Sponsored by: Medallion
Healthcare is at a turning point. AI, policy shifts, and operational pressure are reshaping what readiness looks like.
Join Elevate, Medallion's virtual event on September 1, where healthcare leaders will explore where AI is driving real efficiency, how teams are adapting, and what it takes to scale while keeping patient experience at the center.
Good morning, readers —
Let’s get to it:
Open source models shock the world, again
Specialized health AI gets a win
Anthropic teams up with the largest healthcare company in the world
25 new tools/partnerships, 9 funding updates, new AI jobs & link-worthy content
Read time: 6 minutes
Our Picks ✨
Highlights if you’ve only got 2 minutes…
1/
Open source models shock the world, again
In what some are calling a DeepSeek 2.0 moment, Moonshot AI, a Chinese AI lab, released Kimi K3, a 2.8 trillion-parameter open model that is already competing with leading frontier systems on several benchmarks. Its listed pricing is roughly 40% below comparable Claude models, putting more pressure on a market where OpenAI and Anthropic have captured much of the value.
That could be good news for almost everyone else. Lower model margins mean more spending flowing toward chips, power, data centers, cloud infrastructure, and software, while customers get more intelligence per dollar. It also helps explain why Nvidia has been so supportive of open source.
Across the broader AI market, companies may increasingly use a mix of models. Cheaper systems can handle routine work, while premium models take on the hardest tasks. Over time, software could simply route each request to the best model for the job.
Healthcare AI could benefit heavily. More competition should bring lower costs, greater vendor choice, and more control over sensitive data. Health systems and startups could run models on their own infrastructure, fine-tune them for clinical workflows, and build specialized products without sending most of the value back to a frontier lab.
K3 still uses a lot of tokens, so the sticker price may understate its true cost per task. Long story short, there is a real battle between open and closed models, geopolitics aside, and the long-term answer will likely be a mix of both. Either way, AI keeps getting cheaper and more accessible, giving healthcare companies more room to build and capture value. Onwards! (link)
2/
Specialized health AI gets a win
A new ARISE-led study tested 45 models and four clinical AI tools across 1,100 specialist-consult cases, and the results look encouraging for specialized healthcare AI. AMBOSS LiSA, Doximity Ask, OpenEvidence, and Glass Health all beat the leading general-purpose models, scoring roughly 80% to 86% on a severity-weighted measure compared with about 66% to 70% for GPT, Claude, and Gemini.
AI also helped physicians perform better, though the models still scored higher on their own. In many cases, doctors received useful recommendations from the AI and simply left them out of their final plans. That points to a workflow problem as much as a model problem.
Safety remains a real concern. More than 80% of errors came from omissions, such as missing a necessary test, treatment, referral, or follow-up. Even so, the study is broadly good news for the industry. Specialized clinical systems are showing clear advantages, and AI is already adding value before the human-AI workflow is fully worked out. (link)(tweet)
3/
Anthropic teams up with the largest healthcare company in the world
Anthropic is pushing deeper into healthcare, with Optum now deploying Claude across parts of its business. The partnership gives Anthropic a foothold inside UnitedHealth Group, the largest healthcare company in the world by revenue, and puts Claude closer to the administrative workflows that shape how care is delivered.
Optum says it will use Claude to reduce administrative burden and make everyday interactions clearer for members, while keeping clinicians in control. The announcement was light on specific use cases, but the scale alone makes it notable. Optum sits across claims, care delivery, pharmacy, analytics, and health services.
Anthropic also partnered with UST to embed Claude into CarePath, a claims and care management platform. That adds another distribution channel, but Optum is the bigger signal.
Anthropic has spent the year building out Claude for Healthcare, prior auth, claims appeals, care coordination, life sciences, drug discovery, and public health. Becoming an AI partner to UnitedHealth Group suggests that push is starting to land with the biggest players in the market. (link)(linkedin)
Tools & Partnerships 🔧
Latest on business, consumer, and clinical healthcare AI tools and partnerships…
TOOLS
Moonshot AI unveils Kimi K3 AI model: The open-weight model rivals leading frontier AI systems with a 1M-token context window and advanced reasoning, coding, and research capabilities at a lower cost. (link)
Clinicians are adopting AI faster than employers: Heidi Health's global survey found 86% of clinicians now use AI regularly, with most adopting tools independently despite ongoing concerns about hallucinations, privacy, and clinical judgment. (link)
Anthropic launches enterprise AI services firm: Anthropic partnered with Blackstone and Hellman & Friedman to launch Ode, embedding AI engineers into enterprises to build Claude-powered systems. (link)
Amigo AI debuts clinical copilot: The platform provides real-time documentation and clinical decision support to reduce clinician administrative burden. (link)
Mayo Clinic surpasses 150 deployed AI models: Mayo continues expanding AI across clinical and operational workflows to improve care delivery and efficiency. (link)
UCSF Health launches AI accelerator: UCSF partnered with venture firms to build and validate healthcare AI solutions directly within clinical environments. (link)
CHAI launches public health AI initiative: The Coalition for Health AI introduced a national program to help public health agencies safely evaluate and implement generative AI. (link)
Actava unveils 1T-parameter clinical AI model: CURA 1T is designed specifically for enterprise healthcare applications and claims strong performance on medical benchmarks. (link)
Queensland Institute pilots AI cancer screening: Researchers are testing an AI system that detects subtle cancer markers in pathology slides to enable lower-cost molecular screening. (link)
White House launches AI cybersecurity initiative: Gold Eagle uses AI to rapidly identify and coordinate responses to critical cybersecurity vulnerabilities across healthcare and other sectors. (link)
Autonomize AI launches Genie AI: The healthcare-native platform lets clinicians build autonomous AI workflows using natural language. (link)
El Camino Health adds millions in revenue with AI: Predictive AI tools optimized operating room scheduling and patient appointments, generating millions in additional revenue. (link)
Study questions AI-drafted patient messages: Dartmouth researchers found clinicians often spend longer editing AI-generated portal responses than writing messages from scratch. (link)
Voice AI boosts specialty care enrollment: A real-world study found conversational AI agents increased specialty care program enrollment by 340%. (link)
Carta Healthcare expands AI for oncology registries: Lighthouse for Cancer Registry Abstraction combines AI with human review to automate oncology data collection. (link)
Trice Imaging launches AI reporting: The company introduced AI-powered reporting within Trice Workspace to streamline prenatal imaging documentation and clinical workflows. (link)
PARTNERSHIPS
Mayo Clinic + Scale AI: Mayo Clinic partnered with Scale AI to deploy clinical AI tools that streamline documentation review, detect patient safety events, and reduce administrative burden. (link)
Boston Children’s Hospital + OpenEvidence: Boston Children's launched a multiyear collaboration with OpenEvidence to study how clinical AI tool usage impacts care delivery patterns. (link)
GE HealthCare + Catholic Health: GE HealthCare and Catholic Health formed a 10-year, $500M partnership to deploy imaging, diagnostic, and AI technologies across 40+ sites. (link)
CVS Health + Google Cloud: CVS Health expanded its partnership with Google Cloud as it prepares to launch Health100, an AI-powered consumer health platform designed to serve as a personalized healthcare assistant across pharmacies, providers, and insurers. (link)
Novant Health + K Health: Novant Health partnered with K Health to deploy clinical AI tools, including PatientGPT and an embedded digital assistant, to provide patients with 24/7 integrated virtual primary care. (link)
NHS England Midlands + Heidi Health: NHS England Midlands partnered with Heidi Health to deploy an AI clinical documentation tool across 15 trusts and 1,239 GP practices, streamlining workflows for 70,000 clinicians. (link)
Spanish Peaks Regional + Iterate AI: Spanish Peaks Regional Health Center launched a trial of Iterate AI's secure, private LLMs to automate workflows and fight medical claim denials. (link)
Samsung + Lark Health: Samsung partnered with Lark to scale home-based health ecosystems, deploying clinically validated AI coaching and remote monitoring tools to address senior preventative care. (link)
SpotitEarly + HITLAB: SpotitEarly partnered with HITLAB to accelerate the U.S. launch of its AI-powered breath-testing technology for early cancer detection. (link)
Deal Desk 💰
Spotlight on latest capital raises, M&A, and investments…
FUNDING
Neko Health, a Stockholm-based developer of AI-powered preventive full-body scanners, raised $700M in Series C funding led by Lightspeed Venture Partners and O.G. Venture Partners to fuel its expansion into the US. (link)
Chai Discovery, an AI drug discovery company, raised $400M at a $3.8B valuation. Index Ventures led, joined by Kleiner Perkins, Sequoia Capital, Dimension, Bain Capital Ventures, Thrive Capital, OpenAI and Yosemite. (link)
Miles Wang, an OpenAI researcher leaving to found an AI model company for drug discovery, is in talks to raise $200M at a $2B valuation. Lightspeed is in discussions to lead. (link)
Bunkerhill Health, an SF-based agentic AI platform for health systems, raised $55m from seed to Series B funding. Khosla Ventures led the B, joined by Sequoia Capital, Felicis, Optum Ventures, and Y Combinator. (link)
Hemispheric, an Israeli startup building a foundation AI model for the human brain, emerged from six years in stealth with $52M in funding. (link)
Corner Health, an L.A.-based platform for Nurse Practitioners to start their own practices, raised $32.5m in Seed and Series A funding. Oak HC/FT led the A, joined by First Round Capital and Zigg Capital. (link)
SONATA, a NYC, SF, and LA-based membership-based preventive-care provider, raised $7M in funding from Lux Capital, BoxGroup, Sunflower Capital, and others. (link)
Eli Lilly and Company (NYSE: LLY) invested in Oura, a smart ring maker. (link)
M&A
Lyric + Concert Genetics: Lyric acquired Concert Genetics to integrate its platform into Lyric42, enhancing AI-enabled healthcare payment accuracy and transparency for complex oncology, diagnostic, and genetic testing. (link)
market snapshot as of 7/19/26
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👇
