Novartis

Ortent Life Sciences Radar, research beta. Evidence to 6 October 2026, rule set v0.4.3. Scope: AI and computation in drug discovery (T1), and research-lab digitalisation and scientific data systems (T2). Each ID in brackets links to its online source, and the source index at the end of this brief lists them.

Disclosure: Andrew Wyatt, Ortent's principal, was Chief Growth Officer at Sapio Sciences until April 2026. Software vendors, including Sapio and its competitors, appear in this brief only where public sources name them.

Status: research beta. Ortent has reviewed this brief; it has not been independently validated. It summarises activity identified in public sources. Search coverage varies by company and is set out under Coverage limits, and zero identified items does not mean no activity. It does not predict buying intent or measure operational maturity. For this radar, process and CMC development are grouped with manufacturing and excluded from the research-lab theme (T2); such work appears as context where it helps.

Summary

The reviewed sources identify seven external discovery collaboration announcements with a stated AI or computational component involving Novartis between November 2024 and July 2026, three of them in the year to 5 October 2026 [>NVS-E01, >E03, >E04, >E05, E07, >E12, >E13]. A reaction-data collaboration with CAS was announced on 29 September 2026 [>NVS-E16]. Company material describes Data42, a research data platform in Biomedical Research [>NVS-E17]. The 6 October recruitment snapshot contains 16 classified R&D requisitions across the two themes [NVS-R01a to R14]. These observations identify programmes and functions for further investigation; they do not establish procurement opportunity, deployment completion or net hiring.

Announced commitments

Every item in the table was announced first by the partner or reported first in media. Novartis confirmed three of the nine in its own documents, through the 2025 annual report published on 4 February 2026 [>NVS-E03, >E04, E07]. We found no Novartis confirmation of the other six in the reviewed sources [>NVS-E01, >E02, >E05, >E12, >E13, >E16].

The CAS collaboration is a T2 item and the Schrodinger informatics licence is tagged to both themes; the other rows are T1 [>NVS-E16, >E02]. Under rule set v0.4.2 the Schrodinger collaboration and the expanded licence are separate initiatives, because the release calls the collaboration a new agreement and uses "expanded" only for the software licence [>NVS-E01, >E02].

The counts are the same by event date and by first disclosure, because every counted row has an exact date. In T1, five initiatives had a qualifying commitment change in the year to 4 October 2025 and three in the year to 5 October 2026. The earlier figure was four under rule set v0.4.1, before the Schrodinger collaboration was counted apart from the licence. In T2 there was one in each year: the Schrodinger informatics licence, then the CAS collaboration. No negative changes were found in either theme. Separately, one T1 milestone (the San Diego groundbreaking) and two T2 newly observed capabilities (SPT Labtech sample transport and the Data42 leadership) were disclosed [>NVS-E09, >E15, >E17].

Coverage was partial for every event type in both themes and both years, and no cell is marked comparable. No comparison between the two years is made.

Identified recruitment

On 6 October 2026 the reviewers screened 51 archived Novartis job descriptions from the Workday careers site and classified 38 requisitions from their duties [NVS-R01a to R34]. This is a screened candidate set, not a full census of the careers site [reviewer note B/g1]. Of these, 16 are R&D roles in the two themes. Split by theme they are 8 T1 and 10 T2, because two roles are tagged to both themes and counted once in each [>NVS-R04, >R05]. They cover 12 distinct postings, because three roles are posted under several requisition numbers in different locations [NVS-R01a, R01b, R12a, R12b, R13a to R13c]. A further 13 are enterprise AI and data roles, four are plant or quality roles, two are postdoctoral fellowships classed as research training, and three are out of scope. These are open requisitions on one day, not hires, headcount or a trend.

The R&D postings name roles across discovery data science, lab automation and research data: - AI scientists for AI-driven target identification in Oncology Data Science, posted in Basel and Cambridge [NVS-R01a, R01b] - applied AI for small-molecule hit generation and lead optimisation within design-make-test-analyse cycles [>NVS-R02], a computational medicinal chemist in the CADD team in La Jolla, San Diego [>NVS-R03], and a data science lead for Lab-in-the-Loop hit finding in San Diego [>NVS-R05] - AI image analysis of high-content screening and organoid images [>NVS-R04], and machine learning with mechanistic modelling for compound progression in PKS Modeling & Simulation [>NVS-R08] - an HPC infrastructure architect in IT Cloud & Hosting in Hyderabad, whose duties cover scientific workloads; this is a borderline classification [>NVS-R09] - a Basel scientist splitting time between lab automation and liquid handling support and assay work, and an HTS and automated cell culture robotics engineer in Cambridge [>NVS-R10, >R11] - Data42 forward deployed engineers building data pipelines and applications on Palantir Foundry in Cambridge and Basel; this is a borderline classification, because the posting names no lab or discovery data [NVS-R12a, R12b] - a Director of technology and engineering in Biomedical Research Data & Digital, leading AI-enabled software across scientific workflows and laboratory processes, posted in Basel, Hyderabad and Prague [NVS-R13a to R13c] - a senior principal data scientist in the Biologics Research Center to drive the Biomedical Research data strategy and scientific data governance across research sites [>NVS-R14]

Posting dates are Workday start dates, which reset when a role is reposted, so they may be refresh dates [NVS-R01a to R34].

LinkedIn Premium company insights, read through an Ortent account on 6 October 2026, show Research openings down 38% over six months and total headcount up 6% over a year (aggregate LinkedIn member data). These are generic function figures. They do not measure AI or lab-automation hiring and are context only.

Technology mentions in recruitment

Mentions are classified from the exact posting wording. They do not establish installation, an incumbent contract or company-wide use. - Named as existing systems: the Data42 engineer posting places the role "at Data42" and says the engineer will "rapidly prototype, iterate, and productionize solutions on Palantir Foundry" [NVS-R12a; the same text appears in R12b]. - Category stated as the employer's own: the HPC architect posting says Cloud & Hosting Services "has established AI & HPC Engineering as the single engineering and delivery channel for high-performance computing". This is role-specific wording about one unit. It names no product [>NVS-R09]. The same posting asks the architect to design CPU and GPU nodes, which does not state that GPU nodes are in use (unclear) [>NVS-R09]. - Integration responsibilities: no R&D posting was classified as an integration or environment statement. The only one is a site robotics role at the Schaftenau manufacturing site, which is to connect with PLC/SCADA, MES and SAP [>NVS-R17]. - Preferred experience: the HTS automation engineer posting rates knowledge of Biosero Orchestrator and Green Button Go scheduling software as highly desirable. It also lists HighRes Biosolutions and Cellario, CellTrio RoboCell, Beckman Echo and GNF Director platforms, and experience integrating automated systems to LIMS "either in-house or commercially available" [>NVS-R11]. - Lists of examples: Genedata Screener, Spotfire and GraphPad Prism as screening and analysis software [>NVS-R10, >R11]; Schrodinger, CCG and OpenEye as examples of computational drug design tools [>NVS-R03]; image analysis tools and a deep learning stack, and Claude Code or Copilot [>NVS-R04]; Slurm or LSF as workload schedulers [>NVS-R09]; PyTorch, scikit-learn and Keras as machine learning libraries [NVS-R01a, R08]. - Outside R&D: a Morrisville quality control role covers "building, maintaining, and optimizing LabWare LIMS that supports Quality Control operations" and Empower CDS [>NVS-R15].

Deployment evidence

• A Novartis story of 28 July 2026 says AI use "is playing out in focused initiatives across Novartis R&D", naming target identification and validation, generative chemistry and preclinical safety. It describes what the models can do and reports a qualitative benefit, with teams spending less time sifting through data. We found no adoption or benefit metric for it in the reviewed sources [>SRC-NVS-AIRD].
• Christian Diehl, Chief Data & Digital Officer, Biomedical Research, writes that he developed Data42 within Biomedical Research and that it "underpins our growing ecosystem of AI-driven research" by connecting clinical, molecular, preclinical and biology data (Novartis story, 23 September 2026). We found no adoption metric for Data42 in the reviewed sources [>NVS-E17] [>SRC-NVS-DATA42].
• An SPT Labtech blog of 22 September 2025 describes its pneumatic sample transport system in Novartis analytical labs at the Basel and Cambridge sites, and says Novartis processes up to 3,000 vials a day worldwide. The source is vendor-authored, cites an undated interview and states no deployment date [>NVS-E15].
• The Schrodinger release says the expanded licence enables Novartis to deploy Schrodinger's drug discovery technologies across its research sites. That is agreement wording; we found no report of the extent of deployment [>NVS-E02].
• Novartis broke ground on 6 February 2026 on a San Diego Biomedical Research centre of about 466,000 sq ft with AI-enabled discovery capabilities, expected to house about 1,000 employees and to be operational in 2029. It is in build [>NVS-E09].

Context

Fiona Marshall, President of Biomedical Research, is quoted in the Schrodinger, Monte Rosa, Relation and Chai releases [>NVS-E01, >E05, E07, >E13]. A Novartis story says she and Shreeram Aradhye, President of Development, have defined the R&D strategy for integrating these technologies [>SRC-NVS-AIRD]. John Tallarico, Head of Discovery Sciences, is quoted in the Orionis release [>NVS-E12]. No appointment date for Christian Diehl as Chief Data & Digital Officer was found in the reviewed sources [>NVS-E17].

The Board Science and Technology Committee reviewed AI use cases during 2025, including "New Target Identification" and AI-augmented chemistry [>NVS-E10]. Novartis states that data, digital and AI will be a standing Board topic from 2026 [>NVS-E11]. The 2025 annual report says Novartis launched Lean Digital Core, a transformation programme for core business processes and technologies in Germany, the UK, the Netherlands and the US [>SRC-NVS-AR2025]. Thirteen enterprise AI and data requisitions sit outside R&D, including AI Foundations engineering, an Agentic Lab and semantic and knowledge engineering [>NVS-R19 to >R31].

The San Diego site was first announced on 10 April 2025 within a USD 23 billion five-year US plan, with no AI or lab-automation wording at that point [>NVS-E08].

Swissinfo (Keystone-SDA) reported on 2 September 2026 that Novartis is cutting up to 130 jobs in Basel by the end of 2027. It is discontinuing small-scale biologics production in Kleinbasel and relocating cell-bank, analytical testing and technical development labs to the Novartis Campus [>NVS-E19]. The source does not link this to AI in discovery or lab digitalisation, and we identified no qualifying pause or cancellation in the two themes in the searched sources. That is a search result, not evidence that none occurred.

Coverage limits

• Coverage is partial for every event type in both themes and both years. No FY2026 annual report exists yet, so the latest year relies on the FY2025 report and interim material.
• The novartis.com newsroom has no keyword filter and was not enumerated. Some quarterly media releases and investor presentations were not re-read.
• No dedicated acquisition web search was run for Novartis. Reviewer A ran one web search per event type spanning both years, not one per year [reviewer notes B/g1, A/g1].
• Patents and preprint servers were not searched. Publication counts are raw query counts and are not comparable across companies.
• fiercebiotech.com blocked automated access. A May 2026 media report on Biomedical Research roles could not be archived, is not verified and is not used in this brief [>NVS-E18].
• The Isomorphic Labs page links to a Novartis release that was not retrieved. The ProFound and Monte Rosa rows rest on media copies, because the primary releases were not read [>NVS-E03, >E04, >E05].
• Reviewer A found no end or renewal notice for a 2019 five-year Microsoft AI alliance, so no row is recorded [reviewer note A/g1].
• Technology mentions were not extracted from the enterprise or out-of-scope requisitions [reviewer note B/g1].
• The archived Novartis postings show no openings field, so vacancy status is not recorded on every Novartis row and postings listed with zero openings could not be identified.

Two questions to investigate (Ortent's interpretation)

1. Biomedical Research has a Chief Data & Digital Officer and a Data & Digital function. It is recruiting engineers for Data42 on Palantir Foundry, a Director for AI-enabled software across scientific workflows and laboratory processes, and a data strategy and governance lead [>NVS-E17, R12a, R12b, R13a to R13c, >R14]. External partners have announced reaction-data work with CAS and an expanded Schrodinger informatics licence [>NVS-E16, >E02]. What is the remit of Data & Digital over these platforms and partner services? Who holds decision rights on additions to research data and lab software, and how are the interfaces between Data42, partner platforms and laboratory systems governed?
2. Novartis describes an R&D approach that combines collaborations with AI companies and tools developed internally [>SRC-NVS-AIRD]. Partners have announced model and platform access (Chai, Schrodinger), research computing is recruited within IT Cloud & Hosting, and enterprise AI platform roles sit outside R&D [>NVS-E13, >E02, >R09, >R19 to >R25]. Who decides between building, partnering and buying for discovery AI tools? How is access to research models and compute procured and funded, and what is the interface between Biomedical Research teams and the enterprise AI platforms?

Sources for the main claims

• NVS-E01, E02 Schrodinger and Novartis: https://s203.q4cdn.com/609444515/files/doc_news/Schrdinger-Announces-Multi-Target-Collaboration-and-Expanded-Software-Licensing-Agreement-with-Novartis-2024.pdf
• NVS-E03 Isomorphic Labs collaboration expansion: https://www.isomorphiclabs.com/articles/isomorphic-labs-announces-novartis-collaboration-expansion
• NVS-E04 ProFound Therapeutics (BioSpace): https://www.biospace.com/deals/novartis-makes-potential-750m-bet-per-cardiovascular-disease-target-in-another-flagship-pact
• NVS-E05 Monte Rosa (BioPharm International): https://www.biopharminternational.com/view/monte-rosa-and-novartis-partner-to-target-undruggable-immunology-pathways-with-degraders
• NVS-E07 Relation Therapeutics: https://www.manilatimes.net/2025/12/09/tmt-newswire/globenewswire/relation-announces-strategic-collaboration-with-novartis-to-advance-therapeuti
• NVS-E12 Orionis Biosciences: https://orionisbio.com/2026/06/novartis-collaboration-2026/
• NVS-E13 Chai Discovery: https://www.aol.com/articles/chai-discovery-announces-collaboration-novartis-110000000.html
• NVS-E16 CAS: https://www.cas.org/press-releases/cas-announces-collaboration-with-novartis-to-advance-ai-ready-reaction-informatics
• NVS-E08 to E11, E15, E17, E19 and all requisitions: URLs and archived copies in the v0.4 tables
• SRC-NVS-AIRD: Novartis story, 28 July 2026 (kittest file A_NVS-13)
• SRC-NVS-DATA42: Novartis story, 23 September 2026 (kittest file A_NVS-15)
• SRC-NVS-AR2025: Novartis Annual Report 2025 (kittest file A_NVS-01)

Source index

Each ID in brackets links to the online source for that row. Job posting links come from a snapshot of the company job site on 6 October 2026; postings that have closed since are marked and no longer linked. Where a company site blocked access, the link goes to the copy that was read.