GSK

GSK plc is a London-headquartered pharmaceutical company (NYSE: GSK), best known for the Shingrix and Arexvy vaccines, the respiratory medicine Trelegy and the HIV medicines of its majority-owned ViiV Healthcare business. It was formed in 2000 from the merger of Glaxo Wellcome and SmithKline Beecham, separated its consumer health arm as Haleon in 2022, and has been led since January 2026 by Chief Executive Luke Miels. It appears here not as an AI vendor but as a large pharmaceutical AI adopter, and it does not sell an AI product.

GSK's defining pattern is data-first. Its research strategy rests on the bet that human genetics and functional genomics, read with machine learning, pick better drug targets, and it runs a dedicated in-house artificial intelligence and machine learning group, led by Kim Branson, to do that reading. In November 2021 GSK said more than 70 percent of its early-stage programs had genetic support, and that its goal was to double development success rates. Its outside deals mostly buy the raw material for that group, meaning patient data, lab-generated cell data and licensed biology models, rather than handing discovery to a partner.

The deals follow that shape. In October 2022 GSK paid a 70 million dollar initial payment for a three-year collaboration with Tempus AI, gaining access to its AI-enabled platform and library of de-identified patient data to improve clinical trial design, speed enrollment and identify targets, building on a relationship begun in 2020. In December 2024 it signed two collaborations with the London biotech Relation in fibrotic diseases and osteoarthritis, paying 45 million dollars upfront, including a 15 million dollar equity investment, for machine learning on single-cell data from human tissue. In July 2026 Relation expanded the work into immunology, metabolic and bone disease, for up to 110 million dollars, generating large-scale cell perturbation data to feed its Morgan foundation model.

In January 2026 GSK took a five-year, non-exclusive license to Noetik's virtual cell foundation models for non-small cell lung cancer and colorectal cancer, with 50 million dollars in upfront and near-term milestone payments plus annual licensing fees, treating a biology model as a subscription rather than a service. In July 2026 it signed with Boltz, whose scientists will fine-tune Boltz's biomolecular models on GSK's proprietary data to build exclusive models for its discovery pipeline, with GSK keeping ownership of the compounds and no terms disclosed. In November 2025 it committed 45 million pounds to six antimicrobial resistance programs with the Fleming Initiative and Imperial College London, several of them built on AI-driven discovery. Its September 2025 pledge of 30 billion dollars in United States research and manufacturing included a 1.2 billion dollar tranche that bundles a new biologics factory with AI and digital technology at its five existing American plants, with no separate figure for the AI.

The honest framing: GSK's value rests on its vaccines, its HIV franchise and its specialty medicines, and AI is a lever on target selection and trial design on top of them. Most of what is public is spending on inputs, data access and model licenses, and GSK has not attributed an approved medicine to them; doubling success rates is a stated goal, not a measured result. Among peers, Gilead procures specialists to do discovery work, Eli Lilly productized its own models through TuneLab, Moderna built hundreds of models in-house, Merck & Co. standardized on one enterprise platform with Google Cloud and Novo Nordisk spreads work across OpenAI, Amazon Web Services and Anthropic, while GSK concentrates its spend on buying biology data and licensing models for its own team to use.