Pfizer

Pfizer is one of the world's largest drugmakers, headquartered in New York City (NYSE: PFE) and founded in Brooklyn in 1849 by Charles Pfizer and Charles Erhart. Its best-known products include the Comirnaty COVID-19 vaccine developed with BioNTech, the Paxlovid antiviral, the Prevnar pneumococcal vaccines, the blood thinner Eliquis co-marketed with Bristol Myers Squibb, and tafamidis, sold as Vyndaqel and Vyndamax, for transthyretin amyloid cardiomyopathy (ATTR-CM). It appears here not as an AI vendor but as a large pharmaceutical AI adopter, and it does not sell an AI product.

Its defining pattern is license-and-customize. Rather than building its own foundation models, as Eli Lilly and Moderna have, Pfizer licenses frontier biomolecular models from specialist labs and pays those labs to retrain them on Pfizer's own historical experimental data, producing private versions no competitor can use. What Pfizer contributes is decades of assay results, not model architecture. The habit predates the current wave: a collaboration with CytoReason, which builds computational disease models, was expanded in September 2022 with a 20 million dollar equity investment and up to 110 million dollars over five years. In June 2025, XtalPi expanded its work with Pfizer on a platform pairing physics-based simulation with machine learning to predict how strongly small molecules bind.

Two 2026 deals made the pattern explicit. In January 2026, Pfizer signed a strategic collaboration with Boltz, a public benefit corporation formed by the MIT team behind the open-source Boltz models. Boltz will refine its Boltz-2 and BoltzGen models on Pfizer's historical data into exclusive models for structure prediction, small-molecule binding affinity and biologics design, and Pfizer keeps full ownership of any compound discovered with them. In June 2026, Pfizer licensed the platform of Chai Discovery, gaining early access to the Chai-3 antibody-design model plus a custom model trained on Pfizer's proprietary data. Neither deal disclosed financial terms. Holding private models from two competing labs is itself a hedge.

Inside the company, Pfizer has a Chief AI and Analytics Officer, Berta Rodriguez Hervas, and ran its second annual internal AI Festival in October 2025, with 54 sessions across seven countries. It has made AI fluency a development goal for all colleagues. The company says AI has cut the time to a first-draft clinical study report by 40 percent, and that it uses machine learning to predict which patients may become eligible for breast cancer trials, with a stated goal of raising trial enrollment by more than 20 percent.

The honest framing: Pfizer's value rests on its marketed medicines, its vaccines and its late-stage pipeline, and AI is a lever on top of that, not the product. None of the licensed discovery models has yet been credited with a disclosed clinical candidate, the operational figures are the company's own claims, and the enrollment number is a target rather than a measured result. Its support for algorithms that flag likely ATTR-CM patients in health records has an obvious commercial side too, since its own drug treats that disease. Among catalog peers, Gilead procures specialists' services, Eli Lilly productized its own models through TuneLab, Moderna built hundreds of models in-house, Merck & Co. runs one enterprise AI platform with Google Cloud, and Novo Nordisk spreads its work across OpenAI, AWS and Anthropic. Pfizer's distinct bet is that its data, trained into other labs' models, is the asset worth owning.