AWS expands tools for imaging, genomics and clinical data with 2025 upgrades
AWS also introduced the Trainium 3 chip and new AI-optimised servers designed to handle resource-intensive healthcare workloads such as genomics pipelines, biomarker discovery, molecular-simulation tasks and large-scale clinical-trial data processing.
- Dec 3, 2025,
- Updated Dec 3, 2025 12:47 PM IST
Amazon Web Services (AWS) announced a series of upgrades at re:Invent 2025 that are expected to influence how hospitals, diagnostics firms, pharmaceutical companies and research institutions use cloud-based tools for imaging, genomics and clinical-data processing.
Setting the broader shift in motion, AWS Chief Executive Officer Matt Garman said: “Two years ago, people were building AI applications. Now, people are building applications that have AI in them.”
The company expanded its Amazon Nova model suite to include multimodal capabilities that can analyse medical images, structured clinical records, text and video in a single workflow. These capabilities are expected to support faster radiology interpretation, clinical-note summarisation, patient-triage systems and automated documentation tools. Garman described the larger direction of these capabilities, noting: “The time for simple copilots is over… we are moving into the Age of the AI Agent.”
AWS also introduced the Trainium 3 chip and new AI-optimised servers designed to handle resource-intensive healthcare workloads such as genomics pipelines, biomarker discovery, molecular-simulation tasks and large-scale clinical-trial data processing. Emphasising AWS’s infrastructure footprint, Garman said, “AWS has the broadest AI infrastructure in the world.” Discussions at re:Invent sessions suggested that the upgraded compute layer could help reduce the time and cost of population-scale genomics and data-heavy drug-development programmes.
The company extended its agentic-AI frameworks to regulated sectors, enabling healthcare organisations to automate tasks such as clinical-document generation, EMR population, trial-data curation, imaging-workflow routing and real-world-evidence analysis. Explaining the rationale for custom model development, Garman asked: “What if you could integrate your data at the right time during the training of a frontier model, and then create a proprietary model that is just for you?”
Additional upgrades to analytics, vector search, data-governance features and privacy-first data pipelines aim to help organisations manage sensitive patient information while scaling digital-health and research workflows. Presentations at the conference highlighted early use cases in hospitals and life-sciences companies testing the new data-layer tools for imaging and clinical-data workloads.
While AWS did not announce healthcare-exclusive products, the combined set of model, compute and governance upgrades indicates a broader effort to strengthen its position in health and life sciences. Calling the overall upgrade cycle a significant step, Garman said: “This marks a big leap forward in the journey towards unlocking the value of AI.” Industry executives said providers continue to face limits in “scalable, compliant compute”, and AWS’s 2025 stack is intended to address these gaps as organisations increase investment in digital diagnostics, genomics and clinical-data automation.
“By extending Nova’s multimodal capabilities, advancing Trainium silicon, and strengthening governance layers, AWS is assembling a unified architecture capable of supporting population-scale genomics, compliant clinical automation, and high-throughput imaging,” said Prabhu Ram, Vice President (VP), Industry Research Group (IRG), CyberMedia Research (CMR).
"Together, these advances remove long-standing compute and integration bottlenecks, enabling hospitals and life-science organisations to run genomics, imaging, and clinical-data pipelines at true enterprise scale," he said.
Amazon Web Services (AWS) announced a series of upgrades at re:Invent 2025 that are expected to influence how hospitals, diagnostics firms, pharmaceutical companies and research institutions use cloud-based tools for imaging, genomics and clinical-data processing.
Setting the broader shift in motion, AWS Chief Executive Officer Matt Garman said: “Two years ago, people were building AI applications. Now, people are building applications that have AI in them.”
The company expanded its Amazon Nova model suite to include multimodal capabilities that can analyse medical images, structured clinical records, text and video in a single workflow. These capabilities are expected to support faster radiology interpretation, clinical-note summarisation, patient-triage systems and automated documentation tools. Garman described the larger direction of these capabilities, noting: “The time for simple copilots is over… we are moving into the Age of the AI Agent.”
AWS also introduced the Trainium 3 chip and new AI-optimised servers designed to handle resource-intensive healthcare workloads such as genomics pipelines, biomarker discovery, molecular-simulation tasks and large-scale clinical-trial data processing. Emphasising AWS’s infrastructure footprint, Garman said, “AWS has the broadest AI infrastructure in the world.” Discussions at re:Invent sessions suggested that the upgraded compute layer could help reduce the time and cost of population-scale genomics and data-heavy drug-development programmes.
The company extended its agentic-AI frameworks to regulated sectors, enabling healthcare organisations to automate tasks such as clinical-document generation, EMR population, trial-data curation, imaging-workflow routing and real-world-evidence analysis. Explaining the rationale for custom model development, Garman asked: “What if you could integrate your data at the right time during the training of a frontier model, and then create a proprietary model that is just for you?”
Additional upgrades to analytics, vector search, data-governance features and privacy-first data pipelines aim to help organisations manage sensitive patient information while scaling digital-health and research workflows. Presentations at the conference highlighted early use cases in hospitals and life-sciences companies testing the new data-layer tools for imaging and clinical-data workloads.
While AWS did not announce healthcare-exclusive products, the combined set of model, compute and governance upgrades indicates a broader effort to strengthen its position in health and life sciences. Calling the overall upgrade cycle a significant step, Garman said: “This marks a big leap forward in the journey towards unlocking the value of AI.” Industry executives said providers continue to face limits in “scalable, compliant compute”, and AWS’s 2025 stack is intended to address these gaps as organisations increase investment in digital diagnostics, genomics and clinical-data automation.
“By extending Nova’s multimodal capabilities, advancing Trainium silicon, and strengthening governance layers, AWS is assembling a unified architecture capable of supporting population-scale genomics, compliant clinical automation, and high-throughput imaging,” said Prabhu Ram, Vice President (VP), Industry Research Group (IRG), CyberMedia Research (CMR).
"Together, these advances remove long-standing compute and integration bottlenecks, enabling hospitals and life-science organisations to run genomics, imaging, and clinical-data pipelines at true enterprise scale," he said.