What the new model is designed to do

The co-developed foundation model is tailored specifically for medical conversations and will power Abridge’s expanding “clinician intelligence platform.” Its stated clinical functions include:

• Automated clinical note generation: The model will convert real-time patient-clinician conversations into structured, billable notes, extending Abridge’s existing ambient AI capability .
• Billing and revenue-cycle support: By embedding clinically grounded reasoning into the documentation workflow, the model aims to streamline billing and payer interactions .
• Real-time clinical decision support: The model is designed to provide decision assistance at the point of care, using the clinical conversation as input .
• Life sciences and payer connectivity: Abridge’s recently unveiled platform positions the model as a hub connecting care delivery, payment, and evidence-based treatment, including pharmaceutical trial screening .

Trained on Nvidia’s Nemotron open model family and deployed on Abridge’s infrastructure, the model is the first foundation model purpose-built for clinical conversations rather than adapted from general-purpose LLMs .

Abridge’s existing platform: what the model plugs into

Founded in 2018 by cardiologist Dr. Shiv Rao, Abridge uses ambient AI—automatic speech recognition combined with generative AI—to listen to patient-clinician conversations and produce structured, billable clinical notes in real time . Key platform characteristics that frame the partnership:

• Enterprise scale: The platform is deployed in more than 250 health systems, including Northwestern Medicine, Emory Healthcare, and Johns Hopkins .
• EHR integration: Abridge integrates directly into Epic workflows—including Haiku for mobile and Hyperspace for desktop—without requiring clinicians to change how they work .
• Clinician intelligence vision: On June 11, 2026, Abridge unveiled a broader “AI-native clinician intelligence platform” that organizes clinical, financial, and evidence-based workflows around the patient, extending beyond documentation into billing, payer adjudication, and drug-discovery collaboration with organizations like Eli Lilly .
• Industry trust: The platform is rated #1 Best in KLAS for ambient AI .

The new Nvidia model will serve as the AI engine behind this architecture, replacing or augmenting Abridge’s existing models in all those workflows.

How the Nvidia-Abridge relationship deepened

The co-development deal is the culmination of a multi-stage relationship that evolved from venture investment to foundational technology partnership:

• March 2024 – Initial research collaboration: Abridge and Nvidia announced that Nvidia would provide compute resources, foundation model expertise, and GPU-accelerated infrastructure to help Abridge scale its multilingual clinical conversation platform across the U.S. healthcare system. The announcement also disclosed an investment from NVentures, Nvidia’s venture capital arm .
• February 2025 – Series D investment: Abridge raised a $250 million Series D co-led by Elad Gil and IVP, with participation from NVentures, Lightspeed Venture Partners, CVS Health Ventures, Redpoint Ventures, and others. The round valued Abridge at $2.75 billion post-money . The funding coincided with Abridge surpassing 100 health-system deployments .
• June 2026 – Clinical foundation model: The partnership shifted from infrastructure provision to co-developing a purpose-built healthcare AI model, marking Nvidia’s deepest integration yet into clinical workflow AI .

NVentures’ participation across both the early-stage investment and the Series D round gives Nvidia a financial stake in Abridge’s growth, while the model partnership locks one of the fastest-scaling clinical AI platforms onto Nvidia’s AI stack .

Why this matters for Nvidia’s healthcare strategy

The Abridge partnership is not an isolated move—it is one piece of an aggressive, multi-front push by Nvidia into the $10 trillion healthcare and life sciences industry . That strategy now spans three distinct domains:

• Clinical workflow (Abridge): By co-developing the documentation model, Nvidia gains a direct foothold in the point-of-care clinical AI market, where ambient documentation is rapidly becoming standard. Abridge’s platform supports more than 100 million clinical conversations annually .
• Drug discovery (Eli Lilly): In October 2025, Nvidia and Eli Lilly announced a partnership to build what they call the most powerful AI supercomputer owned by a pharmaceutical company—an “AI factory” powered by more than 1,000 of Nvidia’s Blackwell Ultra GPUs . In January 2026, the companies deepened that relationship with a $1 billion, five-year AI co-innovation lab in the San Francisco Bay Area, focused on drug discovery and development using Nvidia’s next-generation Vera Rubin chips .
• Diagnostic imaging (GE HealthCare): Nvidia and GE HealthCare expanded a 16-year relationship in March 2025 to advance autonomous imaging, beginning with X-ray and ultrasound applications. GE HealthCare is using the Nvidia Isaac for Healthcare simulation platform to train, test, and validate autonomous imaging workflows .

Together, these partnerships position Nvidia as an infrastructure and model-building partner across the drug pipeline—from early-stage discovery (Lilly), to diagnostic imaging (GE), to real-world clinical care delivery and documentation (Abridge). Industry estimates cited in coverage of the Abridge deal suggest that 70% of healthcare organizations are actively using AI, giving Nvidia a substantial adoption tailwind across each of these bets .

For Abridge, the partnership provides the compute resources, model lineage, and venture backing to build a healthcare-specific foundation model at a scale that few startups could achieve independently. Fortune described Abridge’s ambition as building “the operating system for medicine,” and the Nvidia partnership both funds and technically enables that vision .

What the new model is designed to do

The co-developed foundation model is tailored specifically for medical conversations and will power Abridge’s expanding “clinician intelligence platform.” Its stated clinical functions include:

• Automated clinical note generation: The model will convert real-time patient-clinician conversations into structured, billable notes, extending Abridge’s existing ambient AI capability .
• Billing and revenue-cycle support: By embedding clinically grounded reasoning into the documentation workflow, the model aims to streamline billing and payer interactions .
• Real-time clinical decision support: The model is designed to provide decision assistance at the point of care, using the clinical conversation as input .
• Life sciences and payer connectivity: Abridge’s recently unveiled platform positions the model as a hub connecting care delivery, payment, and evidence-based treatment, including pharmaceutical trial screening .

Trained on Nvidia’s Nemotron open model family and deployed on Abridge’s infrastructure, the model is the first foundation model purpose-built for clinical conversations rather than adapted from general-purpose LLMs .

Abridge’s existing platform: what the model plugs into

Founded in 2018 by cardiologist Dr. Shiv Rao, Abridge uses ambient AI—automatic speech recognition combined with generative AI—to listen to patient-clinician conversations and produce structured, billable clinical notes in real time . Key platform characteristics that frame the partnership:

• Enterprise scale: The platform is deployed in more than 250 health systems, including Northwestern Medicine, Emory Healthcare, and Johns Hopkins .
• EHR integration: Abridge integrates directly into Epic workflows—including Haiku for mobile and Hyperspace for desktop—without requiring clinicians to change how they work .
• Clinician intelligence vision: On June 11, 2026, Abridge unveiled a broader “AI-native clinician intelligence platform” that organizes clinical, financial, and evidence-based workflows around the patient, extending beyond documentation into billing, payer adjudication, and drug-discovery collaboration with organizations like Eli Lilly .
• Industry trust: The platform is rated #1 Best in KLAS for ambient AI .

The new Nvidia model will serve as the AI engine behind this architecture, replacing or augmenting Abridge’s existing models in all those workflows.

How the Nvidia-Abridge relationship deepened

The co-development deal is the culmination of a multi-stage relationship that evolved from venture investment to foundational technology partnership:

• March 2024 – Initial research collaboration: Abridge and Nvidia announced that Nvidia would provide compute resources, foundation model expertise, and GPU-accelerated infrastructure to help Abridge scale its multilingual clinical conversation platform across the U.S. healthcare system. The announcement also disclosed an investment from NVentures, Nvidia’s venture capital arm .
• February 2025 – Series D investment: Abridge raised a $250 million Series D co-led by Elad Gil and IVP, with participation from NVentures, Lightspeed Venture Partners, CVS Health Ventures, Redpoint Ventures, and others. The round valued Abridge at $2.75 billion post-money . The funding coincided with Abridge surpassing 100 health-system deployments .
• June 2026 – Clinical foundation model: The partnership shifted from infrastructure provision to co-developing a purpose-built healthcare AI model, marking Nvidia’s deepest integration yet into clinical workflow AI .

NVentures’ participation across both the early-stage investment and the Series D round gives Nvidia a financial stake in Abridge’s growth, while the model partnership locks one of the fastest-scaling clinical AI platforms onto Nvidia’s AI stack .

Why this matters for Nvidia’s healthcare strategy

The Abridge partnership is not an isolated move—it is one piece of an aggressive, multi-front push by Nvidia into the $10 trillion healthcare and life sciences industry . That strategy now spans three distinct domains:

• Clinical workflow (Abridge): By co-developing the documentation model, Nvidia gains a direct foothold in the point-of-care clinical AI market, where ambient documentation is rapidly becoming standard. Abridge’s platform supports more than 100 million clinical conversations annually .
• Drug discovery (Eli Lilly): In October 2025, Nvidia and Eli Lilly announced a partnership to build what they call the most powerful AI supercomputer owned by a pharmaceutical company—an “AI factory” powered by more than 1,000 of Nvidia’s Blackwell Ultra GPUs . In January 2026, the companies deepened that relationship with a $1 billion, five-year AI co-innovation lab in the San Francisco Bay Area, focused on drug discovery and development using Nvidia’s next-generation Vera Rubin chips .
• Diagnostic imaging (GE HealthCare): Nvidia and GE HealthCare expanded a 16-year relationship in March 2025 to advance autonomous imaging, beginning with X-ray and ultrasound applications. GE HealthCare is using the Nvidia Isaac for Healthcare simulation platform to train, test, and validate autonomous imaging workflows .

Together, these partnerships position Nvidia as an infrastructure and model-building partner across the drug pipeline—from early-stage discovery (Lilly), to diagnostic imaging (GE), to real-world clinical care delivery and documentation (Abridge). Industry estimates cited in coverage of the Abridge deal suggest that 70% of healthcare organizations are actively using AI, giving Nvidia a substantial adoption tailwind across each of these bets .

For Abridge, the partnership provides the compute resources, model lineage, and venture backing to build a healthcare-specific foundation model at a scale that few startups could achieve independently. Fortune described Abridge’s ambition as building “the operating system for medicine,” and the Nvidia partnership both funds and technically enables that vision .