AI deployment in healthcare is being driven from the top, shadow AI is proliferating across clinical and administrative functions, and the infrastructure required to support secure, compliant AI workloads at the point of care is not yet in place, according to Nutanix.
As AI moves from the data centre to the bedside, where up to 75% of healthcare data is expected to be generated, the stakes around infrastructure readiness, data sovereignty, and clinical governance have never been higher.
“Healthcare organisations across APJ (Asia-Pacific and Japan) are under growing pressure to adopt AI, but clinician demand is colliding with the readiness of the infrastructure underneath it,” said Daryush Ashjari, Nutanix CTO and VP of solution engineering in APJ.
“The impact extends beyond IT; it can affect the availability of critical systems, access to data, and ultimately, the continuity of patient care. For healthcare leaders, the priority is to shift from reactive management and build a unified, hybrid approach that bridges the gap between data sovereignty compliance and the real-time, low-latency insights required at the patient’s bedside,” said Ashjari.
Findings show that shadow AI is widespread and largely unmanaged, with 79% of healthcare organisations encountering AI applications or agents being implemented by employees in non-IT functions, and 83% believing that AI tools and agents operating outside official oversight create business risk.
The same proportion, 83%, say silos between business units and IT make it difficult to effectively execute technology initiatives, deepening the governance challenge as AI adoption scales.
Infrastructure is not ready for AI at the point of care, with 88% of healthcare IT leaders viewing their current infrastructure as not fully ready to support deploying AI workloads on-premises. This is a significant gap given that AI inference at the point of care, rather than via cloud-only processing, is increasingly seen as essential to eliminating latency risks in clinical settings.
Single-patient rooms can generate up to 7TB of data annually, and high-density device environments such as ICU beds can include 15 to 20 connected devices, requiring local, low-latency AI processing to maintain clinical continuity.
AI is accelerating container adoption as healthcare modernises its application strategy, with 86% of healthcare organisations saying AI is meaningfully accelerating their adoption of containers, which enable AI models to be deployed locally at the bedside in secure, portable environments.
Also, 81% expect the level of application containerisation to increase at their organisation, and 80% are already building new applications in containers. Containers allow hospitals to keep data where it is generated, within their own walls, while enabling real-time AI-driven insights without compromising network performance.
AI agents are seen as transformative for healthcare operations, with 58% of healthcare IT leaders expect AI agents to improve productivity and efficiency, 57% anticipate agents will transform business processes and operations, and more than half (55%) see potential for AI agents to create new products, services, or revenue streams.
Looking three years ahead, 57% of organisations anticipate using agentic AI or autonomous agents, alongside generative AI (62%) and predictive analytics or machine learning models (55%).
Data sovereignty is a must-have, not a nice-to-have, with 72% of healthcare organisations saying data sovereignty is a high priority or a must-include when making infrastructure decisions.
In addition, 54% run containerised applications on-premises or on private clouds today, and 54% feel the need to run infrastructure within a single country due to customer or stakeholder expectations. This reflects the sensitivity of protected health information (PHI) and the compliance requirements governing where this data can be stored and processed.
AI adoption is being driven from the top, with scale coming fast as 55% of healthcare organisations anticipate having more than five AI-enabled applications within three years, including 12% who expect to be running more than 10.
Further, 63% currently run AI applications on managed service providers, with hybrid deployment models expected to remain the norm as organisations look to support AI centrally and at the point of care.
Nutanix said healthcare organisations are accelerating into AI without the infrastructure beneath them to support it safely. Closing the gap from the data centre to the bedside requires a fundamental rethink of how healthcare IT is architected, governed, and scaled.
