How NVIDIA & HOPPR AI Champions AI for Medical Imaging

NVIDIA's reasoning and Gen AI models NV-Reason and NV-Generate, designed for medical imaging development, are now available on HOPPR AI Foundry, a platform built for developers and researchers to train, evaluate and fine-tune medical imaging AI within a secure compliant environment.

This integration could signal a shift in how AI models are developed for specialised imaging applications, bringing together advanced computational capabilities with domain-specific infrastructure.

HOPPR is a US$30m start-up founded in 2019 that aims to bridge technology and clinical insight to make AI development in medical imaging faster and more transparent.

The company combines the expertise of clinicians, AI engineers and scientists to accelerate the deployment of foundation models that could make a meaningful impact in clinical practice.

Developers can use multimodal reasoning capabilities within NV-Reason for chest X-ray interpretation workflows in AI development.

The model provides structured analytical reasoning steps alongside model outputs, such as diagnosis or follow-up recommendations, offering clarity into how model interpretations are developed.

Meanwhile, NV-Generate creates realistic 3D medical images along with paired segmentation masks and anatomical annotations.

For example, an image could identify regions of interest from a scan – such as tumours, organs or tissues – and explain why it recommends a specific diagnosis.

This approach to explainability in AI-generated outputs could address one of the key challenges in deploying AI systems in high-stakes environments where transparency is essential.

Infrastructure and compliance considerations

HOPPR AI Foundry is built on NVIDIA A100 and H100 GPUs and is designed to support the development of medical imaging AI while maintaining strong data controls and compliance with international rules.

These include the Digital Imaging and Communications in Medicine standard for storing, transmitting and printing medical images.

The platform's architecture reflects the growing importance of regulatory compliance in AI development, particularly in sectors handling sensitive data.

The Foundry also supports model fine-tuning through HOPPR's Forward Deployed Services, which is a partnership model that combines HOPPR's machine learning and industry expertise with customer teams for imaging applications.

This collaborative approach could enable organisations to adapt foundation models to specific use cases without building entirely new infrastructure.

Expanding Foundation Model Access

The availability of NVIDIA open models within the HOPPR AI Foundry expands the range of foundation models available to developers building AI applications for medical imaging.

According to David Niewolny, Director of Business Development for Healthcare and Medical at NVIDIA: "The next generation of medical imaging AI will combine multimodal reasoning with the ability to generate high-fidelity clinical data.

"Platforms like the HOPPR AI Foundry enable developers to train and deploy medical imaging on NVIDIA accelerated computing with the performance and scale required for healthcare innovation."

Dr Khan Siddiqui, CEO and Co-Founder of HOPPR, says: "Medical Imaging AI is entering a new era where models can reason about images and generate new clinical data to accelerate application development.

"The HOPPR AI Foundry brings together secure infrastructure, curated datasets, fine-tuning tooling and advanced AI models to help developers build the next generation of imaging AI applications."

The NVIDIA-HOPPR integration was announced at NVIDIA GTC 2026, taking place between 16 and 19 March in San Jose, California.

The event features 1,000 sessions with 2,000 speakers and more than 450 sponsors, highlighting the scale of interest in AI infrastructure and application development across industries.