AI in healthcare is accelerating across radiology, drug discovery, medical devices and digital twins of the human body, according to a new NVIDIA survey.
The company’s second annual State of AI in Healthcare and Life Sciences report suggests the industry is shifting from experimentation to execution, with organisations reporting return on investment in core areas such as medical imaging and drug discovery.
It also shows growing use of open source software and AI models for targeted applications, alongside exploration of agentic AI to speed knowledge retrieval and research paper analysis.
Seventy per cent of respondents said their organisations are actively using AI, up from 63 per cent in 2024.
Sixty-nine per cent reported using generative AI and large language models, up from 54 per cent.
Forty-seven per cent said they are using or assessing agentic AI, where systems can act autonomously rather than simply respond to prompts.
Eighty-two per cent said open source software and models are moderately to extremely important to their AI strategy.
Among executives, 85 per cent said AI is helping increase revenue and 80 per cent said it is helping reduce costs.
Adoption rose across all industry segments, including digital healthcare, pharmaceutical and biotechnology, payers and providers, and medical technology and tools.
Digital healthcare led at 78 per cent, followed by medical technology at 74 per cent.
The most common workload was generative AI and large language models, cited by 69 per cent of respondents.
AI for data analytics and data science ranked second, followed by predictive analytics. Agentic AI ranked fourth, with 47 per cent reporting use or assessment.
Across the sector, AI is being applied in ways aligned with core functions.
Sixty-one per cent of medical technology respondents said they use AI in medical imaging, for example to help radiologists identify areas of concern more quickly.
Among pharmaceutical and biotechnology respondents, 57 per cent said AI is being used in drug discovery.
The leading overall use cases were clinical decision support, medical imaging and workflow optimisation.
The report indicates strong return on investment.
Fifty-seven per cent of medical technology respondents reported ROI from AI in medical imaging.
Nearly half of pharmaceutical and biotechnology respondents, 46 per cent, cited drug discovery and development among their top ROI use cases.
Among digital healthcare providers, 37 per cent identified virtual health assistants and chatbots as their leading ROI area, while 39 per cent of payers and providers cited administrative tasks and workflow optimisation.
AI is also improving back-office productivity and scaling into patient interaction and administrative functions.
Eighty-five per cent of respondents said their AI budgets will increase this year, while 12 per cent said spending will remain the same.
Forty-six per cent expect budgets to rise by more than 10 per cent.
“Over the next 12-18 months, the most visible and scalable impact of AI will come from logistics and administrative streamlining,” said John Nosta, president of NostaLab, a healthcare think tank.
“That’s where adoption curves are already steep — scheduling, documentation, coding, utilization management and care coordination.”
Dr Annabelle Painter, clinical AI strategy lead at Visiba UK, said: “Scaling generative AI in healthcare starts with focusing on real clinical and operational problems, rather than the technology itself.
“The organisations seeing impact are those that embed AI into existing workflows instead of layering AI on top as a separate tool.”
She added: “Healthcare organisations that successfully integrate AI are those that explicitly fund and prioritize evaluation as a core operational function, ensuring AI delivers measurable improvements in safety, quality and patient care over time.”
On open source, Nosta said: “Open models will shape the intellectual field. They are essential for exploration and for keeping the field honest.
“But in clinical environments where safety, liability and accountability are nonnegotiable, proprietary systems will remain necessary for validation, integration and trust.
“The key insight here is that discovery will be open, and deployment will demand stewardship.”
