Bullet Thesis: the next $1tn company will be in consumer healthcare

13 Feb 2025

Hi, it’s Michael from firstminute capital. Along with my colleagues Lina, Sam, Lorcan and Adriana, I am investing in European pre-seed and seed B2B software companies like this, this and this. In a previous life I was a foreign correspondent for the Financial Times and the founding editor of Sifted. Here is our latest thesis. Thanks to our VP Steve Crossan for the idea & insights.

Over the past six months, study after peer-reviewed study has shown that large language AI models are, in effect, better doctors than doctors.

In one study, LLMs given real world medical cases diagnosed 90% of them right, when their human counterparts got only 74%. In another, LLMs got 60% right compared to 30% for humans.

Sure, LLMs are not full-blown doctors yet. They can’t do a physical exam or hold the hand of a worried elderly patient, for instance.

But the advances in what they can do on the digital side in recent months have been stunning. Try just loading into ChatGPT some blood test results - the answers are already excellent.

We can see this as part of a wider trend in AI today, which is that AI models can increasingly do the jobs of people at MBA and PhD level, replacing scientists, doctors and lawyers — rather than just lower level desk work like junior sales reps or content marketers.

Agents, by which I mean AI models that can take action, are making this more dramatic. Recently, Stanford researchers created a virtual laboratory where 100 AI scientists—large language models with defined scientific roles—successfully collaborated to design antibody fragments.

What does it mean though that we can have a world class doctor in our pocket? One that may soon be able to make proactive choices for us?

One idea is that it will be the unlock for the next $1tn company, which will at its heart be an AI doctor and medical concierge that lives in all our pockets and becomes the portal from which we access all health services. It will be the Apple, Visa, and Amazon of healthcare.

This has been the vision of many for a long time.

"I dream of a time when people will be able to access what I call 'medical intelligence,' which is the notion that we should be able to access the world's best medical knowledge and not just rely on the one doctor," says Mark Hyman, founder of the Cleveland Clinic Center for Functional Medicine.

But today it might just be possible. Google’s chief health officer Karen DeSalvo, who is leading much of Google’s work on a “personal health” large language models, describes what she sees this year as an “inflection point in AI, where we can see its potential to transform health on a planetary scale”.

And while a trillion-dollar consumer health company might sound crazy, consider that five of the world's ten largest companies are consumer brands (Apple, Amazon, Alphabet, Meta & Tesla). And healthcare remains one of the largest industries without a dominant tech player.

Consumer healthtech today

Some of the big companies that have emerged in consumer healthtech today are:

Wearables companies & brands: Oura Ring (sleep and fitness tracking, valued at $5.2 billion), the Apple Watch (which as a division within Apple saw $37 billion of revenue in 2024)
Personalized nutrition tests: Zoe (valued at $270 million) and Viome (which has raised $186m since launching in 2016).
D2C health companies: Hims ($2.4 billion valuation) and Manual (has raised $60m+) that give us bespoke vitamin and treatment regimes.
Telehealth companies: Doctolib ($6.5bn valuation), Ro ($6.6bn valuation), and Noom ($3.6bn valuation) providing virtual healthcare services so we can see a doctor on demand from our living rooms. In the mental health space there were companies like BetterHealth (acquired for $4.5bn).
Digital wellness apps: Headspace ($1.2bn valuation in 2021) and Calm ($2bn valuation) were the winners from this previous wave of apps that could help with sleep, meditation, or quitting smoking or healthier eating.
Whole body scanning: companies like Prenuvo (raised a $70m Series A), Ezra (which has raised $41 million), SimonMed and in Europe Neko Health (which raised $260m Series B and is backed by Spotify founder Daniel Ek).

All of these companies speak to a few mega themes in consumer healthcare:

Consumer interest in "wellness", longevity and preventative medicine, with more people focused on maintaining their health and living longer.
Decentralization of healthcare, with people increasingly being consumers who are making their own health choices and picking their own services.
Personalization and quantification of healthcare, with more data leading to more personalized solutions.

The problem with all of these companies above, or at least the blocker to them becoming a $1tn company, is that while they are all part of the same bucket of themes, all of them remain "point solutions"—highly specialized services that require consumers to actively seek them out.

Neko is for body scanning, Calm for meditation, Hims for hair loss supplements, Oura ring for sleep tracking - none have broken through to "own" the whole health journey.

As a result, none are anywhere close to becoming the size and scale of an Apple or an Amazon. That is where this new company comes in.

The app to rule them all

What could change all of that is an app that, at its core, is two things:

The primary repository for all customer's medical information, which lives essentially on our phones and devices. See here some early sign of how it might work.
An LLM, and a set of agents, which help users navigate all their healthcare needs based on this data.

On the data side, imagine a world where, when visiting healthcare providers, users can instantly share their complete medical history. After consultations, notes are added directly to their personal medical records on their phone. It not only contains users’ medical records, but everything from the sphere of preventative medicine and "wellness" as well through wearables or at home tests.

On the AI side, the main attraction of the product is that sitting above that data is an AI doctor on a user's phone, which provides continuous, personalized attention that human doctors cannot match. It would:

Order and interpret regular tests, including bloodwork and other diagnostics
Interface with wearables like glucose monitors and Oura rings
Track exercise and nutrition
Serve as an opinionated health co-pilot

The AI would navigate the healthcare system for users—ordering lab tests, recommending specialists, and providing guidance on complex health decisions. For example, if someone had cancer, the AI could help guide within the established systems.

AI transcription and analysis of doctor visits would ensure nothing is missed or forgotten—a crucial feature for anyone who has experienced the challenge of remembering complex medical discussions.

This approach inverts the traditional healthcare model: the system's components (specialists, tests, surgeons) become tools managed by users and their AI co-pilots.

This company could potentially build an Amazon-like marketplace, Visa-like payment system, and Apple-like hardware ecosystem.

Lots of people in the last 10 years have tried to solve this patient data problem by allowing patients to own their own data (e.g Citizen Health), but have faced challenges including the interoperability withing the current healthcare system, but also a lack of incentive for patients to upload their own data. Maybe this AI doctor is the unlock for this?

Showing the way

Showing the way here perhaps are a growing number of startups in the "medical concierge" space, which is in effect a service to help people manage their way through traditional and non-traditional health providers.

Function Health is one new one, boasting a waiting list of 300k people and celebrity investors like Matt Damon, Zac Efron, Kevin Hart, Pedro Pascal and Colin Kaepernick along with Andreessen Horowitz. For $499 per year consumers get access to more than 100 lab tests and interpret them.

Meanwhile Superpower offers a similar service, boasts a waiting list of 100,000 and has raised a $4m pre-seed from Long Journey ventures and angels like Cameron and Tyler Winklevoss, Balaji Srinivasan, and Evan Charles Moore the co-Founder of Doordash.

"We're building the healthcare system we want for ourselves and our loved ones. One that is holistic, data-driven, and can support the problems that have classically been neglected by healthcare: nutrition, sleep, inflammation, gut health, hormones, toxins, and longevity, amongst others," says Superpower co-founder Max Marchione.

This is in effect a digital version of an analogue industry of medical concierge for well off clients, which specialises in connecting their clients to the best possible private medical care. "We're like medical fixers," says Kirsty Ettrick, CEO of London Medical Concierge, "Charlie's Angels for the medical world."

More established companies include Eden Health, which raised $103M, and offers bespoke solutions via employers.

Maybe these are the companies that ultimately pivot their business to becoming an AI-first concierge service. Maybe it's Google, which has long been working on AI doctor product, or Apple, which is a dominant forced thanks to it’s Apple Watch. Maybe it's Neko health, which starts with the full body scanning.

Or maybe - and most excitingly - it's a new company starting today.

The time is now

Whoever builds it, there is maybe a trillion dollar "AI doctor in your pocket / medical concierge" business to be built, and what's exciting is that the time to build it is now. This vision doesn't require technological breakthroughs or consumer behaviour—the essential components already exist or are very close:

LLMs have reached sufficient capability (or are close)
Agentic systems are nearly there
Health monitoring hardware continues to improve
Consumers have shown huge appetite for preventative medicine

Challenges to Address

Several big hurdles remain, of course

Healthcare systems will continue to be complex, and interfacing with them will remain challenging. While an LLM could excel at interpreting testosterone levels and recommending solutions, complex conditions like cancer still require traditional system intervention.
How do you get healthcare professionals to by into this new customer led health journey?
For serious conditions, the AI could help navigate treatment options, recommend clinical trials, and find appropriate specialists—but ultimate treatment must occur within established medical systems.
Regulation about what AI doctors will be able to say and do will need to evolve. Also data privacy is and will always be a huge issue.
LLM hallucinations pose risks—how much control over health decisions can we safely delegate to an AI co-pilot?
AI cannot replicate the human connection and physical touch crucial to healing and well-being. Although work on embodied and AIi and robotics will perhaps change that.

Despite these challenges, I believe the potential of having an AI doctor in your pocket creates an unprecedented opportunity.

The company that successfully builds this health co-pilot and becomes the central source of truth for healthcare could achieve trillion-dollar status while dramatically improving global health outcomes. The opportunity is here—who will build it?