Merck KGaA, Darmstadt, Germany

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), in the Healthcare business sector only. The Life Science business sector (MilliporeSigma), which sells lab tools and software to others, is out of scope: its product launches are excluded and its deals count only where Healthcare is the user. This company is not Merck & Co (MSD), and no MSD material is used. 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 one T1 initiative with an AI-discovery commitment change involving the Healthcare business sector first disclosed in the year to 5 October 2026, and one in the year before [>MKG-E01, >E04]. The first is a group high-performance computer, launched with Lenovo and Equinix, which the company says is to streamline drug discovery in Healthcare and support AI and machine-learning models; its launch day is not stated [>MKG-E01]. The second is a co-development with imec of a microphysiological systems platform intended to generate training data for AI-driven drug discovery, to be validated in Healthcare laboratories [>MKG-E04]. We identified no commitment change in company lab digitalisation (T2) in either year in the searched sources. The 6 October recruitment snapshot contains 3 relevant R&D requisition postings across the two themes, none with ambiguous vacancy status [>MKG-R15, >R26, >R28]. These observations identify programmes and functions for further investigation; they do not establish procurement opportunity, deployment completion or net hiring.

Announced commitments

Merck KGaA published the HPC release itself and issued the imec release jointly with imec [>MKG-E01, >E04]. Both were read on emdgroup.com, the US and Canada mirror of the group newsroom. The company also mentions the HPC in its Strategic Update deck of 13 May 2026, and Equinix issued a same-day release, which one reviewer read as a copy on finviz [>MKG-E01]. The imec platform also appears in the Q2 2025 interim report and on a company research page [>MKG-E04].

The adjudicator recorded the three group-scope rows as follows. - Group HPC [>MKG-E01]: a new T1 commitment. It is a group asset shared by three sectors. It passes the group test because the release states drug discovery use by Healthcare and AI and machine-learning workloads. It is compute, not a lab or scientific data system, so it is not T2. The release reports the launch as done but gives no launch day, and no bundled source gives an earlier bound. Under rules D1 and D3 the publication date is not used as the event date, so the event date is unknown with a latest bound of 20 November 2025. It counts only in the first-disclosed view, in W1. Both reviewers had used 20 November 2025 as an exact date; under that reading it would also count in W1 by event date. - imec co-development [>MKG-E04]: a new T1 commitment in W0, jointly issued, found by one reviewer only; the other fetched the release and recorded no row. Healthcare counts as a named user because the platform is to be validated in Healthcare laboratories, although Life Science will sell it. T1 rests on the stated in-silico pipeline and the link to AI training data for drug discovery, not on cellular models alone (AR7a). It is not T2, because use in Healthcare workflows is limited to validation. The adjudicator marked it borderline on scope and theme. If validation in Healthcare laboratories is not enough to make Healthcare the user, T1 W0 would be zero in both views. - M-Boltz Hackathon [>MKG-E02]: recorded as exclude on scope, with theme T1 as assessed. The release of 27 October 2025 reports that the company "hosted the M-Boltz Hackathon together with the Boltz team", with one challenge on antibody-antigen complexes and another on allosteric small-molecule binders, evaluated on "an internal dataset provided by Merck KGaA, Darmstadt, Germany". The adjudicator found C42-4 met in substance: a co-hosted challenge with stated discovery scope and company collaboration through that dataset; NVIDIA and Hugging Face sponsored the prizes, and company funding is not stated. C42-6 does not apply, because the company released no model, dataset or tool. The row fails the group test applied to the HPC: it is a Group Digital Innovation activity, and the release never names the Healthcare sector as user. It says only "drug discovery and development processes" and that Boltz is "already integrated into many research workflows at our company", and Life Science also runs drug discovery businesses for customers. The hackathon dates are not stated (latest bound 27 October 2025). Both reviewers had counted it as a new T1 commitment. Under the opposite scope reading it would be a new T1 commitment counted in W1 by first disclosure only, because its event date is uncertain, making that count two. If C42-4 were also read narrowly, so that a one-off hackathon is not a funded challenge, it would be venture context and no count would change.

By event date, we counted one T1 initiative with a qualifying commitment change in W0 (5 October 2024 to 4 October 2025), the imec co-development, and none in W1 (5 October 2025 to 5 October 2026). By first disclosure, the T1 count is one in W0 and one in W1, the group HPC, whose event date is uncertain. T2 is zero in both views and both years. No milestone or negative change was disclosed in either theme. Existing capabilities first disclosed numbered one in T1 in each year: AIDDISON's internal use in Healthcare early discovery in W0 and the Predictive Modeling Platform in W1 [>MKG-E05, >E07]. These have no stated start date and are not counted as changes. The AIDD automated lab and the oligonucleotide computational design group are observed only through job postings first seen on 6 October 2026, after W1 ends, so neither is counted (C42-17) [>MKG-E06, >E16]. Under the opposite reading of borderline clarification C42-12, which counts group items that state drug discovery without naming Healthcare, T1 W1 would be zero by event date and two by first disclosure, adding the hackathon [>MKG-E02]; existing capabilities first disclosed would be two in T1 in each year and one in T2 in W0 [>MKG-E14, >E03, >E05]. The opposite reading of C42-4 alone changes no count, because the hackathon is excluded on scope, and no row turns on C42-2.

Coverage is partial for both themes in both years. Six of the nine source families are complete, but the company newsroom, results transcripts and partner newswire families are partial in every cell and marked not comparable between the years, so no comparison between the years is made.

Nine rows are recorded as exclude: four group items that fail the group test, three Life Science customer deals or products, and two Healthcare items outside the themes (see Context).

Identified recruitment

On 6 October 2026 the careers.emdgroup.com sitemap listed 862 unique job IDs by both reviewers' counts; the search page reported 865 hits. The site lists postings for all three business sectors. A common title filter selected the same 35 requisitions for both reviewers, and one reviewer added two roles outside the filter as plainly in scope [>MKG-R26, >R28]. We classified 37 postings from their duties. We observed 3 relevant requisition postings in the R&D theme class, 3 T1 and 1 T2; the role tagged both appears in each split [>MKG-R26]. No posting shows an openings field, so vacancy status is not recorded for any of the 3 and none has ambiguous vacancy status. A further 19 are enterprise roles, 5 are plant, quality or manufacturing roles, 5 are internships or graduate roles and 5 are out of scope. These are requisition postings listed on one day, not hires, headcount or a trend.

The R&D postings are all in Billerica, Massachusetts: - a Senior Scientist, Oligo Bioinformatics, in the Oligonucleotide Therapeutics Group (DDTech) at EMD Serono, which the posting says "is expanding", to lead computational design and optimisation of siRNA and ASO sequences, "directly shaping therapeutic programs from target identification through lead optimization" (T1) [>MKG-R15, >E16] - a Principal Scientist, Synthetic & Discovery Chemistry, in AI and Automation in Drug Discovery (AIDD), to plan and run synthesis "within the AIDD automated lab", work with automation engineers to make routes robot-executable and keep compound data in ELN and data systems, as part of a "closed-loop discovery engine" (both themes) [>MKG-R26, >E06]. The posting reports to a Head of Science Operations and also names "selected external customer programs" among its duties. - a Principal Scientist, Translational Quantitative Pharmacology, in Clinical Measurement Sciences, to develop QSP models that "will inform compound progression decisions from lead optimization to clinical proof of concept and beyond" (T1) [>MKG-R28]. The adjudicator marked this borderline, because most of the duties are translational and clinical. Read as out of scope, it would make the R&D count 2, with T1 2 and T2 1.

Enterprise postings include four Jerusalem roles in the Advanced Data Science team, which build machine learning and agentic AI across pharmaceutical R&D but name no discovery activity, so they are enterprise roles under C42-21 [>MKG-R09 to >R12]. Others are a cloud data and AI architect for the Healthcare portfolio in Barcelona [>MKG-R33], group AI/ML and agentic AI platform engineers in Bangalore, Barcelona, Portugal and Cairo [>MKG-R18 to >R20, >R22, >R24, >R25, >R29] and a Linux, Unix and HPC manager in Bangalore whose HPC duties are not tied to R&D [>MKG-R31]. An internship on AI-first workflows for an ADC payload team in Darmstadt is outside the R&D count [>MKG-R05]. The five out-of-scope postings are Life Science and Electronics roles: data and AI programme managers at EMD Millipore, OT engineering for biopharmaceutical R&D environments at MilliporeSigma and three Electronics supply-chain data scientists [>MKG-R32, >R06, >R01 to >R03].

No page shows a date-posted field. Dates in the embedded job data may record a refresh, so they are not reliable.

No LinkedIn company insights were collected for Merck KGaA.

Technology mentions in recruitment

Mentions are classified from the exact posting wording. They do not establish installation, a contract or company-wide use. The table records 30 Merck KGaA mentions: 6 explicit use, 6 integration or environment, 1 required experience, 2 preferred experience and 15 alternatives. - Named as existing systems: the AIDD automated lab, recorded as explicit use because the posting states the employer's own lab; no product or vendor is named [>MKG-R26]. In enterprise postings, Amazon EKS for a multi-tenant agentic AI platform [>MKG-R22, >R29], AWS for group AI/ML platform infrastructure ("you will design, deploy, and continuously improve scalable, resilient infrastructure on AWS") [>MKG-R20, >R25] and ServiceNow [>MKG-R36]. This is role-specific wording. - Integration responsibilities: ELN and data systems, with no product named [>MKG-R26]; bioinformatics platforms for siRNA and ASO design [>MKG-R15]. Outside R&D, the Healthcare cloud architect is to "enable governed solutions across AWS, Snowflake, Palantir Foundry, and related technologies" [>MKG-R33], and the HPC manager covers enterprise high-performance computing [>MKG-R31]. - Required experience: AWS services, for an EMD Serono data and MLOps engineer [>MKG-R08]. - Preferred experience: AgentCore and Amazon Bedrock [>MKG-R22, >R29]. - Lists of examples: BLAST, Ensembl, RNAfold and siRNA design tools [>MKG-R15]; Enamine, ChemDiv and Mcule building block databases [>MKG-R26]; Snowflake, SageMaker, LiteLLM, Azure OpenAI and Palantir Foundry in enterprise postings [>MKG-R08, >R18, >R21, >R24, >R33].

No ELN or LIMS product is named in the R&D postings, and the only category recorded as explicit use is the AIDD automated lab.

Deployment evidence

• The HPC release reports the computer as launched, with Lenovo ThinkSystem servers in an Equinix data centre in Munich and a hybrid of private and public cloud [>MKG-E01]. It is a group asset for all three sectors, and the release does not say how much of it Healthcare uses.
• The Annual Report 2024 says of AIDDISON, a Life Science product launched in 2023: "In addition to external commercialization, we also use it internally in our Healthcare business sector in early stages of drug discovery" [>MKG-E05]. The same passage says "We are continuing to build our automated design-make-test-analyze platform powered by lab automation", but names no sector as its user, so that platform is not recorded in T2. The Annual Report 2025 repeats the passage. The report's first-public date rests on the FY2024 results date and is uncertain.
• A company release of 21 September 2026 names the "Predictive Modeling Platform", "a tool for AI-driven design of drug-like molecules for the Healthcare business sector, recently recognized with an in-house award" [>MKG-E07]. In the same release, the Head of Discovery & Development Technologies in Healthcare R&D says the RDKit toolkit helps Healthcare R&D scientists "turn chemical data into AI-ready representations". Start dates are not stated.
• A posting first seen on 6 October 2026 describes "our Boston-area autonomous make-and-test lab" in AIDD, combining automated synthesis, analytical characterisation and biological readout [>MKG-E06, >R26]. Its start date is not stated. The posting carries Healthcare boilerplate but also mentions external customer programmes, and whether the unit is purely Healthcare is not settled.
• We found no adoption or benefit metric for any of these items in the reviewed sources.

Context

Leadership: on 2 September 2025 the company announced the appointment of David Weinreich as Global Head of R&D and Chief Medical Officer for the Healthcare business sector. The release states no AI, computation or lab-data remit, so the row is recorded as exclude [>MKG-E12]. Arvind Rajpal is quoted as Head of Discovery & Development Technologies at Research and Development within the Healthcare business sector in the September 2026 release, with no appointment date [>MKG-E07]. Laura Matz is quoted as Chief Science and Technology Officer in the November 2025 HPC release [>MKG-E01]. Philipp Harbach is quoted as Global Head of Group Digital Innovation in the October 2025 hackathon release and as Chief Science and Technology Officer in September 2026; no appointment date was found, and the group role has no stated discovery remit, so it is context (C42-7) [>MKG-E02, >E22].

Group governance and enterprise AI, all context with no discovery scope stated: a newly established Data, Digital & IT Executive Council under the joint leadership of the Group Chief Science & Technology Officer and the Group Chief Information Officer (Annual Report 2024) [>MKG-E08]; UPTIMIZE, a group data and AI operating model [>MKG-E17]; Digital Enterprise Solutions, which the Annual Report 2025 says the company formed as "a single Group function that unites Information Technology, the Group Data & AI Organization and Global Enterprise Solutions" [>MKG-E19]; and the Strategic Update of 13 May 2026, which says "We combine expertise in biology, chemistry and physics with data and AI" [>MKG-E25]. The four Jerusalem Advanced Data Science postings are recorded as enterprise AI context [>MKG-E26].

Group items excluded on scope, because no source names Healthcare as user: the in-house protein engineering suite that the hackathon release says uses Boltz, placed with the group Science & Technology Office [>MKG-E03]; "smart" lipid nanoparticles developed with AI across sectors [>MKG-E14]; and Iktos, an AI drug-design company listed in the M Ventures Life Science portfolio with no investment date [>MKG-E15].

Life Science deals and products, recorded as exclude: a memorandum of understanding with Siemens of 23 September 2025 on software, automation, data and AI for customers, whose page title refers to AI drug discovery [>MKG-E10]; a multi-year agreement of 23 January 2025 to automate assay kits on an Opentrons workstation [>MKG-E11]; and the ChemiSphere customer app of 16 December 2025 [>MKG-E27].

Other context: a research collaboration with Saturnus Bio on rare genetic cardiomyopathies, announced on 23 June 2026 with $50 million upfront, states no AI or computational discovery work and is recorded as exclude [>MKG-E28]. The Annual Report 2024 describes a Bio-Convergence project developing models that combine AI with human-derived cells and tissues to improve the translatability of drug testing; it is context and is not linked to the imec deal, because the report does not name imec [>MKG-E18]. M Ventures led a GBP 6.6 million round in Abselion in December 2024 and co-led a EUR 33 million round in Cyllene Therapeutics in July 2026; neither investee states an AI discovery platform [>MKG-E23, >E24]. An Innovation Cup challenge for AI agents offered a 10,000 EUR prize with registration in June 2026; its challenge topics are not stated as drug discovery and the page is undated [>MKG-E09]. The company gave its Heinrich Emanuel Merck Award for computational sciences to Alán Aspuru-Guzik in June 2025 [>MKG-E20], and hosted the RDKit user group meeting in September 2026 [>MKG-E21].

We identified no qualifying pause or cancellation in the two themes in the searched sources. Both reviewers ran a negative-event web search for each theme. That is a search result, not evidence that none occurred.

Coverage limits

• Complete families: annual reports, consortium announcements, academic collaborations, the corporate venture arm, one web search per event type and one negative-event search per theme are checklist complete for both themes and both years.
• Partial families, in every cell and not comparable between the years: the company newsroom, results presentations and transcripts, and partner newswires.
• merckgroup.com redirects every path to the US and Canada mirror emdgroup.com, and the global newsroom archive returned 404 there. Releases published only on the global site, including Europe-only or German-only releases, may be missing. careers.merckgroup.com also redirects to careers.emdgroup.com, whose sitemap lists global postings.
• Call transcripts were not published or located on the investor site. One reviewer read no quarterly report other than the H1 2026 report.
• Business Wire blocked access and GlobeNewswire was unreachable. Fierce Biotech, StreetInsider, BioWorld, merckmillipore.com, ucl.ac.uk, the Recursion investor site, Pharmaceutical Technology, R&D World and mwe.com also blocked access, and the IHI site refused its robots.txt. endpts.com disallows Claude agents in robots.txt. jobs.merckgroup.com and ligand-ai.eu were refused by the proxy, and bioalps.org and trysignalbase.com reset the connection.
• Searches not run (search budget): a newswire search naming Business Wire, GlobeNewswire and PR Newswire; two follow-ups on the 2023 BenevolentAI and Exscientia (now Recursion) collaborations and Iktos; a follow-up on Boltz, the MIT Jameel Clinic and the Acceleration Consortium; and date checks for the hackathon and for the HPC launch. The two date checks could move the HPC, and the hackathon if its scope were reversed, into the event-date view [>MKG-E01, >E02]. Neither reviewer could check the 2023 BenevolentAI and Exscientia collaborations for milestones or terminations in either window.
• A media report of a cut of about 20 roles at the Billerica R&D site in August 2026 names no AI, computational or lab-data function and is not verified [>MKG-E13]. Fierce Biotech blocked access and only a generated summary was saved. An aggregator claim that the company acquired Valo Health is not verified; it has no bundled source and neither reviewer found a company source, so it is not in the tables.
• Annual-report first-public dates rest on results dates and are uncertain [>MKG-E05, >E08, >E14, >E17, >E18, >E19]. Rows drawn from job postings take 6 October 2026 as their first public date, because no page shows a date-posted field [>MKG-E06, >E16, >E26].
• Patents are not among the searched source families. Publication counts are raw Europe PMC query counts, are not events, and cannot separate Healthcare authors from Life Science authors.
• Recruitment: both reviewers enumerated 862 job IDs and title-screened them; only the 37 postings classified are in the table. One reviewer opened twelve R&D bench titles outside the filter and recorded no rows for them.
• Two independent reviewers searched Merck KGaA, and the final rows are their source-verified union. Both worked from the same newsroom mirror, annual reports and careers sitemap, so the union does not show that both did not miss the same events. The imec co-development and the AIDD lab posting were found by one reviewer only [>MKG-E04, >E06].

Two questions to investigate (Ortent's interpretation)

1. AI and automation for discovery appear in several units. The AIDD unit runs an automated make-and-test lab in the Boston area, its chemistry role reports to a Head of Science Operations, and the posting names external customer programmes [>MKG-R26, >E06]. Healthcare R&D has a Discovery & Development Technologies function, whose head is quoted in the release that names the Predictive Modeling Platform, and an Oligonucleotide Therapeutics Group in DDTech [>MKG-E07, >R15, >E16]. The group Science & Technology Office developed the Protein Design Platform of protein AI tools, and AIDDISON is a Life Science product used internally in Healthcare [>MKG-E07, >E03, >E05]. What is the remit of each unit for discovery AI, lab automation and lab data in Healthcare? Who decides on ELN and data systems for the automated lab, and how are the interfaces between Healthcare R&D, the group Science & Technology Office and Life Science products governed?
2. The HPC is a group asset for three sectors, launched with Lenovo and Equinix on a hybrid cloud, and the release quotes the Chief Science and Technology Officer (CSTO) [>MKG-E01]. The imec platform is co-developed for validation in Healthcare laboratories and sale by Life Science [>MKG-E04]. Group data and AI governance sits with an executive council jointly led by the Group CSTO and Group CIO, and with Digital Enterprise Solutions [>MKG-E08, >E19], while postings for the group AI/ML platform describe its infrastructure on AWS [>MKG-R20, >R25]. Who holds decision rights and budget for discovery compute and AI tools used by Healthcare R&D: the Healthcare sector, the CSTO's office or Digital Enterprise Solutions? How is the cost of shared assets such as the HPC set for Healthcare, and does buying an external AI discovery tool pass through group IT procurement, Healthcare R&D or another route?

Sources for the main claims

• MKG-E04 Merck KGaA and imec, MPS co-development (emdgroup.com copy): https://www.emdgroup.com/en/news/mps-with-imec-20-05-2025.html
• MKG-E01 Group high-performance computer with Lenovo and Equinix (emdgroup.com copy): https://www.emdgroup.com/en/news/one-hpc-11-20-2025.html
• MKG-E02, E03 M-Boltz Hackathon (emdgroup.com copy): https://www.emdgroup.com/en/news/protein-folding-hackathon-27-10-2025.html
• MKG-E05, E08, E14, E17, E18 Annual Report 2024: https://www.emdgroup.com/content/dam/web/corporate/non-images/investors/reports-and-financials/earnings-materials/2024-q4/us/2024-Q4-Report-NA.pdf
• MKG-E07, E21, E22 RDKit conference release: https://www.emdgroup.com/en/news/rdkit-21-09-2026.html
• MKG-E06, R26 AIDD posting 303648: https://careers.emdgroup.com/us/en/job/303648/Principal-Scientist-Synthetic-Discovery-Chemistry
• Other event rows and all requisitions: URLs and archived copies in the Merck KGaA tables. No SRC source files are cited for Merck KGaA.

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.