AI Drug Discovery Shifts Toward Wet-Lab Data

Illumina (ILMN) [1] and Revvity (RVTY) [2] both said on their third-quarter earnings calls, held between October 29 and November 3, 2026, that the bottleneck in AI drug discovery is standardized wet-lab data, and that drugmakers and AI biology companies are now paying tool vendors to generate it.

Why AI drug discovery models are short of lab data

Both companies sit at the start of the drug-research chain. Illumina sells gene sequencers and the reagents that run on them. Revvity sells high-content screening systems, which automatically image cells across thousands of wells and analyze how their shape changes, along with reagents and detection products. Historically, the data these tools produced was scattered across individual research projects, each with its own conditions and formats. Illumina said on its call that models trained on this kind of data are often better at predicting where the data came from than at predicting the biology the experiments actually reveal [1].

Customers therefore need large, standardized and reproducible datasets, and those datasets do not exist at scale. Illumina launched the Billion Cell Atlas early in 2026 to produce "perturbation" data: researchers deliberately change a gene in a cell and then use sequencing to record how the cell responds [1]. Because this data can only be produced batch by batch in a lab, a customer ends up buying sequencing, screening instruments and reagents in turn, and the demand can spread to other suppliers along the same experimental workflow.

Illumina's backlog turns into revenue as Revvity's AI demand broadens

Illumina's management called the lack of functional datasets one of the biggest bottlenecks in applying AI to biology, and said there is big demand from customers asking it to help generate large datasets [1]. In the third quarter it said the Billion Cell Atlas program kept delivering data and was converting more of its backlog to revenue [1]. The disclosures do not give a dollar figure for those orders or that revenue.

Revvity's shift started earlier and comes with more detail. In the second quarter of 2026, it first disclosed a clear increase in orders directly tied to AI, from traditional pharma and biotech customers as well as new customers that had never bought from it before, and high-content screening order velocity outpaced near-term production capacity [3]. By the third quarter, the CEO said AI-related incremental demand had begun to reach parts of the portfolio beyond cellular analysis, and the CFO said backlog grew across almost every instrument group, with Life Sciences organic growth of 9% [2]. The company has added capacity and is building more for next year [2].

One company sells data programs and the other sells instruments and reagents, but both attribute the demand to customers needing more wet-lab data [1][2].

How lab data demand may move to reagents and later preclinical steps

When data has to come out of experiments, orders go to vendors that can produce standardized data reliably and at scale. Installed instruments also keep consuming reagents. Revvity's CEO described the order in which demand travels: first reagents, from high-content screening to high-throughput screening, and then, as drug candidates approach an Investigational New Drug (IND) filing, the application submitted to regulators before human trials begin, imaging and in vivo detection; the company expects that step in 2027 [2]. The CFO added that AI demand is hitting some areas of the industry more than others, which could widen the growth gap between Revvity and the broader market [2].

This view has a clear limit. Revvity's CEO also acknowledged a strong fundamental recovery in the base market that is helping the instruments business [2], so the AI contribution cannot yet be separated out. Three things are worth tracking: whether Revvity's backlog and reagent growth continue, whether Illumina discloses the revenue size of the Billion Cell Atlas, and whether in vivo and imaging demand shows up in 2027 as expected.

Other companies exposed to AI lab-data demand

• Waters (WAT): Sells liquid chromatography-mass spectrometry systems and columns used to separate and measure drug molecules. Its CEO said on the third-quarter call that AI is creating more samples that must be measured, driving more runs and greater column consumption, and that dozens of AI-native biotech companies became first-time customers this year [4]. Waters has not separately quantified the AI contribution.
• 10x Genomics (TXG): Sells single-cell analysis tools and calls Flex Apex the standard assay for large-scale perturbation experiments, the same data-generation step as the Billion Cell Atlas. Its second-quarter total revenue grew only 3%, and instrument revenue fell 47% on a platform transition, so this demand is not yet visible in revenue [5].
• Charles River (CRL): Provides preclinical drug testing, supplies wet-lab data for Eli Lilly's TuneLab AI model, and runs the studies candidates need before an IND filing, the next step Revvity described. The company itself says the number of AI-native clients is still very small [6].

Sources

[1] Drillr · Illumina (ILMN) · 2026-10-29 · Q3 earnings call

[2] Drillr · Revvity (RVTY) · 2026-11-03 · Q3 earnings call

As customers look to build larger biological data sets to train and refine their AI models, they are going to need exponentially more data from the wet lab.

[3] Drillr · Revvity (RVTY) · 2026-08-04 · Q2 earnings call

[4] Drillr · Waters (WAT) · 2026-11-03 · Q3 earnings call

[5] Drillr · 10x Genomics (TXG) · 2026-08-06 · Q2 earnings call

[6] Drillr · Charles River (CRL) · 2026-08-05 · Q2 earnings call

This is only meant to help you spot industry changes and companies that may be overlooked - not stock recommendations.