JPM25 AI Highlights: What NVIDIA’s partnerships signal, and the big AI policy shift

Welcome to this week’s special edition of the Roundup, featuring health AI insights from the J.P. Morgan Health Care 2025 conference in San Francisco. At almost every AI talk or panel, there’s at least one person who can’t resist mentioning how long they’ve been in AI (“WAY before it was cool”). JPM was no different, but I caught some truly interesting discussions about AI in healthcare that I’m excited to dig into in this newsletter. We’ll cover:

• Breakdown of NVIDIA’s healthcare partnership announcements

• 5 of the most interesting takes at JPM about what’s next in health AI

• Some wild funding rounds - health AI companies have already announced $1B in new funding (and it’s only mid-Jan)

📢 Headlines

NVIDIA Partners with Industry Leaders to Advance Genomics, Drug Discovery and Healthcare (NVIDIA Newsroom) 

NVIDIA made waves on day 1 of #JPM25 announcing major new partnerships in healthcare and life sciences:

• IQVIA: To build custom foundation models and agentic AI flows on IQVIA’s 64 petabytes+ of data* to speed research, clinical development and access to new treatments

• Illumina: To make the human genome more accessible to researchers, pharma, and life sciences to aid drug discovery

• Mayo Clinic: To build custom foundation models on Mayo Clinics’ dataset of 20 million whole-slide images with 10 million associated patient records* to enable personalized diagnostics and treatments.

• Arc Institute: To collaborate on open-source AI research to better understand diseases.

Sidebar on my takes about the NVIDIA announcement:

• Setting NVIDIA's impressive tech aside for a minute, look at those *data numbers I called out above. IQVIA’s data alone is equivalent to:

  • 10.6 billion songs (60x Spotify’s library)

  • 21 million hours of HD video content

  • 50+ complete copies of every book in the U.S. Library of Congress

  I heard some skepticism about these partnership announcements being merely hype (and sure, there’s always a marketing angle) but there’s a deeper story here: it’s clear that healthcare companies, with their massive troves of untapped data, are highly valuable for developers of AI tech and models.

• Speaking of models, notice the focus on custom foundation models in two major partnerships (IQVIA and Mayo). Prediction #4 from last week’s Roundup about 2025 Health AI Predictions is already playing out—this could be the year we see more bespoke models solving healthcare-specific problems. It’s worth watching this space to see if NVIDIA’s computational powers and technology, combined with large repositories of healthcare data, can finally crack persistent challenges like diagnostic accuracy and model explainability that have held back early health AI applications.

Ok, on to other headlines…

Trump revokes Biden's executive order on responsible AI development (MobiHealthNews)

President Trump revoked Biden’s 2023 executive order on responsible AI development, which directed HHS to create a safety program for AI in healthcare and required developers of high-risk AI systems to share safety test results with the government before public release. (Interestingly, this rollback aligns with sentiments at #JPM25, where several startups and VCs spoke about their hope for lighter regulation).

New AI tool for fighting health insurance denials could save hospitals billions, and help patients (CNBC)

Healthcare payments company Waystar has developed a tool to automatically draft appeal letters for denied claims, aiming to help providers recover funds from over 450 million annual denials.

Faro Health and Recursion Partner To Advance AI-powered Clinical Trial Design (Faro Health)

Faro Health and Recursion have teamed up to leverage AI for better clinical trial design. Faro Health’s platform will help Recursion Pharma, a clinical-stage biotech company, streamline how it plans and runs clinical trials for its growing list of potential new drugs.

🧪#JPM25 Roundup - What are the most exciting opportunities in AI?

At multiple JPM events, industry experts weighed in on the oft-asked question: what are the most exciting areas in healthcare for AI? Here were 5 of the most interesting takes:

👩‍⚕️ AI for enhancing patient-provider interaction. Using AI to analyze video and audio recordings of primary care visits, automate data entry into EHRs, perform sentiment analysis to gauge how the patient is feeling, and even assess physical aspects like gait and posture - to help physicians focus on their patients and prevent burnout. - Marylyn Ritchie, Vice Dean of Artificial Intelligence and Computing, Penn Medicine

🧬 AI for drug discovery. In the past 5-7 years, we’ve seen a boom in AI algorithms for identifying new targets and leads in early stages of drug development. To get to the next level, companies need to prove success in the clinic by working with drug development experts who understand how to design clinical trials for the best possibility of success. - Sally Zheng, Centenarian Fund

🩺 AI to support clinical decision-making. The technology for clinical decision support systems is not as mature as imaging AI and other areas, but startups have the opportunity to build on existing innovation here. - Preetha Ram, Pier 70 Ventures

💊 AI for drug repurposing. Identify generic drugs that are sitting on pharmacy shelves right now and that can work for patients with rare diseases. - David Fajgenbaum, cofounder of Everycure

🧓 AI for the aging population, longevity, and drug development. Huge number of companies are going to have drug candidates, and there’s an opportunity for leveraging AI to condense the timeframe to get these drugs through regulatory approval quickly. - Ed Dua, IASV Seed Ventures

💸 Funding

Truveta raised $320M to build the largest, most diverse database of genotypic and phenotypic information to date with the goal to accelerate drug discovery, enhance clinical trials, and train AI models.

Innovacer raised $275M to enhance its AI capabilities and copilot offerings for healthcare professionals.

HippocraticAI raised $141M for its staffing marketplace of AI agents to conduct low-risk patient-facing services.

QVentus raised $105M to build AI teammates for care teams to reduce administrative burden.

SlingshotAI raised $30M to build a foundation model for mental health from a16z, Felicis, and others.

Collate raised $30M to automate paperwork for life sciences companies.

Oxford Cancer Analytics raised $11M to develop blood tests for early lung cancer detection using advanced proteomics and AI.

That’s all for this week! See you next time 👋