What Comes Next
StatusGoStrategy in Motion AI in Healthcare
 

In This Issue

→  The Rise of AI Agents: Five Milestones
→  The State of the Hospital
→  AI Meets the Clinic
→  Follow the Money
The Rise of AI Agents · Five Milestones

AI agents are here

In under a year, agents moved from local experiments to mass-market personal assistants and the first attempts at standards for transacting. Capability is no longer the question. What is left is the hard part: oversight, access, and accountability.

 
Nov 2025
OpenClaw
Open-source
Autonomy can run locally
 
Aug 11
Grok Bot
xAI
The always-on cloud coworker
 
Aug
Instinct
Startup
Each agent gets its own inbox
 
Sep 8
Muse
Meta
The mass-market personal agent
 
Oct 6
Personal Agent Protocol
Sierra + Meta
A proposed standard for transacting

Sources: company announcements and launch coverage, Nov 2025 to Oct 2026. OpenClaw originated as Warelay in November 2025 and took the OpenClaw name in January 2026.

The State of the Hospital
1
The big systems are shrinking to profitability
Providers

Per Medicare hospital cost reports for FY2024, compiled by the Lown Institute, the ten largest systems took 25% of US hospital profits, $29 billion, with HCA's hospitals alone reporting $16.5 billion, more than the other nine combined. Most of the ten own fewer hospitals than in 2020 (Ascension -29%, CHS -31%), shifting toward outpatient and surgery centers. Lown Institute

2
The US is losing hospitals faster than it builds them
Access

A JAMA study counted 432 US hospital closures and 216 openings from 2010 to 2025. Two findings stand out: urban and rural hospitals closed at similar rates, and 117 of the closed sites kept outpatient services. Closures clustered among safety-net, small and for-profit hospitals; new openings skewed urban and for-profit. The hospital map is being redrawn, not simply shrinking. Washington Post

AI Meets the Clinic
3
OpenAI's physician network, and the evidence that isn't there yet
Clinical AI

OpenAI has contracted hundreds of physicians across specialties and 49 languages to evaluate example ChatGPT health conversations for safety and quality, over 700,000 responses reviewed by July. The reach is wide: OpenAI says 300 million people use ChatGPT for health advice weekly. But evaluation volume is not outcome evidence, and OpenAI concedes none yet shows the answers help patients. Business Insider

4
Google tests a diagnostic AI on real patients
Clinical AI

In a Lancet feasibility study, Google's AMIE chatbot took histories from 98 primary care patients before their visits, under physician supervision. There were zero safety stops; supervisors noted one hallucination and added information in five cases. Clinicians rated the conversations favorably. It is a supervised, single-center test of safety and quality, not evidence of better outcomes. Google Research (The Lancet)

5
Why Teladoc is betting on bundling
Digital Health

Teladoc, nursing a $200 million 2025 loss and a cut 2026 outlook, is bundling AI into its Solo platform. SoloScribe, an ambient note-writer, is available now; SoloVitals, which reads heart and respiratory rates from a camera, is in private preview, with broader release expected in December. The strategy is consolidation, fewer point solutions. Commercial results are still to come. Healthcare Dive

6
Health systems and workers, shaping AI together
Workforce

Health systems are bringing workers into AI decisions. Kaiser gave its unions half the seats on the task force overseeing AI investments; HCA had front-line nurses help shape a handoff tool; Mayo ran a nurse-led contest that drew 450+ ideas. The expectation is better usability and adoption. Whether co-design delivers that is still to be shown. Becker's

7
A government Slack room where AI lobbies health policy
Policy

A KFF Health News investigation describes a 1,700-member CMS Slack where AI and tech firms discussed health-AI policy with federal officials, largely out of public view. Per KFF, one official hoped CMS could be a "sales engine" for health apps, and legal experts it quoted say the group resembles a federal advisory committee that skipped the public, balanced process. CBS News / KFF

Follow the Money
8
The AI bottleneck moves from chips to power
Infrastructure

Morgan Stanley says the AI bottleneck is shifting from chips to electricity. It estimates US data-center developers face a 34% shortfall in the power their planned buildout needs through 2028. Nvidia and Broadcom look insulated, their 2027 forecasts intact; the exposure falls on makers of memory, optics and power components, whose orders slip if sites can't be energized. Reuters

9
OpenAI's $50 billion, and the $18 billion that wasn't there
Markets

OpenAI told investors its annualized revenue is about $50 billion, $18 billion below the $68 billion reported in September. The difference reflects how revenue was counted, the higher figure included partners' gross revenue, not falling demand. The clarification hit a jittery market: Nvidia fell 3% that day amid a broader AI-stock pullback, as suppliers' plans rest on OpenAI's private numbers. CNBC

10
SoftBank chases $100 billion more for AI
Deals

SoftBank's Masayoshi Son is reportedly seeking up to $100 billion from Gulf investors for a new fund that would buy companies and run them using AI, the Financial Times reported; talks are exploratory and SoftBank declined to comment. It is separate from capital already committed: a completed $30 billion OpenAI stake and a record $11.1 billion bond sale. SRN News

One More Thing · The 2026 Nobel Prizes

Stepping outside AI and healthcare for a moment, this year's laureates.

Field Winner(s) For what
Medicine Deisseroth, Hegemann, Nagel Optogenetics, using light to switch neurons on and off
Physics Francis Halzen The IceCube observatory and the first high-energy cosmic neutrinos
Chemistry Kagan and Soai Non-linear effects and autocatalysis in asymmetric synthesis
Literature Anne Carson A bold, inventive reinvention of the classical tradition
Peace Navi Pillay Efforts to promote peace and international law
Economics Hall and Woodford Microeconomic foundations of macroeconomic stabilization

NobelPrize.org