Before comparing how these companies approach healthcare AI, here’s the essential context on the wider healthcare and life sciences ecosystem:
If you haven’t already seen it, I mapped out what the new healthcare system will look like in this post and article (where I dive deeper into it).
Before looking at each of these companies separately, you need to understand that Healthcare 2.0 will be decentralised. Analyse the map I created and understand how it all flows. The patient will be at the core, and the true moat will be in empowering that patient with an LLM and data that flows freely across the entire ecosystem spanning healthcare and life sciences. Here is the map:
Decentralised Healthcare 2.0
Once you grasp that, you’ll see why some takes framing "OpenAI vs. Anthropic vs. Amazon" as a winner-takes-all battle are missing the point. Each of these companies is targeting different parts of this Healthcare 2.0 map. While they have overlapping ambitions, they possess distinct moats and limitations.
Here is how they are carving out their territories in the new Healthcare OS:
OpenAI
The Strategy: Leveraging massive patient distribution to converge consumer pull with enterprise governance.
The Moat: 40M+ daily users asking medical questions. Their priority is empowering their distributed (not localised) patients, which is why they are betting heavily on bringing AI directly to the patient.
The Nuance: Many argue OpenAI is purely "consumer" while Anthropic is "enterprise." This is wrong. While OpenAI has the consumer distribution advantage, they are aggressively packaging an enterprise layer too. I have been predicting they will soon merge these sides, creating a seamless network effect and interoperability marketplace.
The Execution: Patient Layer - ''ChatGPT Health'': Through partners like b.well Connected Health, they are linking EHR data directly to patients, connecting the PHR with lifestyle apps like Apple Health, and leveraging that massive user base. Health System Layer- ''ChatGPT for Healthcare,'': They are building out AI Clinical Decision Support and workflow tools (discharge summaries, patient instructions, prior auth). This also includes governance: They are introducing centralised access and visibility (SAML SSO, SCIM), which I view as the early shape of "Clinical AI Governance as a Service," starting first around their own workspace.
The Limitations: OpenAI’s enterprise offering currently lacks deep healthcare provider infrastructure and workflow capabilities compared to Anthropic. However, partnerships with the likes of Cedars-Sinai and Stanford Medicine Children's Health will help them catch up quickly. On the consumer side, as a UK resident, I don’t have access to ChatGPT Health, but looking at reviews from others online (like Myoung Cha's commentary and Josh Mandel, MD post ), the patient-facing interface still struggles with reliability, accuracy and safety - baring in mind it's of course non-HIPAA compliant. OpenAI's approach remains: build fast and ship fast.
Anthropic
The Strategy: The "Infrastructure First" wedge.
The Moat: Enterprise-grade AI expertise. Because they lack OpenAI’s consumer distribution advantage, their wedge is enterprise-first: tighter governance, clearer access boundaries, and a posture built specifically for health systems.
The Execution: Infrastructure & Protocols: Anthropic is positioning Claude as the responsible medium for AI scaling. Their Model Context Protocol (MCP) natively connects with ICD-10 registries, PubMed, FHIR standards, and CMS coverage data. Patient Layer: They are utilizing HealthEx to link patients to their EHR data and interface with lifestyle apps, but with tighter guardrails and explicit patient control over access compared to competitors. Workflows: They are focusing on "Agent skills" as primitives - automating prior auth workflows and appeals, and plugging directly into EHRs and payer systems.
The Comparison: Unlike OpenAI’s broad approach, Anthropic is HIPAA-ready by design - for both enterprise and consumer interfaces. They are effectively building the "infrastructure of healthcare operations" with deep informatics utility, rather than just raw model capability.
The "Consumer vs. Enterprise" Myth
I’ve seen plenty of takes suggesting this is a binary choice: OpenAI for consumers, Anthropic for enterprise. I disagree.
In the new Healthcare OS, these two layers converge. Patient-facing access and health-system controls are part of the same stack. Both companies are aiming for this convergence, just from different starting points.
OpenAI is doubling down on patient pull-through while building the enterprise tools (validating my prediction of their playbook strategy from 5 months ago here).
Anthropic is using enterprise trust to reach patients within the enterprise's healthcare parameters.
Ultimately, the network effect drawn into my map - interoperability - will be the true moat.
Amazon One Medical
The Strategy: Vertical Integration and the "Closed-Loop Ecosystem."
Many takes suggest Amazon’s approach is "better" than ChatGPT or Claude. Again, this undermines where the ecosystem is heading. Amazon has a fundamentally different playbook.
The Moat: Distribution & Economics. Amazon owns the physical distribution (One Medical) and the pharmacy network. This allows them to treat AI as a loss leader: they don't need to monetize the model itself (via tokens or subscriptions) like their competitors. Instead, they can provide the intelligence for "free" to drive revenue through membership retention and pharmacy transactions.
The Difference: While Anthropic and OpenAI are building tech for other healthcare systems and patients to use, Amazon is building tech around its OWN healthcare ecosystem to create a proprietary Healthcare OS.
The Execution: Agentic AI: They are moving beyond simple chat to Agentic AI - models that don't just answer questions but execute actions, such as real-time appointment booking and prescription auto-fills. Infrastructure: Built on AWS Bedrock, they use strict "red line" guardrails - essentially clinical protocols as code - that automatically stop the AI from treating complex issues and route the patient to a human provider. It is a closed loop: Consumer + Healthcare Provider + Infrastructure all owned by Amazon.
The Current Trajectory
We are moving toward a decentralised care model where Consumer + Healthcare Providers + Infrastructure connect seamlessly.
OpenAI is fighting to mature its infrastructure to match its massive consumer reach.
Anthropic is leveraging its "safe infrastructure" reputation to become the backbone of hospital operations and clinical knowledge.
Amazon is proving the model by building it vertically within its own walls.
The map below visualizes where each of these companies is aiming to fit into the new healthcare operating system today:
Where OpenAI, Anthropic & Amazon fit in Decentralised Healthcare 2.0
The relevance gap
In May 2025, I flagged a warning & pivotal shift here: patients were quietly leveraging AI to bridge accessibility gaps and bypass traditional clinical pathways. As I expanded on in this post, healthcare leaders must act now to formalize pathways that start at home, meeting patients exactly where they are.
A common rebuttal at the time was that health systems “can’t compete with OpenAI.” That missed the point. Once patients adopt these tools, the provider’s role must evolve around that behaviour.
We are now feeling the friction of this shift as the centralised model buckles under new demands. Delay is no longer just an opportunity cost; it is an existential risk. Every healthcare leader must now pause and confront the reality:
Are we creating safe, governed pathways that acknowledge patient-AI use, or are we ignoring the market reshaping around us?
Patient-facing AI is no longer a 'nice to have' - it is the price of admission. Whether you build, partner, or govern, the mandate is clear: operationalise these capabilities now, or risk irrelevance in a decentralised future.
Hi - I’m Dr. Youssef Aboufandi, a physician and digital health consultant. I support healthcare and technology organisations with AI adoption and digital transformation strategy, with a focus on market intelligence and design thinking. If you found this article useful, consider subscribing to the newsletter. If you’d like to work together, message me on LinkedIn.
Subscribe to this newsletter for weekly market intelligence on AI in healthcare.
I’ll see you next Tuesday with the weekly newsletter - covering what’s shaping AI in healthcare.
- Dr. Aboufandi
📩 If something here stood out, hit reply - I read every message.