Nobody announced a healthcare AI race. It just started.
In the span of about three months, OpenAI, Anthropic, Amazon, and Perplexity all launched dedicated health products aimed at the same person. You, sitting at home, trying to understand a lab result your doctor spent four minutes explaining before moving on to the next patient.
The timing is not a coincidence. Three in five American adults have used AI for health purposes in the past three months. 230 million people talk to ChatGPT about their health every single week. The moment before you decide whether to call your doctor, book an appointment, or just go back to bed is now the most valuable piece of digital real estate in the industry. Every major AI company wants to own it.
Here’s what each of them is actually building, and why it matters who wins.
OpenAI: The Biggest Audience, the Broadest Play
ChatGPT Health launched January 7 and is the most consumer facing of the four. You connect your medical records and wellness apps, Apple Health, wearables, lab data, and ChatGPT becomes a personalized health assistant that explains your test results, spots trends in your data, prepares questions for upcoming appointments, and translates medical jargon into plain English.
The scale is the story. OpenAI already has 800 million global users and roughly a quarter of them were already asking health questions before the dedicated product even launched. ChatGPT Health is less of a new product and more of an official acknowledgment of what people were already doing anyway. OpenAI is also killing off products like Sora to refocus resources on exactly these kinds of high value use cases.
OpenAI also launched a separate enterprise product called ChatGPT for Healthcare built specifically for hospital and clinic workflows. One product for patients, one for the systems that treat them.
Anthropic: The Enterprise Bet
Claude for Healthcare launched four days later and it’s playing a different game. While ChatGPT Health focuses on patients understanding their own information, Claude is targeting the entire healthcare ecosystem at once. Providers, payers, and patients.
The standout angle is enterprise. Claude connects to the CMS Coverage Database for insurance coverage determinations, ICD-10 for medical coding, the National Provider Identifier Registry, and PubMed for access to 35 million pieces of biomedical literature. The big use case Anthropic keeps leading with is prior authorization, the paperwork nightmare where doctors spend hours submitting documentation to insurers just to get treatments approved.
Anthropic CPO Mike Krieger put it plainly: “Clinicians often report spending more time on documentation and paperwork than actually seeing patients.” Claude for Healthcare is designed to eat that paperwork. If you want to see what Claude can do outside of healthcare, we did a full review of Claude Pro covering the general product.
On the consumer side it works similarly to ChatGPT Health. Connect your records, understand your results, prepare for appointments. But the real bet here is B2B. Anthropic wants to be the AI layer inside the healthcare system itself, not just the thing patients turn to before they interact with it.
Amazon: The Vertical Integration Play
Amazon Health AI is the most structurally different product of the four and probably the most ambitious. It launched in January for One Medical members and expanded in March to Amazon.com and the Amazon mobile app, putting it in front of hundreds of millions of people overnight.
Here’s what makes it different. Amazon doesn’t just answer your question. It does the next thing too. Health AI can explain your lab results, book an appointment, renew your prescription through Amazon Pharmacy, connect you to a One Medical provider for a video visit, and route you to specialists at Rush University and Cleveland Clinic if needed. The entire chain, question, appointment, prescription, specialist, happens inside one product connected to one provider network.
The company that already has your credit card, your shopping history, your Alexa, and your prescription delivery now wants your health records too. The vertical integration play is complete on paper. Whether patients trust it enough to hand over that data is a different question.
Perplexity: The Open Platform
Perplexity Health launched last week and takes the opposite approach from Amazon’s walled garden. Instead of building its own provider network, Perplexity is positioning itself as the interface that sits across everything. Your wearables, your lab results, your health records, regardless of where that data comes from.
The same week it launched, lab testing company Function announced a direct connector letting members pipe lab results straight into Perplexity Health. That’s the model. Become the universal health data layer rather than owning any piece of the actual healthcare delivery.
It’s the lightest touch play of the four. No appointments. No prescriptions. No provider network. Just a smarter way to understand what your health data actually means. Though the trust angle just took a hit since Perplexity is now facing a class action lawsuit alleging they secretly shared user conversations with Meta and Google. Handing them your health data hits different after that.
So Who’s Actually Winning
Nobody yet. And the race is less about who has the best AI and more about who you already trust with the most data.
If you live inside the Amazon ecosystem, Health AI has a structural advantage. The data is already there. If you’re a ChatGPT power user, ChatGPT Health is the path of least resistance. If you work inside a health system, Claude for Healthcare’s enterprise integration is the most serious play for your workflow. If you want to understand your health data across multiple sources without handing everything to one company, Perplexity is the move.
The deeper question nobody is asking loudly enough is what happens when these systems are wrong. All four companies disclaim liability for the accuracy of health outputs. All four say they’re not intended for diagnosis or treatment. All four tell you to see a healthcare professional for serious guidance.
Which is true. But three in five Americans are already making health decisions based on what AI tells them, with or without a disclaimer at the bottom of the screen.
The race to own the first click in healthcare is really a race to own the moment when someone decides whether something is serious. That’s a lot of trust to hand to a system that, legally speaking, isn’t responsible for what it tells you.
