Skip to main content

Dyania Health and NVIDIA bring clinical reasoning AI on-premises

16 NOVEMBER 2026·2 MIN READ·2 SOURCES

Dyania Health announced a collaboration with NVIDIA to put its Synapsis AI clinical-reasoning system to work inside health systems. The companies say patient records stay within each health system, with the system running on its own infrastructure.

Dyania Health and NVIDIA bring clinical reasoning AI on-premises

Key takeaways · 4

  • 01

    Dyania and NVIDIA are working to deploy Synapsis AI inside health systems, using local infrastructure rather than transferring patient records outside the system.

  • 02

    Synapsis AI is built on NVIDIA AI Enterprise; Dyania uses NeMo for training and reinforcement learning, and TensorRT and TensorRT-LLM for inference.

  • 03

    The Cleveland Clinic study had Synapsis answer 88 clinical questions for each of 7,812 patients who received aortic valve replacement between 2005 and 2021.

  • 04

    The resulting model estimates valve-failure risk at three, five and 10 years, but the tool still requires external validation from clinicians.

How the collaboration is structured

Dyania supplies the clinical-reasoning engine and implementation experience, while NVIDIA supplies accelerated computing and software intended for on-premises operation.[1] Synapsis AI is built on NVIDIA’s AI Enterprise software stack.[1] Dyania says it trains its proprietary clinical models using NVIDIA NeMo for model training and reinforcement learning, then uses NVIDIA TensorRT and TensorRT-LLM for inference on hospital infrastructure.[1] The companies say patient records remain inside the health system rather than being transferred outside it.[1]

What Synapsis AI does

Synapsis AI analyzes patient histories across years of notes, operative and imaging reports, and lab results.[1] Dyania says the system answers clinical questions with evidence drawn from the patient’s chart and can process records across a health system’s patient population on that system’s own infrastructure.[1] That capability underpinned the collaboration’s first work: a Cleveland Clinic study of patients who received aortic valve replacements from 2005 through 2021.[1] The analysis used up to 16 years of care records, much of it in narrative notes and echocardiogram reports.[1]

The valve study and its limits

In the Cleveland Clinic study, Synapsis AI answered 88 clinical questions for each of 7,812 patients.[1] The questions covered valve characteristics, coexisting conditions, lab trends, medications, follow-up ultrasound measurements and later clinical events.[1] Researchers used the resulting dataset to train a machine-learning model to predict valve failure and the risk timelines for repeat procedures.[1] The model estimates an individual’s risk at three, five and 10 years, updating its outlook with each new ultrasound.[1] The tool is intended to inform care, but it still requires external validation from clinicians.[1]

What comes next

The companies plan to bring Synapsis AI to health systems in the United States and internationally, publish joint case studies, and integrate it with NVIDIA’s healthcare software stack.[1] They are evaluating NVIDIA Confidential Computing, Dynamo, Multi-Instance GPU and Blackwell Ultra GPUs for possible use in deployments.[1] Dyania says it has invested more than 25,000 physician hours in annotating medical cases for Synapsis AI.[2] Fierce Healthcare reports that Cleveland Clinic began working with Dyania a year earlier on clinical-trial recruitment using its AI platform.[2]

For health-system teams, the collaboration puts clinical-record analysis and model inference on the institution’s own infrastructure, while the companies say records stay within the system. The valve study shows a potential research workflow, but its model is not presented as clinically validated; teams considering similar tools should account for the stated need for clinician validation.

Why it matters
For teams

AI training for your team

kju.ai is built for teams: daily sessions, team analytics, and rollout support for your organization.

How this developed

  1. 16 November 2026

    Dyania Health and NVIDIA bring clinical reasoning AI on-premises

Sources

AI fluency, one session a day, built for your work.