|
||||||||||||||||||||||
|
“The pricing from some of the other labs is very extreme and has very high margins... We think there’s a real ability to offer frontier or very high-level intelligence at a much more affordable cost.” Mark Zuckerberg, CEO, Meta, in a Bloomberg interview ahead of the Muse Spark 1.1 launch, pricing Meta’s new API at roughly a quarter of OpenAI and Anthropic. Briefs.co |
||||||||||||||||||||||
| By The Numbers | ||||||||||||||||||||||
|
HEALTHCARE SECTOR DASHBOARD The vitals we're watching | July 13, 2026
Sources: CDC NHIS, CMS 2026 Trustees Report, BLS, Peterson-KFF, Kaufman Hall, HHS/AP, Yahoo Finance. Teal favorable, red unfavorable, slate neutral. Figures current as of July 2026. |
||||||||||||||||||||||
| Provider Update | ||||||||||||||||||||||
Williamson Health, an independent Tennessee system, voted to sell to Ascension Saint Thomas, choosing a $700M bid plus $250M in capital over ten years. Ascension beat a matching HCA offer and an Optum management-only proposal. The deal closes in 2027 or 2028. Independent systems keep folding into national players. Becker’s Hospital Review |
||||||||||||||||||||||
In a UK study, ChatGPT 5.5 and Gemini Pro each scored 94% on Neurology Specialty Certificate Examination questions, while Claude Opus 4.6 scored 79.5%. The sample was small and peer review light, and the models predate current versions. For clinical leaders, frontier models clear specialist-exam thresholds, though real-world validation lags. Cureus |
||||||||||||||||||||||
El Camino Health’s CIO reports concrete AI returns: OR-scheduling tools added a couple million dollars, radiology AI flagged 700 incidental lung nodules in three months, and ambient documentation cut chart-closing to under 24 hours. EHR AI use jumped 500%. For provider leaders, AI ROI is moving from promise to proof. Becker’s Hospital Review |
||||||||||||||||||||||
| Payer Update | ||||||||||||||||||||||
In NEJM Catalyst, Venrock’s Bob Kocher and USC co-authors argue AI will likely raise US health spending near term. Under fee-for-service payment and consolidated markets, efficiency gains get spent, not saved. Only value-based payment and policy reform bend the curve. A rare case cutting against the author’s own interests. NEJM Catalyst |
||||||||||||||||||||||
Medicare’s 2026 Trustees Report moved the Part A trust fund’s depletion date up a quarter to 2033, when it could cover 89% of benefits. The 75-year deficit widened to 0.56% of payroll. GLP-1 and specialty drugs drove Part D higher. Medicare Advantage now covers 51% of the program’s 69M beneficiaries. CMS |
||||||||||||||||||||||
New federal data shows ACA enrollment fell by 2.6 million in February versus a year earlier, the first full 50-state view since enhanced premium tax credits expired January 1. Ohio and Oklahoma each lost over 32%. Analysts cite doubled premiums; HHS attributes much of the drop to a fraud crackdown. AP |
||||||||||||||||||||||
| Investment in Digital and AI | ||||||||||||||||||||||
|
THE AI MARKET SCOREBOARD Market cap and 2026 YTD performance, as of July 12, 2026
Figures as of July 12, 2026 and subject to change; markets were volatile through June. YTD reflects 2026 price performance; SpaceX is shown since its June 12 IPO at $135. Sources: CompaniesMarketCap, StockTitan, Yahoo Finance, company filings. |
||||||||||||||||||||||
Done Global’s founder Ruthia He received six years and its clinical president two years for a $90M scheme. The subscription telehealth firm pushed 37 million Adderall pills, paying clinicians to sign prescriptions in seconds, with refills continuing after patient deaths. Insurers were defrauded of $12.3M. Warning about prescribe-fast telehealth economics. US DOJ |
||||||||||||||||||||||
OpenAI’s GPT-5.6 Sol nearly matched Claude Fable 5 on benchmarks at $1.04 per task, roughly one-third the cost. Meanwhile Microsoft is swapping OpenAI and Anthropic out of Copilot for its cheaper in-house MAI models, aiming to eliminate third-party costs. Premium AI is being squeezed on price from every direction. The Decoder: Sol · The Decoder: Copilot |
||||||||||||||||||||||
Chinese open-weight models have drawn over 30% of US companies’ token use on OpenRouter every week since February, peaking at 46%, up from an 11% yearly average. They run 60% to 90% cheaper than leading US models. Cost, not capability, is driving the shift toward good-enough models. AI Weekly |
||||||||||||||||||||||
|
How was this week’s newsletter?
|

