Boehringer Buys AI Hope

von Adrian Kempf | 27.09.2026 | English

Artificial intelligence is increasingly permeating drug discovery: algorithms like Owkin's K Pro analyze molecular and genetic data to identify new targets in oncology and immunology.

Boehringer Ingelheim Licenses Owkin's AI "Scientist" K Pro – What's Behind the Deal

Boehringer Ingelheim announced a licensing agreement with the French AI company Owkin in early September 2026. At the center is K Pro, a so-called "agentic AI Scientist" (an AI that doesn't just analyze data but independently generates hypotheses and plans analysis steps) for oncology and immunology research. I took a closer look at the deal – and at how realistic the expectations for AI-assisted drug discovery actually are right now.

What Boehringer Is Actually Licensing

Owkin is transferring three components to Boehringer Ingelheim: the K Pro platform itself, existing multimodal oncology datasets, and newly generated immunology data. K Pro allows researchers to analyze medical data through natural-language queries, without requiring programming skills. The technology was first presented in 2025 at the ESMO congress (the European conference for medical oncology).

This followed a 2025 pilot project based on Owkin's MOSAIC dataset – a collection of tissue samples built at a cost of roughly 50 million US dollars. According to Owkin, the pilot delivered "spatial insights into the tumor microenvironment" for a specific gene target. The newly signed agreement is intended to accelerate the discovery process across multiple indications in oncology and immunology. Financial details – contract value, milestone payments, duration – were not disclosed by either party.

Boehringer is not the first major customer: AstraZeneca committed to a three-year K Pro deal in May 2026, and Sanofi to a five-year alliance in June 2026. Owkin is positioning itself deliberately as an AI infrastructure provider to several Big Pharma companies at once.

The Context the Press Releases Don't Mention

What's missing from Boehringer's and Owkin's announcements is the company's own backstory. In 2025, Owkin underwent a fundamental strategic shift: away from long-term, project-based pharma contracts and toward a scalable SaaS model (Software-as-a-Service, meaning recurring license revenue instead of individual projects) built around agent-based AI products like K Pro. The stated goal: 400 million US dollars in revenue within five to six years. Reports suggest the transition was turbulent internally, with criticism of poor communication, high turnover, and multiple strategy changes in a short period. In the process, Owkin also spun off parts of the company, including Bioptimus, for specialized biomedical AI models.

That doesn't mean the technology doesn't work. But it does mean that Boehringer is buying into a still relatively young product approach from a company that has just reinvented itself – and whose proof of K Pro's success at industrial scale is still pending.

The Bigger Question: What Has AI Actually Delivered in Drug Discovery So Far?

The Boehringer-Owkin deal fits into an industry-wide trend that increasingly invites critical scrutiny. In 2024 alone, AI-driven drug discovery companies raised 8.9 billion US dollars across 264 financing rounds, according to market data. A study published in the journal Pharmaceuticals puts total investment in AI health technologies between 2022 and 2026 at over 100 billion US dollars.

These sums stand against a sobering result: as of August 2026, not a single fully AI-discovered drug has received full FDA approval. According to reports, forecasts for the first approval have been pushed back continuously for three years running.

Concrete examples show where the limits lie: BenevolentAI's candidate BEN-2293 for dermatitis failed both primary endpoints in a Phase IIa trial. Insilico Medicine's candidate rentosertib did reach Phase IIa and orphan drug status for a rare lung disease, but a final approval success is still pending. The cited study sums it up this way: AI demonstrably shortens and cuts the cost of early phases such as target identification and lead optimization (by an estimated 30 to 70 percent), but in clinical testing "the algorithm's advantage erodes" – approval probability remains at 8 to 12 percent, in line with conventional drug development.

On top of this comes a methodological issue the authors call a "validation crisis": only about 6 percent of the underlying AI publications make their source code available, and roughly a third share the datasets used. This makes independent verification of results considerably harder – a point that also remains relevant for assessing partnerships like the one between Boehringer and Owkin, as long as efficacy data from such collaborations isn't published in externally validated form.

The market itself already corrected in 2024: BenevolentAI, Exscientia, and Recursion, three prominent AI drug discovery companies, saw significant valuation losses; Recursion subsequently acquired Exscientia – a signal of emerging consolidation after initially inflated expectations.

Assessment

The deal between Boehringer Ingelheim and Owkin makes strategic sense: anyone wanting to keep pace in oncology and immunology research can hardly avoid AI-assisted analysis tools, and Boehringer joins AstraZeneca and Sanofi in a growing group of Owkin customers. At the same time, the broader body of research shows that the leap from faster early-stage analysis to actually approved drugs hasn't been made yet – by any provider, not just Owkin. Whether K Pro delivers more for Boehringer than comparable systems do for its competitors will only be shown by concrete, externally verifiable trial results – not by the announcement itself.

Sources

Owkin: "Owkin to License K Pro AI Scientist and Multimodal Oncology and Immunology Data to Boehringer Ingelheim," BioSpace, September 2026, https://www.biospace.com/press-releases/owkin-to-license-k-pro-ai-scientist-and-multimodal-oncology-and-immunology-data-to-boehringer-ingelheim
"Boehringer signs AI drug discovery deal with Owkin," pharmaphorum, September 2026, https://pharmaphorum.com/news/boehringer-signs-ai-drug-discovery-deal-owkin
"Owkin Signs AI Deal With Boehringer Ingelheim for Oncology, Immunology Research," Bloomberg, September 2, 2026, https://www.bloomberg.com/news/articles/2026-09-02/boehringer-signs-deal-for-owkin-ai-joining-astra-and-sanofi
"Boehringer Ingelheim Licenses Owkin's K Pro AI Scientist and Patient Data," unite.ai, September 2026, https://www.unite.ai/boehringer-ingelheim-licenses-owkins-k-pro-ai-scientist-and-patient-data/
"Owkin signs AI drug discovery deal with Boehringer for cancer, immunology," The Star, September 2, 2026, https://www.thestar.com.my/tech/tech-news/2026/09/02/owkin-signs-ai-drug-discovery-deal-with-boehringer-for-cancer-immunology
"Owkin K Pro and Boehringer AI Drug Discovery," Briefs, September 2026, https://www.briefs.co/news/boehringer-ingelheim-taps-owkin-s-k-pro-to-hunt-cancer-and-i/
Owkin: "Owkin Launches K Pro: the First Agentic AI Co-Pilot for Biopharma Powered by Biological Reasoning Models," company announcement, https://www.owkin.com/newsfeed/owkin-launches-k-pro-the-first-agentic-ai-co-pilot-for-biopharma-powered-by-biological-reasoning-models
"Q&A: Owkin's five-year Sanofi deal bets on 'purpose-built' AI agents," R&D World, June 2026, https://www.rdworldonline.com/qa-owkins-five-year-sanofi-deal-bets-on-purpose-built-ai-agents/
"An 'AI scientist' can tackle drug R&D. What does that mean...," PharmaVoice, https://www.pharmavoice.com/news/ai-pharma-scientist-drug-research-owkin/818113/
"Startup News: Lessons, Mistakes, and Benefits from Owkin's Pivot to AI Agents in 2025," femaleswitch.com, https://www.femaleswitch.com/startup-news/tpost/3hsomno4x1-startup-news-lessons-mistakes-and-benefi
"AI Drug Discovery Has $8.9 Billion in Hype and Zero FDA Approvals: When Does the Bill Come Due?," Clinical Trial Vanguard, https://www.clinicaltrialvanguard.com/opinion/ai-drug-discovery-has-8-9-billion-in-hype-and-zero-fda-approvals-when-does-the-bill-come-due/
"AI in Drug Discovery: Clinical Failures, Regulatory Reality, and the Validation Crisis Behind the Hype," Pharmaceuticals (MDPI), Vol. 19, No. 6, Article 916, https://www.mdpi.com/1424-8247/19/6/916
"AI Drug Discovery FDA Approvals: The 2026 Reality Check," IntuitionLabs, https://intuitionlabs.ai/articles/ai-drug-discovery-fda-approvals
"AI-Discovered Drugs in Clinical Trials 2026: Full Pipeline," IntuitionLabs, https://intuitionlabs.ai/articles/ai-discovered-drugs-clinical-trials-2026
"AI in drug discovery: predictions for 2026," Drug Target Review, https://www.drugtargetreview.com/ai-in-drug-discovery-predictions-for-2026/1865962.article
"Boehringer wächst zweistellig" ("Boehringer grows in double digits"), transkript.de, 2026, https://transkript.de/artikel/2026/boehringer-waechst-zweistellig/
Owkin, company profile, Wikipedia, https://en.wikipedia.org/wiki/Owkin

Adrian Kempf
Pharma Communications Writer, Graphic & Media Designer
Kirchzarten im Dreisamtal, Germany

Usage & Citation Permission

Full or partial reproduction of this text is permitted with source attribution. Please cite as: "Adrian Kempf, pharmavisuell.de"

This text may be forwarded, published, shared on social media, or posted publicly.

Editorial abridgment is permitted as long as the meaning and message of the text are not distorted.

This usage permission applies to the specialist articles published on pharmavisuell.de. Individually commissioned texts are subject to separately agreed usage rights.