Bayer

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), for the Pharmaceuticals division of Bayer AG only. Crop Science and Consumer Health items are excluded, and a Leaps by Bayer investment counts in T1 only where the investee's stated platform is AI or computational discovery for human therapeutics. 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 three T1 initiatives with an AI-discovery commitment change involving Bayer's Pharmaceuticals division first disclosed in the year to 5 October 2026, and one in the year before [>BAY-E01, >E02, >E03, >E12]. The three are collaborations with Cradle, Vanderbilt University Medical Center (VUMC) and Iambic; the one is the acquisition of Tavros by Bayer's subsidiary Vividion [>BAY-E01, >E03, >E02, >E12]. An expansion of Bayer's TetraScience Scientific Data Foundry deployment is the only T2 commitment change identified, in the year to 5 October 2026 [>BAY-E10]. The 6 October recruitment snapshot contains 5 relevant R&D requisition postings across the two themes, none with ambiguous vacancy status [>BAY-R01 to >R05]. These observations identify programmes and functions for further investigation; they do not establish procurement opportunity, deployment completion or net hiring.

Announced commitments

Bayer issued three of these five rows jointly with the partner [>BAY-E01, >E02, >E03]. The Tavros acquisition was announced by Vividion on the Bayer newsroom, with a Bayer executive quoted [>BAY-E12]. The TetraScience expansion was announced by TetraScience with a Bayer lead quoted, and we found no Bayer confirmation of it [>BAY-E10]. Bayer's main site outside the newsroom was blocked to both reviewers, so any confirmation published only there was not reachable.

The Tavros acquisition counts in T1 because Bayer's Annual Report 2025 says it expands Vividion's "proprietary methods for computer-based genomic screening" [>BAY-E12]. Vividion is wholly owned and not separately listed, and Bayer reports the acquisition itself, so the group scope clarification C42-12 does not exclude it. The annual report also gives terms the release did not: an upfront consideration of around €19 million and up to around €29 million on research and development milestones [>BAY-E12]. The VUMC deal counts in T1 only because of the later Bayer sentence on AI-driven analytics; on the January release alone it would be context, as access to external data (AR7a) [>BAY-E03]. The TetraScience expansion counts in T2 because the release states added scope across pharmaceutical R&D and new self-service tooling (rule 1a); the crop science share is out of scope [>BAY-E10]. Cradle is T1 only, because the release describes an AI design platform, not a lab or scientific data system that Bayer deploys [>BAY-E01].

By event date, we counted one T1 initiative with a qualifying commitment change in W0 (5 October 2024 to 4 October 2025), the Tavros acquisition, and three in W1 (5 October 2025 to 5 October 2026): Cradle, VUMC and Iambic. By first disclosure, the T1 count is also one in W0 and three in W1. T2 is zero in W0 and one in W1 in both views: the TetraScience expansion [>BAY-E10]. No milestones were disclosed in either theme. Existing capabilities first disclosed in W1 numbered five in T1 and one in T2: the FinnGen and PRECISE partnerships, the Head of Data Science & AI, an in-house AI platform and the Machine Learning Research group, and Bayer's existing use of the TetraScience Scientific Data Foundry [>BAY-E04 to >E08, >E11]. These have no stated start date and are not counted as changes. Under the opposite reading of borderline clarification C42-2, counting the in-licensed MOMA and Dewpoint programmes as T1 commitments, T1 W0 would be three in both views [>BAY-E13, >E14]; under a reading of borderline clarification C42-12 that covers any arm's-length subsidiary such as Vividion, the Tavros acquisition would be excluded and T1 W0 would be zero in both views [>BAY-E12]. Two further calls are borderline: read as context data access, VUMC would make T1 W1 two [>BAY-E03]; read as an expansion with no qualifying added scope under rule 1a, the TetraScience expansion would make T2 W1 zero [>BAY-E10].

Coverage is partial for both themes in both years. The consortium, academic collaboration and negative-event search families are complete in every cell, and the annual report family is complete in W1 only. The company press release, results, partner newswire, corporate venture and per-event-type search families are partial in every cell. The table marks the annual report, partner newswire and consortium families as not comparable between the years, so no comparison between the years is made.

Two in-licensed programmes and two incubator items are recorded as context (see Context).

Identified recruitment

On 6 October 2026 the talent.bayer.com job data listed 598 requisitions by one reviewer's count, and the jobs.bayer.com feed listed 619 postings by the other's. Each reviewer's common title filter matched 29 requisitions. The two reviewers classified 36 requisitions from their duties, 33 of them screened by both. We observed 5 relevant requisition postings in the R&D theme class, 5 T1 and 2 T2; the two roles tagged both appear in each split [>BAY-R01, >R02]. No Bayer posting shows an openings field, so vacancy status is not recorded for any of the 5 and none has ambiguous vacancy status. A further 9 are enterprise roles, 5 are internships or working-student roles, 1 is a manufacturing role and 16 are out of scope, 12 of them because the posting belongs to Crop Science or Consumer Health. These are requisition postings listed on one day, not hires, headcount or a trend.

The T1 postings describe AI and modelling roles in discovery: - two postings for a Head of Machine Learning Research, one in Whippany or Cambridge, Massachusetts and one in Berlin. The role sits on the Data Science & AI leadership team, is to build and lead a team of about 25, covers ML molecular design from hit identification to candidate nomination and makes "build/buy/partner decisions" across generative design, property prediction, structure-based ML and synthesis planning [>BAY-R01, >R02] - a Senior Principal Scientist for quantum mechanics and modelling in Wuppertal-Aprath, to apply quantum mechanics, physics-based modelling, machine learning and AI to compound design across the discovery portfolio, working closely with the Data Science & AI unit [>BAY-R03] - a Principal Machine Learning Scientist in Berlin, for machine learning in the characterisation and design of biomolecules to accelerate drug discovery [>BAY-R04] - a Principal Machine Learning Scientist in Wuppertal and Berlin, in the Machine Learning Research group, for biological foundation models and virtual cell and virtual patient models, with target discovery among the stated uses [>BAY-R05]

The two Head of Machine Learning Research postings are also tagged T2. The US posting is to work with data engineering to "secure the assay, structural, and DMTA data foundation" and to set up MLOps practice; the Berlin posting says the holder will build "a scalable data and MLOps foundation with curated and harmonized assay, structural, and DMTA data" [>BAY-R01, >R02]. No other posting is tagged T2.

None of the five internship or working-student postings is tagged to the themes. They include a drug substance digital process chemistry intern in Wuppertal, whose team combines synthesis with automation and data science but whose listed tasks are bench work, and a data science co-op in Berkeley for bioprocess monitoring across process development and GMP manufacturing [>BAY-R06, >R07]. Enterprise postings include AI roles in the Warsaw enterprise data and analytics platform, which serves Finance, Supply Chain, HR, Procurement, Legal and Communications, an AI engineer for SAP S/4HANA in Bangalore and an AI governance role [>BAY-R13, >R15, >R14, >R18]. Out-of-scope postings include Crop Science machine learning research roles, a commercial decision intelligence director, a Biological R&D software engineer for medical-device cloud connectivity and a process development lab technician whose duties name no data systems [>BAY-R22, >R23, >R28, >R35, >R29].

Posting dates come from a structured field in the page data or an unlabelled column on the search page, and may record a refresh, so they are not reliable.

No LinkedIn company insights were collected for Bayer.

Technology mentions in recruitment

Mentions are classified from the exact posting wording. They do not establish installation, an incumbent contract or company-wide use. The table records 57 Bayer mentions: 13 explicit use, 2 integration or environment, 1 required experience, 1 preferred experience, 4 planned implementation, 34 alternatives and 2 unclear. - Named as existing systems: none in the R&D postings. Every explicit-use mention is in an enterprise, manufacturing or out-of-scope posting. AWS, Azure and Databricks are named as the stack of the Warsaw enterprise data and analytics platform, and LangChain, LangGraph and PydanticAI as the frameworks for agent workflows there [>BAY-R13, >R15]. SAP S/4HANA and SAP Joule are named for the Bangalore AI engineer [>BAY-R14], AWS and Databricks for manufacturing data at Weimar [>BAY-R20], and Terraform and CloudFormation for medical-device connectivity [>BAY-R35]. This is role-specific wording. - Integration responsibilities: the US Head of Machine Learning Research is to drive closed-loop DMTA cycles, "including integration with high-throughput and automated synthesis where available"; no system is named [>BAY-R01]. Outside R&D, the Weimar data engineer's systems take data from MES, SAP, equipment databases and historians [>BAY-R20]. - Planned implementation: an MLOps practice "appropriate to a regulated R&D environment" and a data and MLOps foundation for assay, structural and DMTA data [>BAY-R01, >R02]. Outside R&D: adoption of the Microsoft AI ecosystem in a Colombian commercial unit and a move from ThingWorx to AWS IoT for device connectivity [>BAY-R19, >R35]. - Required experience: managing high-performance infrastructure and working with external IT providers [>BAY-R03]. - Preferred experience: Azure, in an enterprise posting [>BAY-R19]. - Lists of examples: ORCA and xTB [>BAY-R03]; mmseqs and foldseek [>BAY-R04]; PyTorch, pandas and scikit-learn [>BAY-R04, >R05]; lab automation, self-driving lab concepts or high-throughput chemistry integration [>BAY-R01]; cloud, graph database and agent framework examples in enterprise postings [>BAY-R07, >R12, >R16]. - Unclear: "automation and data science" in the process chemistry team [>BAY-R06], and knowledge graph solutions in an enterprise posting [>BAY-R12].

No ELN, LIMS or scientific data platform is named in any of the 36 classified postings, and no category (ELN, LIMS, cloud) is recorded as explicit use in an R&D posting.

Deployment evidence

• The TetraScience release of 20 November 2025 states that "Bayer uses the Scientific Data Foundry to centralize and harmonize scientific data from laboratory instruments and external CRO and CDMO partners" [>BAY-E11]. This is a vendor release quoting a Bayer Regulatory Science Transformation Lead, who says the expanded partnership "is delivering measurable value". No measure is given, the release covers pharmaceutical and crop science R&D without a split, and the start date of the deployment is not stated. An earlier announcement may exist beyond the newswire depth searched.
• At Bayer's Pharma Media Day on 1 April 2026, Sai Jasti, Senior Vice President and Head of Data Science & Artificial Intelligence, Bayer Pharmaceuticals R&D, stated an ambition to use AI to increase R&D productivity by 40 percent by 2030 [>BAY-E06]. The same release describes "integrating AI platform architecture with anonymized patient-centric data" so that Bayer's data scientists can validate AI models and scale solutions [>BAY-E07]. The 40 percent figure is a target, not a result.
• A posting first dated 28 July 2026 places a Principal Machine Learning Scientist in a Machine Learning Research group working on biological foundation models [>BAY-E08, >R05]. The Head of Machine Learning Research postings describe a team of about 25 still to be built [>BAY-R01, >R02].
• Bayer's Annual Report 2025 says combining the Tavros platform with Vividion's chemoproteomics capabilities "continues to greatly enhance" Vividion's efforts to generate drug targets [>BAY-E12].
• The Cradle release says Bayer selected Cradle from among several vendors following a successful proof of concept [>BAY-E01]. The collaboration itself is described in the future tense.
• We found no adoption or benefit metric for any of these items in the reviewed sources.

Context

Leadership named in the rows: Sai Jasti as head of Data Science & AI in Pharmaceuticals R&D, with no appointment date found [>BAY-E06]; Anastasia Hager, Global Head of Drug Discovery Sciences, quoted in the Cradle release [>BAY-E01]; and Juergen Eckhardt, Head of Business Development and Licensing at Bayer Pharmaceuticals, quoted in the Tavros, VUMC and Iambic releases [>BAY-E12, >E03, >E02]. The Tavros release quotes him saying that Vividion "continues to operate autonomously" under Bayer's arm's-length model [>BAY-E12].

Under borderline clarification C42-2, two in-licensing deals are context, not T1 commitments, because Bayer gains a programme rather than platform or model access. On 8 October 2024 Bayer and MOMA Therapeutics announced an option and licence for an oncology programme from MOMA's platform, which includes "computation-enabled lead optimization" [>BAY-E13]. On 24 October 2024 Bayer licensed a heart disease programme from Dewpoint Therapeutics under a 2019 collaboration, with upfront and milestone payments estimated at $424 million excluding royalties; AI/ML appears only in Dewpoint's description of its platform [>BAY-E14].

Under C42-4 and C42-3, venture items are context. Bayer opened the Co.Lab Berlin incubator on 28 November 2024 and launched Co.Lab AdVenture, to connect incubator start-ups with venture investors, on 8 September 2025; neither states AI scope [>BAY-E19, >E20]. Leaps by Bayer invested in Souffle Therapeutics (2025), SonoThera Therapeutics (June 2026) and Celea Therapeutics (July 2026), none with a stated AI or computational discovery platform [>BAY-E22, >E23, >E24]. A further Leaps investment in Leaf, an agricultural data company, is excluded [>BAY-E25].

Other context: radiology AI collaborations, including data providers for Bayer's radiology AI Innovation Platform (December 2024), an extension of the SubtleGAD imaging collaboration (October 2025) and Systalyze, which an MIT CSAIL article of March 2026 says Bayer plans to make a core technology partner company-wide [>BAY-E15, >E16, >E17, >E18]; a clinical trials alliance with the University of Colorado Anschutz, UCHealth and Children's Hospital Colorado (July 2026) [>BAY-E31]; and an oncology research partnership with Recursion, listed as ongoing in the Annual Report 2025 with no AI scope stated [>BAY-E21]. Enterprise AI: the Q2 2026 board address says Bayer is investing in AI "in both enterprise systems and tools for our people", and the Annual Report 2025 describes MyGenAssist, an in-house AI tool for employees [>BAY-E32, >E35]. Bayer's pharmaceuticals head listed "AI-enabled development" among growth drivers in January 2026, and the Annual Report 2025 lists data science and AI among the measures to improve Pharmaceuticals R&D productivity [>BAY-E33, >E34].

Excluded rows: extensions of the Broad Institute and Tsinghua University research collaborations, neither stating AI or computational work [>BAY-E26, >E27]; the OpenFold Consortium partnership, which is Crop Science [>BAY-E28]; Hurdle biomarker research, which is Consumer Health [>BAY-E29]; and an AWS generative AI prototype for Crop Science process research [>BAY-E30].

We identified no qualifying pause or cancellation in the two themes in the searched sources. The endings found, the Exscientia deal in 2022 and the Mission Bay research hub in 2021, are before both windows, and a Frankfurt site closure is Crop Science (adjudication notes). That is a search result, not evidence that none occurred.

Coverage limits

• www.bayer.com returned a bot-block page to both reviewers for every path outside the newsroom, and pharma.bayer.com redirects there. One reviewer could read the newsroom and enumerated it from its sitemap for both windows; the other could not and worked from PR Newswire, partner and subsidiary sites and Bayer's online reports. The Bayer Pharma AI page could be read only as a generated summary, so its knowledge-graph row is not verified and is not used as evidence [>BAY-E09].
• Investor slide decks and call transcripts were not retrieved, because the bayer.com investor pages were blocked and the document download path is disallowed in robots.txt. Results releases and board speeches on the newsroom, the Q1 2026 quarterly statement and the half-year 2026 report were read. The Q2 2026 call transcript and presentation could still not be obtained by 8 October 2026.
• The Annual Report 2024 was read only as a generated summary or its first 40 pages: the bayer.com PDF was blocked, reports.bayer.com no longer hosts it and annualreports.com disallows the PDF path in robots.txt. The Annual Report 2025 was read in full.
• Business Wire blocked access and GlobeNewswire did not respond. The PR Newswire search reached back only to about October 2025. leaps.bayer.com, Recursion investor relations and Fierce Biotech blocked access, europeanpharmaceuticalreview.com served a verification page, the web archive could not be reached and SEC robots.txt could not be checked.
• Each reviewer ran nine web searches under a capped budget. These were not run: a Leaps by Bayer venture financing search, a FinnGen and PRECISE partnership search (which would also date those partnerships), a Bayer pharma ELN, LIMS or scientific data platform search, and a site-restricted Business Wire and GlobeNewswire search (which would also look for an original TetraScience deployment announcement). Four partnership, acquisition and venture queries ran only as two combined queries, not as written. The T2 counts rest on the newsroom enumeration, the PR Newswire search, the annual and interim reports and one lab automation search per reviewer.
• The Machine Learning Research group row takes its first public date from a structured date-posted field in a posting page [>BAY-E08]. Read as first observed on 6 October 2026, after W1 ends, it would not count as first disclosed in W1, and existing T1 capabilities first disclosed in W1 would be four.
• Patents are not among the searched source families in the coverage table. Publication counts are raw Europe PMC query counts and are not events.
• Recruitment: one reviewer enumerated the talent.bayer.com job data and the other the jobs.bayer.com feed; both title-screened. Only the 36 requisitions classified are in the table. One posting prints a reference code that differs from its system requisition number, and the two were matched on the URL [>BAY-R11].
• Two independent reviewers searched Bayer, and the final rows are their source-verified union. That union contains rows that each reviewer missed, but it does not show that both did not miss the same events. VUMC was found only by the reviewer who could read the newsroom, and the TetraScience expansion only by the reviewer who searched PR Newswire [>BAY-E03, >E10].

Two questions to investigate (Ortent's interpretation)

1. Bayer's scientific data work appears in more than one place. TetraScience says Bayer uses its Scientific Data Foundry for instrument and CRO data across pharmaceutical and crop science R&D, with self-service onboarding for Bayer's digital and IT teams, and the Bayer lead quoted has a regulatory science title [>BAY-E10, >E11]. Separately, the Head of Machine Learning Research is to work with data engineering on the assay, structural and DMTA data foundation and set up MLOps, and a quantum mechanics scientist is to manage high-performance infrastructure with external IT providers [>BAY-R01, >R02, >R03]. No posting names an ELN, LIMS or data platform. What is the remit of the Data Science & AI unit for discovery data systems relative to the digital and IT teams that run the TetraScience deployment? Who decides on additions to the scientific data stack, and how are the interfaces between the Machine Learning Research data foundation and the TetraScience deployment governed, given that the deployment spans the pharmaceutical and crop science divisions?
2. Bayer's external AI-discovery agreements are announced with different executives quoted: the Global Head of Drug Discovery Sciences for Cradle, chosen from among several vendors after a proof of concept, and the head of business development and licensing for Iambic, VUMC and Tavros [>BAY-E01, >E02, >E03, >E12]. Vividion, which acquired Tavros, is said to operate autonomously [>BAY-E12]. The Head of Machine Learning Research is to make build/buy/partner decisions on design platforms, and machine learning scientists are to find opportunities with internal and external AI capabilities [>BAY-R01, >R02, >R04, >R05]. Who holds decision rights over licensing an external AI discovery platform: Drug Discovery Sciences, business development and licensing, or the Data Science & AI leadership team? Does a platform selection like Cradle's pass through procurement, and how is funding for these decisions set against the stated 40 percent R&D productivity ambition [>BAY-E06]?

Sources for the main claims

• BAY-E12 Vividion acquisition of Tavros (Bayer newsroom): https://www.bayer.com/media/en-us/vividion-therapeutics-acquires-tavros-therapeutics-to-expand-functional-genomics-capabilities-and-boost-drug-discovery-platform/
• BAY-E10, E11 TetraScience and Bayer expanded partnership (PR Newswire): https://www.prnewswire.com/news-releases/tetrascience-and-bayer-expand-partnership-to-advance-scientific-data-management-302620343.html
• BAY-E01 Bayer and Cradle collaboration: https://www.bayer.com/media/en-us/bayer-and-cradle-enter-collaboration-to-enhance-ai-enabled-antibody-discovery-and-optimization/
• BAY-E03 Bayer and Vanderbilt University Medical Center collaboration: https://www.bayer.com/media/en-us/bayer-and-vanderbilt-university-medical-center-to-advance-treatments-for-cardiovascular-and-kidney-diseases/
• BAY-E02 Bayer and Iambic collaboration: https://www.bayer.com/media/en-us/bayer-and-iambic-collaborate-to-advance-drug-discovery-with-ai/
• BAY-E04 to E07 Bayer Pharma Media Day 2026: https://www.bayer.com/media/en-us/bayers-strongest-ever-pharma-portfolio-and-promising-pipeline-drive-future-growth/
• BAY-E12 acquisitions note and BAY-E34 Bayer Annual Report 2025: https://reports.bayer.com/annual-report-2025/en/_assets/downloads/entire-bayer-ar25.pdf
• BAY-E13 MOMA Therapeutics: https://www.bayer.com/media/en-us/bayer-and-moma-therapeutics-enter-collaboration-and-license-agreement-in-oncology/
• BAY-E14 Dewpoint Therapeutics: https://www.bayer.com/media/en-us/bayer-and-dewpoint-therapeutics-sign-licensing-agreement-for-heart-disease-program/
• Other event rows and all requisitions: URLs and archived copies in the Bayer tables. No SRC source files are cited for Bayer.

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.