Key Takeaways
- Nvidia has formed a strategic alliance with Abridge, a healthcare AI company, to create specialized AI models for medical conversations.
- The technology will be developed using Nvidia’s Nemotron open model family and operate exclusively on Abridge’s infrastructure.
- Applications include automating clinical documentation and enhancing clinical decision-making processes.
- Abridge, with a $5.3 billion valuation, specializes in AI-powered transcription and summarization of medical appointments.
- Deployment is scheduled for later in 2025; Nvidia holds an existing investment stake in Abridge.
Nvidia (NVDA) has announced a collaborative venture with Abridge, a healthcare-focused artificial intelligence startup, to create a specialized AI model designed for medical consultation scenarios. The Wall Street Journal first disclosed the partnership on June 11.
Shares of NVDA were changing hands near $137 when the news broke.
The upcoming model will leverage Nvidia’s Nemotron open model ecosystem. It’s designed to function solely within Abridge’s software environment, enhancing medical record-keeping and supporting healthcare provider decisions.
Headquartered in Pittsburgh, Abridge develops technology that records and transcribes medical consultations. The platform also creates automated clinical documentation, generates patient summary reports, and verifies medical billing information.
The Case for Open-Source Models
According to Davis Liang, Abridge’s director of applied science, economics played a significant role in selecting Nvidia’s open-source framework. Compact, specialized open models offer reduced operational costs compared to proprietary alternatives and can function on Abridge’s dedicated infrastructure.
Abridge plans to refine the Nemotron models using anonymized clinical information. CEO Shiv Rao emphasized that off-the-shelf models fall short without additional customization — medical AI requires specialized training and validation using actual clinical scenarios.
“Generic models are powerful, but clinical intelligence—it still has to be trained, it has to be shaped, and it has to be evaluated against real-world conditions,” Rao said.
The specialized model is slated for implementation before year-end. It will join multiple other AI models operating within Abridge’s technology stack.
Nvidia’s Healthcare Strategy
Kimberly Powell, who serves as Nvidia’s vice president of healthcare, highlighted that the Abridge collaboration demonstrates the versatility of open models throughout the healthcare sector and life sciences — spanning pharmaceutical development, medical equipment innovation, and digital health solutions.
Nvidia previously invested in Abridge, which secured $300 million in funding during its last financing round at a $5.3 billion company valuation. Since its 2018 launch, the organization has expanded its ambient listening technology throughout numerous major healthcare networks.
Dr. Joon Lee, who leads Emory Healthcare in Georgia, reported implementing Abridge’s technology across more than 3,000 medical professionals within the organization. He anticipates the Nvidia-enhanced model will speed up technological advancement.
This collaboration emerges as competing technology firms pursue comparable strategies. Microsoft recently unveiled a partnership with Mayo Clinic to develop healthcare AI utilizing Mayo’s medical records. OpenAI and Anthropic have also introduced healthcare-specific products.
Abridge hosted a presentation event in New York City on Thursday to reveal additional platform improvements.





