Divergent Compute.AI Economic Think Tank

Markets / Industries / Life-Science Tools

Industry view · Life-Science Tools

The pick-and-shovel trade learns to think

Life-science tools — the instruments, reagents and sequencers that sit upstream of every drug and diagnostic — is a roughly $154 billion market in 2025 that grew up selling hardware. AI is quietly inverting the value: the molecule of the future is software, the instrument is a data-acquisition front end, and the margin is migrating toward whoever owns the interpretation layer.

$154B
Life-science tools market, 2025
Mordor Intelligence
>75%
of labs plan to deploy AI within 2 years
Industry survey via Mordor
~$2-6B
AI-in-drug-discovery market, 2025
Grand View / Roots Analysis
~31%
AI drug-discovery CAGR to 2030
BCC Research

01 · The thesis

The instrument is now a data pipe; the value moved to the model

For two decades the tools industry sold razors and blades: a sequencer or mass spectrometer up front, then a perpetual annuity of consumables. AI does not threaten that model so much as relocate its profit pool. When an interpretation algorithm doubles a sequencer's diagnostic yield — Illumina says its PromoterAI, PrimateAI-3D and SpliceAI stack roughly doubles diagnostic yield versus protein-truncating variants alone — the differentiated, defensible asset is the software, not the optics. The hardware risks commoditisation; the model becomes the moat.

The clearest tell is who the incumbents are partnering with. Thermo Fisher signed a strategic collaboration with OpenAI and reports AI already delivering a 30% increase in effective capacity at major bioproduction sites. Danaher elevated a former Tempus and Google AI leader to Chief Technology and AI Officer in 2025. Illumina is embedding NVIDIA foundation models into DRAGEN. The competitive question is no longer 'whose box is faster' but 'whose box turns raw signal into an answer a clinician or chemist will pay for.'

1Target ID

Foundation models read biology

Protein, DNA and multimodal foundation models now propose targets and structures that wet-lab work used to take years to surface.

AlphaFold 3 / Isomorphic Labs, NVIDIA BioNeMo, Arc Institute
2Discovery

In-silico design compresses the funnel

Physics-plus-AI platforms and generative chemistry shrink the hit-to-lead cycle, shifting spend from benches to GPUs.

Schrödinger, Recursion, Isomorphic
3Sequencing

The sequencer becomes an AI front end

Secondary analysis and variant interpretation, not raw read-out, are where yield and reimbursement are won.

Illumina DRAGEN + NVIDIA, Tempus
4Lab automation

Self-driving labs close the loop

Robotic sample handling plus AI scheduling turn instruments into autonomous, agentic experiment engines.

Thermo Fisher Vulcan, Agilent, 10x
5Interpretation

Data and software capture the margin

Recurring data-and-applications revenue is the fastest-growing, highest-multiple layer of the stack.

Tempus Data, Benchling, Schrödinger software
Pace of AI disruption by stage — Divergent Compute assessment

02 · Public players & exposure

Who routes through, who gets routed around

We plot the listed players on two editorial axes — how exposed each is to AI disruption, against how ready its data, brand and position are to be the answer. The figures in the table are sourced; the placement is our read.

Positioning — editorial assessment, not a sourced metric. Bubble = approximate relative scale.
CompanyStanceThe sourced fact
Thermo Fisher ScientificTMOScaled incumbentReports AI driving a 30% increase in effective capacity at major Pharma services and bioproduction sites, and a strategic OpenAI collaboration embedding its APIs across products and operations.
DanaherDHRAI-native pivotElevated Martin Stumpe (ex-Tempus, ex-Google Cancer Pathology) to Chief Technology and AI Officer in 2025, reporting to the CEO, and added Microsoft's AI Copilot VP to its board.
IlluminaILMNMoat in softwareSays its AI interpretation stack (PromoterAI, PrimateAI-3D, SpliceAI) roughly doubles diagnostic yield; expanding DRAGEN with NVIDIA Biology foundation models.
Tempus AITEMData-first challengerGenerated $1.27B revenue in 2025; its Data and Applications segment alone reached $316.4M, about 25% of total, atop a multimodal oncology dataset.
SchrödingerSDGRPhysics + AIReported $256M total 2025 revenue with $200M from software; drug-discovery revenue surged 295% in Q3 to $13.5M on collaboration milestones.
AgilentAAutomation playPosted $6.95B FY2025 revenue and is targeting the ~$2B automated-lab market by 2030 with AI-driven LC-MS and GC workflows plus an OmixAI proteomics pact.
RecursionRXRXBurn-rate betPost-Exscientia merger ($~688M), full-year pipeline reworked; Q1 2025 revenue was just $15M against a large clinical portfolio, underscoring monetisation risk.
10x GenomicsTXGVolume pressureFY2025 revenue of $642.8M ($598.7M ex-litigation), a ~2% decline, even as AI-driven demand for single-cell and spatial data rebuilds the pipeline.
BrukerBRKRMultiomics scalerTargeting ~$3.5B 2025 revenue; building a spatial-biology and functional-proteomics franchise (NanoString CosMx, timsOmni) for the post-genomic era.
Isomorphic LabsPRIVFrontier disruptorRaised $600M in 2025 (Thrive Capital, GV, Alphabet) to build a next-gen AlphaFold-3-based drug-design engine; first clinical trials guided for end-2026.
The map is Divergent Compute’s editorial positioning, offered as a lens, not a measurement. Every figure in the right-hand column is drawn from a named source — see Sources.

03 · The two clocks

The spend, and the payoff

Three clocks are running at once: how fast models improve, how fast instruments commoditise, and how fast regulators write the rules.

Disclosed company figures: Tempus, Agilent, 10x, Schrödinger FY2025 reports; Isomorphic 2025 raise (PRNewswire / Fierce Biotech).

The capability clock is fast. AlphaFold 3, released in 2024 by Isomorphic Labs and Google DeepMind, predicts the structure and interactions of all of life's molecules — and Isomorphic raised $600 million in 2025 to push a successor engine toward its first clinical trials by end-2026.

The commoditisation clock is the threat to hardware. With >75% of labs planning AI deployment within two years and services the fastest-growing tools segment (~11% CAGR per Mordor), the recurring profit pool is shifting from instruments toward interpretation and data.

The regulatory clock is the wildcard. In January 2025 the FDA issued its first draft guidance on AI to support regulatory decision-making across nonclinical, clinical and manufacturing phases — a credibility framework that will decide how fast AI-generated evidence reaches approval.

04 · Private flagships

The AI-native challengers

The companies attacking this industry AI-first, with disclosed funding where available:

Isomorphic Labs

AI-native drug design

Alphabet's DeepMind spinout aims to design drugs end-to-end from AlphaFold 3, with partnered programs at Eli Lilly and Novartis and internal oncology/immunology work.

$600M raised March 2025 (Thrive Capital, GV, Alphabet); first trials guided end-2026

Tempus AI

Multimodal oncology data

Pairs genomic diagnostics with a clinical-molecular data library that pharma licenses; the Ambry Genetics acquisition extended hereditary and rare-disease testing.

Public (NASDAQ: TEM); $1.27B 2025 revenue, Data segment $316.4M

Schrödinger

Physics-based + AI discovery

Sells computational chemistry software and runs its own AI-designed pipeline, monetising via licenses plus discovery milestones and royalties.

Public (NASDAQ: SDGR); $256M 2025 revenue, $200M software

Recursion

Industrialised wet-lab AI

Combines high-throughput automated biology with ML after absorbing Exscientia; reworking its pipeline toward fewer, higher-conviction programs.

Public (NASDAQ: RXRX); ~$688M Exscientia merger, Q1'25 rev $15M

Benchling

R&D data cloud

The lab-informatics system of record for biotech; the connective software layer where experimental data and AI workflows increasingly converge.

Private; ~$412M raised, $6.1B peak (2021), ~$2.4B secondary mark (2024)

NVIDIA BioNeMo

AI infrastructure layer

Foundation-model platform and toolkit adopted across genomics and drug discovery, partnering with Illumina, Arc Institute, IQVIA and Mayo Clinic.

Part of NVIDIA (NASDAQ: NVDA); platform, not standalone funding

05 · Signals

What moved, and what to watch

Jan 2025

FDA issues first AI draft guidance

Draft guidance on AI to support regulatory decision-making across nonclinical, clinical, post-market and manufacturing — a credibility framework for AI-generated evidence.

Mar 2025

Isomorphic raises $600M

Alphabet's drug-design unit takes its first external capital to build a next-gen engine and move AlphaFold-3-designed candidates toward the clinic.

2025

Thermo Fisher partners with OpenAI

The largest tools company embeds OpenAI APIs across product, service and operations, citing 30% effective-capacity gains at bioproduction sites.

2025

Danaher names a CTO and AI Officer

Promotion of an ex-Tempus/Google AI leader plus a Microsoft AI board hire signals a strategic, top-down AI reorientation.

FY2025

Tempus crosses $1.27B revenue

Diagnostics up ~112% for the year and a $316M data business prove AI-data monetisation at scale in the tools-adjacent layer.

06 · The exposure read

Who’s defensible, who’s at risk

AI rewards clean, structured advantage and punishes friction. The line runs through who owns the data, the brand and the customer — and who is merely a step the technology can route around.

Defensible

  • Owners of proprietary interpretation models — Illumina's DRAGEN/variant-AI stack and Tempus's multimodal data turn raw signal into reimbursable answers, the layer least exposed to hardware commoditisation.
  • Scaled incumbents that can buy or partner their way to AI — Thermo Fisher and Danaher have the cash, install base and bioproduction footprint to operationalise AI for real productivity gains, not demos.
  • Infrastructure and foundation-model providers — NVIDIA's BioNeMo and the AlphaFold lineage sit underneath everyone's pipeline, collecting value regardless of which drug or diagnostic wins.
  • Data-network compounders — platforms like Benchling and Tempus whose value grows with every experiment logged build switching costs that hardware never could.

At risk

  • Undifferentiated hardware vendors — as interpretation moves to software and >75% of labs adopt AI, instruments without a defensible model layer drift toward commodity economics.
  • Cash-burning AI-native biotechs — Recursion's $15M Q1'25 revenue against a large clinical portfolio shows how far validation and monetisation still lag the funding narrative.
  • Single-modality tool makers facing volume pressure — 10x Genomics' ~2% FY2025 revenue decline illustrates how quickly demand can stall even as AI rebuilds the longer-term pipeline.
  • Anyone betting AI evidence sails through regulators — the FDA's still-draft 2025 framework means the path from AI-generated data to approval remains unsettled and slow.
The tools industry will not be disrupted out of existence — every AI model still needs physical samples, reagents and signal. But the profit is migrating from the box to the model on top of it. The winners of 2026-2030 will be the companies that already treat their instruments as data-acquisition front ends and price the intelligence, not the iron.

Sources

Where this comes from

Mordor Intelligence — Life Science Tools Market Size  ·  Thermo Fisher — OpenAI collaboration (IR)  ·  Illumina — AI in Genomics  ·  Illumina + NVIDIA collaboration  ·  Klover.ai — Danaher AI strategy (Stumpe CTO/AI)  ·  Tempus — Q4 & Full Year 2025 Results  ·  Schrödinger — Q4/FY2025 Financial Results (IR)  ·  Agilent — OmixAI proteomics partnership  ·  10x Genomics — Q4/FY2025 results (PRNewswire)  ·  Isomorphic Labs — $600M funding (PRNewswire)  ·  Recursion + Exscientia merger (IR)  ·  Grand View Research — AI in Drug Discovery Market  ·  NVIDIA — BioNeMo / life-sciences partnerships  ·  FDA — Artificial Intelligence for Drug Development