// Account brief

Roche

Life Sciences Radar, October 2026 edition. Evidence published up to 5 October 2026; job sites read on 6 October 2026. This brief covers AI in drug discovery, and research labs and data. Every link opens the public source. This brief covers Roche and Genentech. Chugai, which is separately listed with its own R&D, is out of scope and was not searched.

Disclosure: Andrew Wyatt, Ortent's principal, was Chief Growth Officer at Sapio Sciences, a lab software company, until April 2026. He has no current relationship with, or financial interest in, Sapio or any other supplier. Software vendors, including Sapio and its competitors, appear in this brief only where public sources name them.

Status: Ortent has reviewed this brief; it has not been independently validated. It sets out what the company and others have published. Search coverage varies by company and is described under Coverage limits, and a count of zero means none was identified, not that nothing happened. It does not predict buying intent or measure operational maturity. For this radar, process and CMC development are grouped with manufacturing and excluded from research labs and data.

Scores

Theme Priority (1 to 10) Delivery (0 to 10) Stage Commitments made public, last 12 months R&D roles open on 6 Oct 2026
AI in drug discovery 8 7 scaled 4 24
Research labs and data 5 4 live 1 7

The scores rest only on what the company itself has published about each theme. Commitments and open roles are observed activity: they show what is being started or staffed now, and they never change a score.

Summary

Roche puts AI at the centre of its R&D plans, and AI in drug discovery scores 8 out of 10 for priority and 7 for delivery, at the scaled stage. It reports a hybrid-cloud AI factory of more than 3,500 NVIDIA Blackwell GPUs, a first AI-enhanced pathobiology map deployed in IBD and four more in neurology, and AI taking part in all large-molecule programmes and about 90% of eligible projects. Research labs and data scores 5 for priority and 4 for delivery, at the live stage: Roche describes work on a simpler set of systems and a data foundation, an Autonomous Lab Accelerator, and a Medra physical AI platform that integrates with Genentech's laboratory information management system. Four AI in drug discovery commitments were made public in the last 12 months, including the NVIDIA AI factory expansion and the Earendil Labs antibody collaboration, and two in the year before; the Medra collaboration is the one research labs and data commitment. On 6 October 2026 Roche had 24 R&D roles open in AI in drug discovery and 7 in research labs and data, including lab automation software engineers.

AI in drug discovery

What the company says: Priority 8

Revolutionizing drug R&D with Lab-in-a-Loop Aviv Regev, Head of Genentech Research and Early Development (gRED)

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 5

Our ‘R&D Excellence’ initiative, launched two years ago to accelerate our research and development, is showing strong progress.

Roche Annual Report 2025, 29 Jan 2026, p. 12

Evolve our R&D engine and invest in its excellence Ongoing R&D investments Boston Innovation Center focusing on CVRM and AI in drug development opened

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 62

2026 AI milestones ... >3500 Blackwell GPUs in hybrid-cloud AI factory, incl. 2 new data centers, one in the US and one in the EU

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 62

With the addition of 2,176 NVIDIA Blackwell GPUs, Roche now operates the pharmaceutical industry’s largest announced hybrid-cloud AI factory, totaling more than 3,500 GPUs.

Media release: Roche launches NVIDIA AI factory to accelerate the development of new therapeutics and diagnostics solutions, 16 Mar 2026

In R&D, the NVIDIA BioNeMo platform enhances Roche’s Lab-in-the-Loop, where biological and chemistry experiments are connected with Roche’s AI models.

Media release: Roche launches NVIDIA AI factory to accelerate the development of new therapeutics and diagnostics solutions, 16 Mar 2026

NVIDIA partnership OpenAI pilot partnership AI factory (NVIDIA) Compute & software

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 71

TargetNexus Our AI agent for target assessment on track to be part of 80% of research portfolio decisions by Q4

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 80

we launched the Computational Sciences Center of Excellence (CSCoE) in 2025

Genentech story: AI Fuels Genentech's R&D Ecosystem, 9 Jan 2026

John Marioni, Senior Vice President at gRED and Head of the Computational Science Center of Excellence

Genentech story: AI Fuels Genentech's R&D Ecosystem, 9 Jan 2026

What it shows running: Delivery 7, scaled

Pathobiology maps have scaled and accelerated our pipeline Example: IBD pathobiology maps

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 76

BRAID (biology AI) PrescientDesign (antibodies AI) gRED Computational Sciences (gCS) consolidated ... Recursion collaboration ... NVIDIA partnership OpenAI pilot partnership AI factory (NVIDIA)

Roche Pharma Day 2026 presentation (London, 28 September 2026), 28 Sep 2026, p. 71

AI-enhanced pathobiology maps across our TAs ... +155% First map deployed (IBD) 4 neurology maps deployed increase in LI entries in 2 years

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 77

AI participating: 100% large molecules / 50% small molecules ... ~90% of eligible projects

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 71

40% of Q4 ’25 through Q2 ’26 pipeline decisions had a tracked AI / computational contribution ... Cumulative coverage: 40% (Q2 coverage reached 44%)

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 73

Today, all of our antibody programs and ~90% of eligible small molecule programs integrate AI to streamline discovery

Genentech story: AI Fuels Genentech's R&D Ecosystem, 9 Jan 2026

we used AI to deliver a backup small molecule with a distinct scaffold for an important disease target in just seven months, versus more than two years for the original lead

Genentech story: AI Fuels Genentech's R&D Ecosystem, 9 Jan 2026

Steady increase in research portfolio productivity 2022-2026 Increase from LI-Go to LO-Go to development candidates under nominally flat resources and halving LI CT ... 2022-2025 growth +155% LI Go

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 72

Increase in high quality volume of research portfolio entries Large-scale human and lab data, AI, and new processes increase high-quality volume and speed

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 75

AI-enhanced pathobiology maps across our TAs ... 1 YEAR 6 WEEKS +155% First map deployed (IBD) 4 neurology maps deployed increase in LI entries in 2 years

Roche Pharma Day 2026 presentation (London, 28 September 2026), 28 Sep 2026, p. 77

Commitments made public in the last 12 months: 4

  • Earendil Labs bispecific antibody discovery collaboration: Genentech and Earendil Labs agree research collaboration; Earendil’s AI-driven platform to discover bispecific oncology antibodies. New, announced, 24 Sep 2026. Published by: partner release, company quoted.
  • NVIDIA collaboration and hybrid-cloud AI factory: Roche adds 2,176 NVIDIA Blackwell GPUs; hybrid-cloud AI factory exceeds 3,500 GPUs; BioNeMo supports Lab-in-the-Loop in R&D. Expanded, operational, 16 Mar 2026. Published by: company.
  • Computational Sciences Center of Excellence (CSCoE): Roche and Genentech launched a Computational Sciences Center of Excellence in 2025 to scale AI across research. New, operational, 9 Jan 2026. Published by: company.
  • Medra physical AI lab collaboration: Genentech collaboration with Medra: physical AI platform integrated with Genentech ML infrastructure and LIMS for lab-in-the-loop experiments. New, operational, 11 Dec 2025. Published by: company.

Commitments made public in the year before: 2

Search coverage for this theme: partial. Commitments count once per initiative, however many announcements they get.

R&D roles open on 6 October 2026: 24

These are postings listed on one day, not hires, headcount or a trend. Roles that serve both themes appear under each.

Technology named in these postings

Named as in use: AWS, Kubernetes, Terraform, PyTorch.

Asked for as experience: Helm, LangGraph.

A product named in a job advert shows what a team wants people to know. It is not proof of a purchase or an installed system.

Research labs and data

What the company says: Priority 5

R&D Excellence: Our solutions Moving from implementation to “business as usual” ... Build a simplified system landscape and data foundation

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 49

Our ‘R&D Excellence’ initiative, launched two years ago to accelerate our research and development, is showing strong progress.

Roche Annual Report 2025, 29 Jan 2026, p. 12

Towards autonomous labs for drug R&D Our Autonomous Lab Accelerator to enhance the loop

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 94

Medra physical AI, MHS (Anthropic) Autonomous Lab Accelerator

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 71

Foundational Datasets for Foundation Models initiative Large-scale data generation dedicated for training of AI models of relevant disease biology

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 78

Andy Lash, Head of Engineering, Lab Automation

Genentech story/podcast transcript: Lab of the Future, 14 Jul 2026

What it shows running: Delivery 4, live

Medra's platform integrates with Genentech’s internal ML infrastructure and its laboratory information management system

Genentech story: AI Fuels Genentech's R&D Ecosystem, 9 Jan 2026

Medra physical AI, MHS (Anthropic) Autonomous Lab Accelerator

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 71

Build a simplified system landscape and data foundation Implemented and moving into business as usual Implementation in progress

Roche Pharma Day 2026 presentation, 28 Sep 2026, p. 49

the latest generation of the most standard robotic arm we use in lab automation has an integrated camera

Genentech story/podcast transcript: Lab of the Future, 14 Jul 2026

Multiple data generation streams at scale Ability to iterate (lab and AL/RL) Autonomous lab capabilities

Roche Pharma Day 2026 presentation (London, 28 September 2026), 28 Sep 2026, p. 100

we use a lot of separate software, a lot of separate automation systems that today are not very uniformly connected to one another

Genentech story/podcast: Lab of the Future (Two Scientists Walk Into A Bar), 14 Jul 2026

Commitments made public in the last 12 months: 1

  • Medra physical AI lab collaboration: Genentech collaboration with Medra: physical AI platform integrated with Genentech ML infrastructure and LIMS for lab-in-the-loop experiments. New, operational, 11 Dec 2025. Published by: company.

Commitments made public in the year before: 0

None identified in the reviewed sources.

Search coverage for this theme: partial. Commitments count once per initiative, however many announcements they get.

R&D roles open on 6 October 2026: 7

These are postings listed on one day, not hires, headcount or a trend. Roles that serve both themes appear under each.

Technology named in these postings

Named as in use: AWS.

Asked for as experience: Kubernetes.

A product named in a job advert shows what a team wants people to know. It is not proof of a purchase or an installed system.

Two questions to investigate (Ortent's interpretation)

  1. Lab automation and lab data work sits in several Roche and Genentech units, including the Computational Sciences Center of Excellence, R&D digital technology teams and Genentech lab automation engineering. Who decides on new lab automation platforms and on links to LIMS and ELN, and how is the work split between these units and the research sites?
  2. Compute and external AI deals for discovery are handled in several places: the AI factory, the research teams' own ML infrastructure on AWS, and business development for deals such as Dyno and Earendil. Who holds decision rights over compute and over external AI model or platform agreements for discovery, and how is discovery use of the AI factory funded and allocated?

Coverage limits

The web search budget ran out early, so results presentations, partner newswires, consortium and academic sources were only partly searched, and no results-call transcripts were read. Business Wire, Fierce Biotech, MobiHealthNews, SEC EDGAR and Recursion's investor site blocked access and GlobeNewswire timed out, so partner posts and copies were used where possible. Several Genentech stories were edited after publication, so the first public dates of some items are uncertain. The scores rest on what could be read, and activity disclosed only in unread sources is not reflected.

// GET IN TOUCH

Clarity when it counts.

For a board, advisory or fractional engagement, or a working call on the next stage, send a note. Thirty minutes is usually enough to test fit.

Request a 30-minute intro call