Amgen chief executive officer Bob Bradway is steering the biotechnology firm toward a future defined by the synthesis of high-speed artificial intelligence and decades of specialized molecular research. As of June 3, 2026, the company is actively deploying advanced computational models to accelerate the selection of drug candidates for clinical development by approximately 50 percent.
The strategic foundation for this shift dates back to 2012, when Amgen acquired the Icelandic firm deCODE Genetics. This acquisition provided access to extensive longitudinal genomic data, which Bradway identified as the essential fuel for future machine learning applications in medicine. Despite initial skepticism from peers and industry observers regarding the tractability of such complex genetic datasets, the company persisted in building its own internal infrastructure.
Current operations at the company now include a dedicated Nvidia SuperPod located in Reykjavik to support intensive computational workloads. Researchers have moved beyond off-the-shelf solutions, developing proprietary protein folding models and pursuing zero-shot antibody design. This methodology allows the firm to engineer drug molecules without relying on prior experimental examples, marking a departure from traditional trial-and-error laboratory processes.
Bradway characterizes the ideal leader in this environment as an autodidact, someone comfortable with the rapid iteration cycles inherent to software development. He requires his senior executive team to complete formal AI training to ensure fluency across the organization. This focus on internal capability building aims to prevent over-reliance on external third-party models, allowing the firm to tailor computational tools to specific biological requirements.
The integration of agentic AI represents the next major milestone for the company in 2026. Bradway anticipates that these autonomous systems will handle administrative burdens such as data summarization and material requisitioning. By offloading these tasks, the company intends to preserve the specialized focus of its scientists, ensuring their expertise remains centered on complex molecular analysis.
The firm is also refining its approach to model selection by avoiding reliance on any single platform. Instead of defaulting to widely used tools like AlphaFold, the team identifies the specific limitations of existing models and builds custom layers on top of them. This iterative process ensures that the computational output aligns with the rigorous standards required for pharmaceutical development.
Reconciling the demand for agile, self-taught generalists with the necessity for deep scientific expertise remains a primary leadership challenge. Bradway acknowledges that the workforce experiences anxiety regarding the unknown, particularly as AI capabilities expand into core research functions. He advocates for radical transparency to mitigate these concerns, though he admits that the long-term cultural impact of these tools remains a work in progress.
The financial rationale for these investments contrasts with the broader market skepticism currently observed in other sectors. While many executives express frustration over the delayed return on investment from AI initiatives, Bradway views biotech as a unique environment where long-term patience is a structural requirement. He notes that the industry standard for developing a single molecule often spans over a decade, making immediate financial gratification an unrealistic metric for success.
Industry participants who fail to embrace these computational tools risk significant competitive disadvantages. Bradway maintains that the cost of entering the field late will be substantial, whereas early adoption provides a clear pathway to sustained innovation. The company continues to monitor the performance of its agentic systems as they move from experimental phases into broader laboratory integration.
Future milestones will depend on the ability of these models to maintain accuracy as they scale to more complex therapeutic targets. Observers are watching to see if the reduction in drug discovery timelines translates into higher success rates in human clinical trials. The ultimate test for the firm will be whether the efficiency gains in the laboratory lead to a more productive pipeline of effective, safe medicines for patients.
