Ortent Life Sciences Radar: method (research beta)
The radar compares what 25 large pharma and biotech R&D spenders say in public about two things: AI and computation in drug discovery, and the digitalisation of their research labs and scientific data. Each company gets two scores for each theme. Priority is how much weight the company puts on the theme in its own disclosures, on a scale of 1 to 10. Delivery is how far its published evidence shows the work running, from 0 to 10. Every score opens the quotes and links behind it, and each company also has a full account brief.
The scores come only from public information up to 6 October 2026. They are Ortent's assessment and have not been independently validated. A low score can mean a company says little in public, or that we could not reach some of its sources, so the coverage status sits next to every score. If you know of public information we missed, you can report it on the radar page and we will review it at the next quarterly refresh. The radar does not predict who will buy, and it does not measure how mature a company's operations are.
The companies
The panel is 25 large biopharma groups, chosen by the R&D expense each group reported for its latest complete financial year (FY2025). We describe them as 25 large R&D spenders, not the top 25. Rankings move depending on whether acquired in-process R&D and deal payments are counted. On one basis Biogen, UCB and Merck KGaA Healthcare fall outside the top 25; on the other, Astellas, UCB and Merck KGaA Healthcare do. BioNTech, Otsuka and BeOne Medicines rank inside on both bases and are not covered.
Groups with large non-pharma businesses are covered for the pharma business only: J&J Innovative Medicine, Roche Pharmaceuticals, Bayer Pharmaceuticals and Merck KGaA Healthcare. Separately listed subsidiaries with their own R&D, such as Chugai, are out of scope.
The 25 are AbbVie, Amgen, Astellas, AstraZeneca, Bayer, Biogen, Boehringer Ingelheim, Bristol Myers Squibb, Daiichi Sankyo, Eli Lilly, Gilead, GSK, Johnson & Johnson, Merck KGaA, Moderna, MSD, Novartis, Novo Nordisk, Pfizer, Regeneron, Roche, Sanofi, Takeda, UCB and Vertex.
The two themes
T1 is AI and computation in drug discovery: models, platforms, compute, data acquired to find targets, partnerships and the people who lead the work.
T2 is digitalisation in research labs: lab automation, lab informatics such as electronic lab notebooks and LIMS, and scientific data foundations.
For this radar, process and CMC development are grouped with manufacturing and excluded from the research-lab theme. That covers process chemistry, scale-up, pilot labs and technical development centres, even where the work happens in a lab. This is a choice about scope. It is not a judgement that this work is unimportant or not digital, and the briefs show examples of it as context where they help.
Where one programme spans both themes, such as a design, make and test loop, the automated lab work counts in T2 and the AI that designs molecules or reads the results counts in T1. Clinical development, manufacturing, commercial and company-wide AI programmes are recorded as context, not counted.
Sources and dates
Evidence runs to 6 October 2026. Each brief covers two windows: 5 October 2024 to 4 October 2025, and 5 October 2025 to 5 October 2026.
We use company releases, filings, annual and results materials, company stories and web pages, partner and vendor releases, consortium pages, trade press, and each company's own job site. Every row in the underlying tables cites a stored copy of its source with a SHA-256 hash. A row is used in a brief only when its quoted passage is found in that copy. Each ID in brackets in a brief links to the online source behind that row, and a source index at the end of each brief lists them all. We do not republish our stored copies, because they are other organisations' documents. Job posting links come from the 6 October snapshot, and some postings will have closed. Where a company site blocked access, the link goes to the copy we read.
We never invent a date. Where a source gives no date for an event, the brief says so. A publication date is never used as the date an event happened.
Briefs follow the tense of the source. "Will deploy" is reported as an agreement or an intention, never as a deployment.
The scores
Priority runs from 1 to 10. A passing mention scores 1 or 2. A theme described in a strategy or R&D section scores 3 or 4. A theme with a named owner, an investment or a dated target scores 5 or 6, and so does a headline company priority tied to the theme. Scores of 7 and above need a headline priority with a specific investment or target; 9 and 10 also need it among the company's top three priorities, quantified and repeated across different kinds of document.
Delivery runs from 0 to 10 and follows fixed steps. 1 and 2 are a plan or a funded commitment. 3 is a build or pilot. 4 is a named activity the company says is running. 5 and 6 add a measure of how widely it is used. 7 and 8 need a measured result, with a comparator and a period, that the company ties to the work. 9 and 10 need that result sustained across reporting periods. Where a company says it uses something but names no system or shows no evidence it is running, Delivery shows as "claimed" with no number.
The gap is Priority minus Delivery. A large gap means a company talks about the theme more than it shows evidence of delivering it, at least in public.
Each company and theme has one of three states. Evidenced means we found eligible evidence and scored it. Provisional means a published company document could not be read in full, so the scores may change; Bayer is in this state for both themes. Where we found nothing eligible, the radar says "No eligible company disclosure identified for this theme in the reviewed sources and period." That is not a finding that the company is inactive.
Only evidence dated inside the window counts. Job adverts, vendor marketing, undated web pages, third-party press and private knowledge never count towards a score. They can still appear in a brief as context.
Reading a comparison
Companies differ in how much they publish and in how open their sites are to research. Some block automated access; some host results calls only as video. So a comparison shows who has said and shown the most in public about each theme. It is a good place to start looking, and a poor basis for concluding that a company is doing nothing. Before acting on any score, open the quotes behind it and read the coverage limits in that company's brief.
What the counts in the briefs mean
Counts in the briefs are initiatives identified in the reviewed sources. They are reported two ways: changes known to have happened in each window, and commitments first disclosed in each window. A count of zero means none was identified in the sources searched. Each count sits next to the coverage limits for that company.
Recruitment figures come from one snapshot of each company's job site on 6 October 2026. They are postings listed on that day. They are not hires, headcount or a trend.
The activity counts are not scores and are not used to compare companies, because they depend heavily on how much of each company's output we could search.
Coverage limits
Some sites blocked automated access, some documents could only be read in part, and search budgets ran out for some companies. Each brief lists these limits for its company. Bayer is the clearest case: its Q2 2026 call transcript and presentation could not be obtained, and the Bayer brief says so.
Vendor and software company names appear in a brief only where a public source names them. Source links stay available on every claim.
Review and status
Ortent built and reviewed this research. That is Ortent checking its own work. It is not independent validation, and nothing on this page or in the briefs should be read as a claim that it has been validated. A second, independent scoring pass and an independent domain review are still to be done.
The rules changed during the work. The latest change narrowed the research-lab theme to exclude process and CMC development. Where a brief shows less T2 activity than an earlier draft, that is a change in method. It does not mean the company stopped the work.
Reporting new information
Anyone can report a public source we missed or a correction, using the form on the radar page. We only accept information anyone can read. Nothing is published automatically. Each report is checked against the same rules as the rest of the evidence and, if it holds up, is included in the next quarterly refresh. Companies can use the same form to point us to their own disclosures.
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 the briefs only where public sources name them.
Version
Research beta, 8 October 2026. Activity rule set v0.4.3. Scoring rubric v0.5.1. Files for this release carry the prefix LSBR, from the radar's working name.