Three companies are fighting to become the default clinical reference tool for working physicians. Each has a fundamentally different business model, trust architecture, and bet on how clinical AI wins. ChatGPT for Clinicians (launched today) adds a fourth combatant — but this war has been building for months.
Three Business Models
OpenEvidence: The Pharma Advertising Model. Free. Revenue from pharmaceutical advertising during loading screens at $70-150+ CPMs. $100 million annualised revenue by January 2026. $12 billion valuation. ~$700 million total funding from GV, Sequoia, NVIDIA, Kleiner Perkins, Mayo Clinic. The bet: physician attention at the point of clinical decision-making is the most valuable advertising real estate in healthcare.
UpToDate: The Expert Subscription Model. $500+/year individual. Institutional licences at substantially higher prices. Part of Wolters Kluwer's €1.5 billion Health division (2024 revenue). 7,600+ expert authors. 13,000+ topics. 10,000+ graded recommendations. The bet: expert curation is worth paying for because AI-generated answers, however fast, cannot match the clinical judgment of specialist authors.
DoxGPT: The Professional Network Model. Free via free Doximity membership. Revenue from pharma advertising and healthcare recruiter services on the broader Doximity platform ($540 million total Doximity revenue, FY2025). 85% of US physicians (3 million+ members). The bet: owning the physician network — where doctors already message colleagues, track CME, and browse jobs — is more valuable than owning the content.
The Trust War
UpToDate (institutional trust). Thirty years. 7,600+ expert authors. Content cited in hospital protocols and clinical governance documents. When a clinical governance committee asks "what was this decision based on?", "UpToDate" is accepted. "ChatGPT" is not.
OpenEvidence (source-controlled trust). Proprietary models on peer-reviewed literature only. No general web sources. HIPAA, SOC 2 Type II. GV, Sequoia, Mayo Clinic backing. Trust through constraining what the model can access.
DoxGPT (physician-verified trust). PeerCheck — 10,000+ named physicians led by Eric Topol and Regina Benjamin. Certified answers with reviewer profile links. "I trust it because Dr [name] reviewed it" — a novel trust architecture unique in clinical AI.
The EHR Integration Race
UpToDate has been in major EHRs for years — incumbent advantage. Microsoft partnership (March 2026) extends this into Dragon Copilot (600,000+ users), Microsoft 365 Copilot, and Teams. Abridge scribe partnership. UpToDate Connect API for third-party integration. Strategy: become the knowledge layer underneath every AI interface.
OpenEvidence is the aggressive challenger — Mount Sinai Epic enterprise-wide deployment (March 2026), Sutter Health Epic collaboration (February 2026). Moving fast to embed inside the EHR where clinicians make decisions.
DoxGPT has no EHR integration. Platform strategy: keep physicians in Doximity. Different distribution bet — Doximity already has the audience.
ChatGPT for Clinicians as Wild Card
Free GPT-5 threatens all three — brand recognition none can match. But it lacks EHR integration (OpenEvidence's advantage), expert curation (UpToDate's advantage), physician verification (DoxGPT's advantage), and physician network distribution (Doximity's advantage).
Most likely outcome: ChatGPT captures administrative and general-research usage. The three specialists retain clinical-decision-support usage where trust and verification matter. Co-existence, not winner-take-all.
What This Means for UK Clinicians
This war is US-focused. OpenEvidence, DoxGPT, and ChatGPT for Clinicians are US-first. UpToDate is global but not UK-guideline-specific. None retrieves NICE guidelines, CKS, or SmPC data.
UK clinicians need NICE/CKS/literature/SmPC — not US pharma-funded AI. iatroX exists precisely for this gap. MHRA-registered, UKCA-marked, UK-guideline-grounded, free, with exam preparation and clinical calculators. The UK market is smaller — but underserved by every major clinical AI, and the MHRA regulatory barrier creates a moat US tools have not crossed.
UK clinicians: the clinical AI war is being fought in the US. Your tool is here. Try Ask iatroX. →