially cheaper every year, and only one of those two facts shows up in the forecast.