Google Teams Up With Tech Rivals: Inside the Gemini-Powered Google Health Coach

Google spent May 2026 rebuilding its health strategy from the ground up: retiring the Fitbit app, launching a Gemini-powered AI coach, and — notably — going out of its way to work with the wearables it doesn’t make, including the Apple Watch. Here’s what actually launched, what it means that Google is deliberately courting rival hardware, and how it stacks up against OpenAI, Microsoft, and Anthropic’s competing health AI products.

Key Takeaways

• Google rebranded the Fitbit app to Google Health on May 19, 2026, and launched a Gemini-powered AI Health Coach as part of a new Google Health Premium subscription ($9.99/month or $99/year).
• Alongside the software rebrand, Google launched Fitbit Air, a $99.99 screenless tracker, positioning the pairing as a lower-cost challenger to Whoop’s subscription-hardware model.
• Google Health Premium is bundled at no extra cost for Google AI Pro and Google AI Ultra subscribers in more than 30 countries — folding health coaching into Google’s broader AI subscription strategy rather than treating it as a standalone product.
• Google has explicitly designed the coach to work with data from devices it doesn’t make, including Apple Watch, Garmin, and Oura, through Health Connect and a new Google Health API — a deliberate break from the walled-garden approach most hardware makers default to.
• Google Health isn’t launching into an empty market: OpenAI introduced ChatGPT Health in January 2026, Microsoft launched Copilot Health in March, and Anthropic unveiled Claude for Healthcare the same week — all racing toward what Microsoft has called “medical superintelligence.”
• According to April 2026 research from the West Health-Gallup Center on Healthcare in America, roughly one in four U.S. adults — more than 66 million people — have used an AI tool for health information, though only about a third say they trust the accuracy of what it tells them.
• Google, OpenAI, and Microsoft have each stated their health AI products aren’t intended to diagnose conditions or replace doctors, a distinction that matters given that consumer health AI largely falls outside HIPAA’s regulatory umbrella.

What Happened: Google’s New AI Health Coach

On May 7, 2026, Google announced it was retiring the Fitbit app in favor of a new Google Health app, alongside a Gemini-powered AI coach designed to act as a combined fitness trainer, sleep expert, and wellness advisor.

The rollout began May 19 and reached full availability by May 26, touching the accounts of Fitbit’s existing user base along with anyone using a Pixel Watch. It’s a significant move for Google, which acquired Fitbit in 2021 and has spent the years since figuring out how to fold the company’s fitness-tracking expertise into its broader AI ambitions — a strategy that fits into the broader trends reshaping the technology industry this year, as consumer AI companies race to attach themselves to daily health and wellness habits.

According to cnn.com, The AI Health Coach is built on Gemini and designed to summarize a user’s health records, generate personalized weekly fitness plans, and answer natural-language questions about a person’s fitness and medical data. Google has said the coach can flag trends — noting, for instance, that a recent run contributed meaningfully toward a weekly distance goal — and can surface recommendations that include seeking in-person care, such as reasons to see an orthopedic doctor for issues the app’s sensors alone can’t resolve.

Rishi Chandra, general manager of Google Health, has framed the effort around a simple comparison: elite athletes have entire support teams — a nutritionist, a sleep coach, a trainer — and the AI coach is meant to bring some version of that structure to everyone else. It’s a pitch that echoes across nearly every company entering this space right now, from Google to OpenAI to Microsoft, all making some variation of the same argument: that AI can close the gap between what expert, personalized health guidance costs and what most people can actually access.

From Fitbit to Google Health: What Actually Changed

The rebrand is more than cosmetic. Google Health 5.0 reorganizes the app into four tabs — Today, Fitness, Sleep, and Health — and replaces Fitbit’s old daily step and cardio targets with a personalized weekly cardio target. The Cardio Fitness Score has been renamed VO2 Max, and its underlying calculation no longer factors in demographic inputs like height and weight the way the old score did. Existing Fitbit Premium subscribers were migrated automatically to Google Health Premium at the same price point, while new annual subscribers now pay $99 (up from Fitbit Premium’s previous $79.99 rate).

Alongside the software change, Google launched Fitbit Air, a $99.99 screenless fitness tracker built to be worn continuously. Unlike the Pixel Watch, Fitbit Air has no display — a deliberate design choice aimed at reducing notification-driven distraction, and one that puts it in direct competition with Whoop’s screenless, subscription-based model.

Where Whoop charges roughly $30 a month with hardware bundled into the subscription, Fitbit Air’s core tracking works without any subscription at all; only the AI Coach and deeper predictive analysis sit behind the Google Health Premium paywall. For Google Fit users — the older, largely dormant Google fitness app — migration to Google Health is expected to roll out later in 2026, with the standalone Google Fit app being phased out entirely.

The business logic behind bundling matters here too. Google AI Pro and Google AI Ultra subscribers — Google’s general-purpose Gemini subscription tiers — now receive Google Health Premium at no additional cost across more than 30 countries. That’s a meaningful shift in how Google is positioning health coaching: not as a standalone wellness product competing purely on its own merits, but as one more feature inside a broader AI subscription bundle, similar to how AI-driven predictive data analysis has become a bundled feature across unrelated product categories rather than a separately purchased tool.

How the Gemini Health Coach Actually Works

Google has said the coach was built using what it calls the SHARP evaluation framework — shorthand for safety, helpfulness, accuracy, relevance, and personalization — and developed in collaboration with a Consumer Health Advisory Panel made up of clinicians across multiple disciplines, alongside Google’s in-house research and sports-science staff. Google also brought in outside athletic expertise: NBA champion Stephen Curry and his team reportedly worked with Google Health’s own experts to help shape aspects of the coach’s approach to goal-setting and recovery guidance, and a Fitbit Air Special Edition tied to that partnership has also been released.

Functionally, the coach behaves like a conversational layer sitting on top of a user’s aggregated health data — fitness metrics, sleep patterns, and, for U.S. users who opt in, medical records pulled from a connected healthcare provider portal. Users can ask it to build a workout plan in plain language, request explanations of lab results, or ask it to summarize longer-term trends across sleep and activity data. Google has also introduced “Smart Health Links,” a feature intended to let users securely share specific pieces of health information with a doctor or family member without exporting an entire medical history.

It’s worth being precise about the boundaries here, because they matter for how the product should actually be used. Google, along with OpenAI and Microsoft, has explicitly stated its health AI tools aren’t intended for medical use or to replace a doctor. In practice, that means the coach is positioned to help users understand and organize what they already know about their health and to help them prepare for conversations with actual clinicians — not to diagnose new conditions or substitute for professional evaluation.

Playing Nice With Rivals: Apple Watch, Garmin, and Oura

The most unusual part of Google’s strategy is how openly it’s courting hardware it doesn’t sell. Chandra has described the philosophy directly: users who already own and love an Apple Watch aren’t being asked to switch — Google says it wants to meet people on whatever wearable health tracker they’re already wearing, rather than requiring a Pixel Watch or Fitbit device as a precondition for using the AI coach.

Technically, this works through Health Connect (Google’s existing cross-app health data standard), a new Google Health API, and direct Apple Health integration for iOS users, allowing the app to pull in data from Peloton, MyFitnessPal, continuous glucose monitors from Dexcom and Abbott, and — per Google’s stated roadmap — a widening set of third-party wearables including Garmin and Oura devices.

That’s a notably different posture than the tightly closed ecosystems most wearable makers, Apple included, have traditionally maintained. Whether it holds up as Google Health scales — and whether third-party hardware makers remain willing partners rather than reluctant ones — is one of the more interesting open questions in this launch, particularly for a market where autonomous AI agents are increasingly expected to work seamlessly across a person’s entire device ecosystem rather than staying siloed to a single brand.

Notably, the Google Health Coach’s initial rollout is narrower than the cross-device pitch suggests: it launched first for eligible Fitbit and Pixel Watch users specifically, with support for other hardware, including Apple Watch, Garmin, and Oura data sources, described as “coming soon” rather than available at launch. Anyone can download the app immediately, but people without a supported device are placed in a queue to be notified once broader compatibility ships.

Where Google Actually Stands in the Wearables Market

Google’s position in the hardware side of this race is genuinely modest. According to IDC’s Worldwide Quarterly Wearable Device Tracker, Apple held roughly 21% of the global wearables market in the fourth quarter of 2025, with Xiaomi at about 10.3%, Huawei at roughly 8.0%, and Samsung at approximately 7.7%. Google and Fitbit combined don’t register as a named entry in IDC’s top-five vendor rankings at all — a more pointed way of putting it than simply “lagging behind,” since it means Google isn’t just behind the leaders, it’s currently outside the tracked leadership tier entirely on unit shipments.

That gap is precisely why Google’s public framing has shifted away from competing on hardware volume and toward positioning itself against AI-native rivals instead. The company appears to have made a deliberate strategic choice: rather than trying to out-ship Apple, Samsung, Xiaomi, and Huawei in smartwatch units, Google Health researchers are focused on building software and AI capabilities that work regardless of which wearable a person already owns — effectively trying to win the software layer of health tracking while ceding, for now, the hardware volume race.

The Real Competition: ChatGPT Health, Copilot Health, and Claude for Healthcare

Google’s real competitive set right now isn’t Apple, Samsung, or Huawei — it’s the other AI labs racing to become what Microsoft’s AI health lead has called “the first steps toward medical superintelligence.” OpenAI introduced ChatGPT Health in January 2026 as a dedicated, more heavily privacy-isolated space within ChatGPT that lets users connect medical records and wellness apps, including Apple Health and MyFitnessPal, so responses can be grounded in a person’s own health context. OpenAI has said health-related questions are already among the most common uses of ChatGPT, with over 230 million people globally asking health and wellness questions on the platform every week.

Microsoft followed in March 2026 with Copilot Health, a secure space inside Copilot that aggregates wearable data, electronic health records from more than 50,000 U.S. healthcare providers, and lab results through a system called HealthEx, then serves expert-reviewed content sourced partly from Harvard Health.

Microsoft says Copilot already fields more than 50 million health-related questions a day across its consumer AI products. Anthropic, notably, unveiled Claude for Healthcare the same week as Microsoft’s launch, rounding out a genuinely crowded field of major AI labs all pursuing consumer health simultaneously within a matter of weeks of each other in early 2026 — a competitive dynamic not unlike the broader pattern of how smaller AI labs are competing against the largest players by finding specific, defensible niches rather than trying to match them feature-for-feature everywhere at once.

Interestingly, Apple — the market’s dominant hardware player — is the one company conspicuously behind on the software side. Multiple reports indicate Apple delayed its own AI health coach, reportedly codenamed “Project Mulberry,” along with an associated Apple Health+ subscription, choosing instead to first rebuild the reliability of its underlying biometric sensors, including its Apple Watch heart-rate tracking engine, before layering AI coaching on top. That delay is arguably the single biggest opening competitors like Google currently have: the wearables market leader is, for now, sitting out the AI coaching race it will eventually have to enter.

Why Health Institutions Are Urging Caution

Medical institutions have been consistent in flagging the limits of this entire product category. Health systems including Mayo Clinic and the Duke University School of Medicine have warned that AI tools can produce inaccurate output and lack the ability to reason through a case or order and interpret tests the way a trained clinician can. That caution sits alongside a broader wave of emerging medical technologies entering clinical and consumer use faster than regulatory frameworks can fully evaluate them, and it’s part of why the involvement of pharmaceutical and clinical experts in reviewing these products before launch has become a standard, expected step across the industry, not an optional add-on.

There’s also a regulatory gap worth understanding. Consumer-facing AI health platforms like Google Health Coach, ChatGPT Health, and Copilot Health generally operate outside HIPAA’s traditional framework, since the companies running them aren’t “covered entities” or “business associates” under that law.

That means the primary legal protections governing this category currently come from a patchwork of state privacy laws, the FTC Act, and the FTC’s Health Breach Notification Rule, rather than the more comprehensive federal healthcare privacy regime most people assume applies whenever “health data” is involved. Legal analysts have also noted that several state medical boards are actively examining whether tools capable of interpreting lab results and flagging physiological patterns risk crossing into the unauthorized practice of medicine — an open legal question none of these companies has fully resolved.

How Many People Already Use AI for Health Advice

The demand side of this story is already substantial, regardless of how any individual product performs. According to an April 2026 survey from the West Health-Gallup Center on Healthcare in America, roughly one in four U.S. adults — more than 66 million people — report having used an AI tool or chatbot for health-related information or advice, based on a nationally representative survey of more than 5,500 adults. Most use is supplemental rather than a replacement for care: majorities say they use AI to research symptoms before an appointment or to better understand a diagnosis afterward, and common motivations include wanting a quick answer, curiosity, and — for a meaningful minority — cost or access barriers to seeing an actual provider.

That last point deserves attention. Among survey respondents who’d used AI for health advice, 14% said the information or advice they received led them to skip a doctor’s visit altogether in the prior 30 days — which, projected across the full adult population, represents an estimated 14 million U.S. adults.

Trust in the technology remains genuinely split: roughly a third of recent AI health users say they trust the accuracy of what they’re told, another third distrust it, and the remainder land somewhere in between. That ambivalence is a useful corrective to breathless coverage of this category — people are using these tools in large numbers, but without much confidence that what they’re being told is reliable, which is precisely the gap institutions like Mayo Clinic and Duke have been warning about.

The economic pressure behind AI health adoption is also worth naming directly. Younger adults research more independently before visits, but income is the more revealing variable: among adults earning less than $24,000 annually, roughly a third say they’ve turned to AI specifically because they couldn’t afford to see a doctor, compared with a small fraction of adults in higher income brackets.

Framed that way, products like Google Health Coach aren’t just a wellness convenience for people who already have good healthcare access — they’re increasingly functioning as a stopgap for people who don’t, a dynamic that raises its own set of questions about whether AI coaching is closing healthcare access gaps or, in some cases, papering over them with imperfect information.

What Happens to Your Health Data

Google has stated explicitly that Fitbit and Google Health users’ data won’t be used for Google Ads, and that people retain control over what gets saved, shared, or deleted, along with the ability to disable specific features individually. Microsoft has made a similar commitment for Copilot Health, stating health conversations aren’t used to train its models and sit in an isolated space with additional privacy layers separate from general Copilot use. OpenAI has made comparable claims for ChatGPT Health, describing purpose-built encryption and conversation isolation specific to health data.

Those commitments matter, but they don’t eliminate the underlying risk. Centralizing medical records, wearable biometrics, and ongoing AI health conversations inside a single consumer platform creates, by definition, a higher-value target for bad actors than any of those data sources would represent separately — a concern that echoes broader worries about protecting sensitive personal data as more of daily life moves through always-on AI platforms, and one that sits alongside wider cybersecurity challenges facing sensitive data across the tech industry generally.

It’s also worth remembering that AI-generated content isn’t just a data-security risk; the same underlying generative technology has already been implicated in unrelated harms like AI voice cloning scams, a reminder that the tools capable of synthesizing helpful health guidance are, in other contexts, capable of being misused — a tension the entire industry is still working through, not something unique to any single company’s product.

What This Means Going Forward

Zooming out, this launch fits into a wider pattern: nearly every major AI lab — Google, OpenAI, Microsoft, and Anthropic — has now shipped a dedicated consumer health product within roughly a five-month window in 2026, each betting that health is one of the categories where AI assistants can move from novelty to habitual daily use. That’s a notably faster, more concentrated rollout than most previous AI product categories have seen, and it raises real questions about how AI is reshaping career paths for the human nutritionists, sleep coaches, and personal trainers these products are explicitly designed to approximate at scale, per Chandra’s own framing of the opportunity.

It also raises questions about how this plays out globally, not just in the U.S. market these launches have primarily targeted so far. AI-driven health tools may find their fastest growth not in already well-served healthcare markets but in regions with significant gaps in clinician access — a dynamic already visible in broader patterns of AI adoption in markets like India, where access constraints are more severe than in the U.S. and the appeal of a free or low-cost AI health entry point is correspondingly higher.

The same underlying pattern — AI stepping into gaps left by expensive or hard-to-access professional services — has already played out in other sectors, including AI chatbots transforming fintech, where automated tools have taken on advisory roles once handled exclusively by human professionals.

The competitive questions worth watching from here are fairly concrete: whether Apple’s delayed health coach, once it ships, resets the competitive baseline given Apple’s dominant hardware position; whether Google’s cross-device compatibility promise to

Apple Watch, Garmin, and Oura users actually materializes on the timeline implied at launch; and whether any of these products meaningfully move the trust numbers from West Health-Gallup’s survey, given that unreliable trust, not lack of usage, currently looks like the biggest structural weakness across the entire category — including questions that echo broader industry debates about how major AI labs are governed and held accountable when their products touch something as consequential as personal health decisions, and disputes like Musk’s legal fight with OpenAI that continue to shape how these companies are perceived even outside the health category specifically.

Frequently Asked Questions

What is Google Health, and how is it different from the old Fitbit app?

Google Health is the rebranded version of the Fitbit app, launched May 19, 2026. It reorganizes the interface into four tabs (Today, Fitness, Sleep, Health), replaces daily targets with personalized weekly goals, and adds a Gemini-powered AI Health Coach as a premium feature not previously available in Fitbit.

Does the Google Health AI Coach work with an Apple Watch?

Google has said this is a goal, using Apple Health integration and its Health Connect standard, but at launch the coach was available first only to eligible Fitbit and Pixel Watch users. Support for Apple Watch, Garmin, and Oura is described as “coming soon” rather than available immediately.

How much does Google Health Premium cost?

Google Health Premium costs $9.99 per month or $99 per year. It’s included at no additional cost for subscribers to Google AI Pro or Google AI Ultra in more than 30 countries.

Is Google’s AI Health Coach a replacement for a doctor?

No. Google, along with OpenAI and Microsoft, has explicitly said its AI health products aren’t intended for medical use or to replace a doctor. They’re designed to help users understand their own health data and prepare for conversations with actual clinicians.

How does Google Health Coach compare to ChatGPT Health and Copilot Health?

All three launched within a few months of each other in 2026 and serve a similar function: aggregating wearable, medical record, and lab data, then using AI to generate personalized health insights. They differ mainly in data-source breadth, pricing structure, and which company’s broader AI ecosystem they’re bundled into.

Is Google’s health data used for advertising?

Google has stated that Fitbit and Google Health users’ data will not be used for Google Ads, and that users retain control over what data is saved, shared, or deleted.

How many people already use AI for health advice?

According to an April 2026 West Health-Gallup survey, roughly one in four U.S. adults — over 66 million people — have used an AI tool or chatbot for health-related information or advice, though only about a third of recent users say they trust its accuracy.

Why is Google working with rival companies like Apple instead of just competing on hardware?

Google holds a comparatively small share of the global wearables market and doesn’t register in IDC’s named top-five vendor rankings. Rather than competing directly on device shipments against Apple, Samsung, Xiaomi, and Huawei, Google appears to be positioning its AI coaching software to work across whatever wearable a person already owns.

What is Fitbit Air?

Fitbit Air is a $99.99 screenless fitness tracker launched alongside Google Health, designed for continuous wear without a display, positioned as a lower-cost alternative to Whoop’s subscription-based tracking model.

Are health institutions concerned about AI health coaching tools?

Yes. Institutions including Mayo Clinic and Duke University School of Medicine have warned that AI can be inaccurate and lacks the reasoning and diagnostic capability of a trained medical professional, which is part of why companies including Google explicitly frame these tools as informational rather than diagnostic.

Conclusion

Google’s rebrand from Fitbit to Google Health, anchored by a Gemini-powered AI coach, is a genuinely significant strategic shift — not because Google suddenly leads the wearables hardware market (it doesn’t), but because it represents a bet that the software and AI layer sitting on top of health data matters more than which company made the device collecting it. Actively courting Apple Watch, Garmin, and Oura users, rather than trying to force everyone onto Google-made hardware first, is an unusually open posture for a company competing in consumer devices, and it puts real pressure on Apple’s own, still-delayed AI health coaching plans.

At the same time, this launch didn’t happen in isolation. Google is one of at least four major AI labs — alongside OpenAI, Microsoft, and Anthropic — that shipped dedicated consumer health AI products within roughly the same few months in 2026, entering a market where usage is already substantial but trust remains genuinely mixed.

The clinical caution voiced by institutions like Mayo Clinic and Duke, combined with a regulatory landscape that hasn’t fully caught up to consumer health AI, means the real test for Google Health — and its rivals — won’t be how many people download the app. It’ll be whether the guidance these tools provide earns the kind of trust that turns curious first-time users into people who actually rely on it.