Sponsored by: Ketryx
How does a best-in-class engineering organization build a medical device from scratch?
Meta’s Reality Labs embedded quality directly into existing workflows so engineers could spend more time coding and less time on documentation.
Hear Lucas Fernandez, Director of Medical Devices at Meta, explain how they did it so teams of any size can borrow the playbook. 👉 Register here to join live or receive the recording.
Good morning, readers —
Let’s get to it:
OpenAI's state of enterprise AI in healthcare
OpenEvidence doubles valuation as revenue surges
How safe are today’s models in clinical use?
12 new tools/partnerships, 7 funding updates, new AI jobs & link-worthy content
Read time: 5 minutes
Our Picks ✨
Highlights if you’ve only got 2 minutes…
1/
OpenAI's state of enterprise AI in healthcare
OpenAI just released its first State of Enterprise AI report, drawing on real usage data from more than one million workplace accounts. AI is clearly no longer experimental at work. Across industries, employees sent 30% more messages year over year, while token usage per company jumped 320x as reasoning models moved into real products. About 75% of workers said ChatGPT improved the speed or quality of their work, and the same share reported doing tasks they could not do before.
Healthcare stands out in the report. It was the second fastest growing sector for ChatGPT adoption, with usage up 8x in the past year, trailing only tech. Clinicians and health system staff are using AI to summarize notes, draft documentation, navigate guidelines, and support research. Power users report saving hours each week. It is encouraging to see healthcare leading the pack as AI reshapes workflows, reduces administrative drag, and expands what care teams can handle day to day. (link)(linkedin)
2/
OpenEvidence doubles valuation as revenue surges
OpenEvidence, The chatGPT for clinicians, is back in the spotlight after reports that it is raising $250M at a $12B valuation. That would double its valuation from a $6B Series C closed just two months ago, which itself followed a $210M Series B at $3.5B earlier this year. The growth and velocity is striking. OpenEvidence is now generating more than $150M a year in revenue from just ads, with fewer than one million users and roughly 90% gross margins even after compute costs. That puts it in rare territory among AI application companies crossing both $10B valuation and $100M-plus revenue. It also seems that 2026 will be the year AI ads cross the rubicon.
The new funding is expected to accelerate R&D, especially around multimodal capabilities, while building on recent milestones like its licensing deal with the National Comprehensive Cancer Network and the launch of a HIPAA-secure Dialer. The bigger signal may be business model driven. OpenEvidence shows that AI-native healthcare apps can scale extremely quickly, monetize through ads, and still meet clinical and privacy standards. (link)(linkedin)
3/
How safe are today’s models in clinical use?
A new Stanford and Harvard study takes a hard look at how safe today’s AI models really are in clinical use and the results are more nuanced than the headlines suggest. Researchers evaluated 31 LLMs on 100 real outpatient eConsult cases across 10 specialties, generating nearly 13,000 expert annotations. These were not routine visits but clinical gray zones where primary care doctors were already asking specialists for help.
The findings are sobering. Even top models produced 12 to 15 harms per 100 cases, while weaker models exceeded 40. Most harm came from omission, meaning AI failed to recommend a critical test or follow up rather than suggesting something overtly dangerous. In many cases, the best models slightly outperformed physicians, while average performance was similar.
The key takeaway is not that AI is unsafe, but that “mostly right” systems can be risky in edge cases. Measuring harm directly, not just accuracy, is becoming essential as clinical AI scales. (link)
Tools & Partnerships 🔧
Latest on business, consumer, and clinical healthcare AI tools and partnerships…
TOOLS
Sword Health introduces MindEval for mental health screening: Sword Health launched MindEval, an AI-powered assessment tool designed to help identify mental health needs and guide users toward appropriate care pathways. (link)
Microsoft open-sources AI to map tumors from tissue slides: Microsoft released GigaTIME, an open-source AI model that extracts immune and tumor insights from standard pathology slides, enabling low-cost cancer analysis across 24 cancer types. (link)
Mass General Brigham spins out AIwithCare for clinical trial matching: The new startup uses AI embedded in the EHR to automate patient identification for trials, targeting screening costs that make up about one-third of total trial expenses. (link)
European study shows AI assistants cut clinician documentation time: Europe’s largest study of AI medical assistants found a 29% reduction in documentation time across 375,000 notes, alongside improved clinician well-being and greater presence in patient visits. (link)
National Academy of Medicine launches AI patient safety initiative: The National Academy of Medicine announced a new initiative focused on improving patient safety as AI adoption grows, aiming to set best practices for governance, monitoring, and risk mitigation. (link)
OpenAI releases GPT-5.2 amid competition with Google: OpenAI launched GPT-5.2 with upgrades in reasoning, vision, coding, and hallucination reduction, introducing Instant, Thinking, and Pro tiers as it responds to pressure from Google’s Gemini models. (link)
Freed expands AI clinician assistant with coding and prep tools: Freed added automated ICD-10 coding, an interactive pre-visit AI assistant for clinician preparation, and auto-generated clinical letters for patients to its AI clinician-assistant platform. (link)
Headspace pivots to enterprise mental health with AI chatbot: Headspace is shifting from consumer subscriptions toward employer and payor partnerships, emphasizing its AI chatbot Ebb for everyday emotional support as it restructures its therapist workforce. (link)
DeepHealth launches Breast Suite for imaging and workflow: RadNet subsidiary DeepHealth introduced Breast Suite, an AI platform designed to support breast cancer detection, risk stratification, and imaging workflow management. (link)
PARTNERSHIPS
athenahealth + Microsoft: athenahealth will embed Microsoft’s Dragon Copilot into its EHR to support clinical documentation and reduce administrative workload. (link)
OhioHealth + Abridge: OhioHealth is scaling Abridge’s ambient AI tool across ambulatory care settings to streamline clinician documentation. (link)
Allina Health + Qventus: Allina Health partnered with Qventus to co-develop AI tools aimed at improving operational efficiency and care delivery. (link)
Deal Desk 💰
Spotlight on latest capital raises, M&A, and investments…
FUNDING
Medra, an SF-based platform for physical AI scientists, raised $52M in Series A funding. Human Capital led, joined by Lux Capital, Neo, NFDG, Catalio Capital Management, Menlo Ventures, 776, and Fusion Fund. (link)
Radial, an NYC-based mental health care startup, raised $50M in Series A funding. General Catalyst led, joined by Solari Capital, JSL Health Capital, Founder Collective, BoxGroup, and Scrub Capital. (link)
Valinor, an SF-based clinical trial tech startup, raised $13M in seed funding from CRV, Harpoon Ventures, Amino Collective and Pelion Venture Partners. (link)
Lin Health, a Denver-based provider of behavioral care for chronic pain recovery, raised $11M in Series A funding. Proofpoint Capital led, joined by others. (link)
Subsense, a NYC-based maker of brain-computer interfaces to treat neurodegenerative diseases, raised $10M in new seed funding from Golden Falcon Capital and others. (link)
a2z Radiology AI, a Boston-based medical imaging interpretation startup, raised $4.5M in seed funding led by Khosla Ventures and SeaX Ventures. (link)
Origin Therapy, a Fairfax, Va.-based enabler of private practice pediatric speech and occupational therapy, raised $2.6M in pre-seed funding led by Distributed Ventures. (link)
as of 12/14/25
Other Relevant News 🔍
News, podcasts, blogs, tweets, resources, etc…
AI Job Opportunities 💼
Explore our AI Job Board or contact us to feature roles in our newsletter…
Visuals of the Week 📸
Funny memes, cool pics, and interesting data from around the web…
That’s it for this week friends! Back to reading — I’ll see you next week.
Stay classy,
PS. I write this newsletter for you. So if you have any suggestions or questions, feel free to reply to this email and let me know
How was this week's newsletter? Tap your choice below👇
