The State of Fitness Tracking in 2026

Executive Summary

Methodology

This report synthesizes publicly available data from the following source categories, covering the period approximately 2020–2026:

• ›Peer-reviewed studies indexed in PubMed/MEDLINE covering HRV, recovery science, and wearable validation (n = 18 studies cited)
• ›Market research from Grand View Research, Statista, IDC, and Counterpoint Research on wearable device shipments and market size
• ›Public statements, press releases, and investor materials from Apple, Garmin, Whoop, Oura, and Google
• ›Developer documentation and API references for Apple HealthKit, Google Health Connect, and Garmin Connect IQ
• ›Systematic reviews of digital health and mHealth intervention effectiveness in clinical and sports science literature

Note: All market size figures carry inherent uncertainty. Different research firms use different methodology and segment definitions. Where figures diverge materially, both are cited. No Cora proprietary user data is included in this report — all findings are based on public sources.

Section 1: Wearable Adoption in 2026

Market size

The global wearable fitness technology market — broadly defined to include smartwatches, fitness bands, and smart clothing with biometric sensors — has grown substantially since 2019. Grand View Research valued the broader wearable healthcare market at approximately $18 billion in 2020^[1] and projects continued growth. Statista estimates global smartwatch shipments reached approximately 150 million units in 2023, with fitness trackers adding another ~60–80 million units^[2]. IDC data for 2024–2025 suggests total wearable shipments (wrist-worn devices) approaching or exceeding 200 million units annually^[3].

The total installed base — devices actively in use — is substantially larger. IDC estimated approximately 553 million active wearable devices globally in 2023^[4]. This figure includes legacy devices still in circulation. The addressable market for fitness data applications is therefore far larger than annual shipment numbers suggest.

Market share by brand

Apple Watch dominates the smartwatch segment by revenue. Counterpoint Research data from 2023–2024 consistently places Apple at approximately 28–32% market share of global smartwatch shipments^[5]. Samsung and Garmin hold the second and third positions in most measures, though their relative standings vary by geography. Garmin is disproportionately strong in fitness-focused segments: among triathletes, cyclists, and runners, Garmin device penetration is substantially above its overall market share^[6].

In the dedicated health-wearable segment (subscription-model devices), Whoop and Oura have both reported significant user growth without disclosing precise installed base figures publicly. Whoop announced crossing 1 million members in 2022 and has grown since^[7]. Oura Ring Generation 3, released in late 2021, drove substantial adoption in the sleep-focused segment.

Demographic breakdown

Wearable adoption skews younger and higher-income, though the gap is closing. Pew Research Center data (2020) found 21% of American adults wore a smartwatch or fitness tracker daily^[8]. By age cohort, 18–29-year-olds were the most likely adopters, followed closely by 30–49-year-olds. Usage among 50+ adults has grown, driven partly by health monitoring applications (ECG detection, fall detection, blood oxygen monitoring) that appeal to older users.

Gender balance varies by device type. Fitness bands and sleep trackers (Oura, Fitbit) show close to even gender splits. Traditional GPS sports watches (Garmin, Polar) skew approximately 60–65% male^[6]. Apple Watch sits near parity.

Section 2: What Users Track — And What's Growing

The fitness tracking ecosystem has expanded from a narrow set of metrics (steps, calories, active minutes) to a much richer data layer. Understanding which metrics are gaining mindshare matters for both product development and for interpreting research on wearable effectiveness.

From steps to recovery

Step counting was the defining metric of the first generation of fitness trackers (Fitbit, Jawbone, 2008–2015). It remains widely used but has been displaced as the primary feature driver. According to analysis of app store reviews and wearable company marketing materials, the most frequently emphasized metrics in 2025–2026 product releases are sleep quality, HRV, and recovery scores — metrics that did not appear in most consumer wearables before 2018.

Heart rate zone training — particularly Zone 2 and threshold work — has gained significant mainstream attention, driven partly by long-form media coverage of endurance sports and the influence of researchers like Iñigo San Millán and Peter Attia. Searches for "Zone 2 training" grew substantially from 2020 to 2024 per Google Trends data, and device manufacturers have responded: Garmin, Apple, Polar, and Cora all introduced Zone 2 detection or guidance features in the 2022–2025 period.

The sleep data inflection

Sleep tracking deserves specific mention. Apple Watch added sleep tracking in watchOS 7 (2020). Prior to that, Fitbit and Oura were the primary consumer sleep tracking devices. Since 2020, every major wearable platform now offers sleep stage detection (REM, deep, light). This creates a situation where a meaningful percentage of wearable users have 2–4 years of continuous nightly sleep data, a longitudinal dataset with significant research value.

The combination of sleep quality data with training load data — exactly the pairing that Cora's Body Charge system uses — represents the next frontier in actionable wearable data.

Section 3: HRV Trends and Normative Data

Heart rate variability has transitioned from a clinical cardiology metric to a mainstream consumer fitness signal in roughly a decade. Understanding what normal HRV looks like — and how it changes with age and fitness — is foundational for any application using HRV-based recommendations.

Published baselines

The seminal normative reference works are Shaffer & Ginsberg (2017)^[9], which provides a comprehensive review of HRV metrics and their clinical correlates, and Sammito et al. (2016)^[10], which established age-stratified normative RMSSD values in a healthy German adult population (n = 3,964). Both papers establish that RMSSD — the metric reported by Apple Watch, Garmin, Whoop, and Oura — is the most reliable short-term HRV metric for autonomic nervous system assessment.

Age-related decline

HRV declines with age as a function of reduced autonomic nervous system flexibility. This is well-established: the Sammito data shows a roughly linear decline of approximately 8–10 ms per decade from age 20 to 65 in healthy sedentary adults. Importantly, this age-related decline is modifiable. Endurance-trained older adults (50–70) consistently show HRV values 20–40% higher than age-matched sedentary controls^[11], suggesting that regular aerobic exercise partially offsets autonomic aging.

Fitness-level effects

A 2019 meta-analysis by Dong (2016)^[12] covering 72 studies and over 5,700 participants found that exercise training consistently increases resting HRV, with effect sizes larger for aerobic training than resistance training. The magnitude of improvement correlates with training volume and VO2 max improvement. This relationship is why HRV serves as a useful proxy for cardiovascular fitness trajectory, not just day-to-day recovery status.

For practitioners: if an individual's 30-day HRV trend is rising, they are almost certainly improving aerobic fitness. If it is declining despite consistent training, overtraining, illness, or life stress is the likely explanation.

Section 4: The Evolution of Recovery Science

From RPE to data-driven recovery

Recovery assessment has evolved through several generations of methodology. The original approach — perceived exertion (RPE) and questionnaire-based wellness scores (POMS, REST-Q) — relies on subjective self-report. These measures have validity but are subject to compliance issues and poor sensitivity at detecting recovery deficits before they affect performance.

Objective biomarkers emerged in sports science labs in the 1990s and 2000s: blood lactate, cortisol/testosterone ratios, and inflammatory markers (CRP, IL-6) are sensitive to overtraining but require lab infrastructure that is impractical for daily monitoring. HRV bridged this gap: it is objective, non-invasive, measurable in two minutes at home or automatically during sleep, and has a substantial research base validating it as a proxy for autonomic recovery status.

The controlled evidence for HRV-guided training

Four particularly strong controlled trials deserve attention for practitioners evaluating this approach:

The proliferation of proprietary recovery scores

Every major wearable platform now offers a composite "recovery score" that synthesizes multiple inputs into a single daily number. This is arguably the most consequential product development in consumer fitness technology since the step counter.

Note: Methodologies are proprietary. These descriptions are based on public documentation and press materials. Direct numeric comparisons across platforms are not valid.

The fragmentation of these scores is both a user education challenge and a product opportunity. A user who switches from Whoop to a Garmin cannot port their historical baseline or meaningfully compare the two scales. Applications that sit above the hardware layer — reading from Apple Health or Google Health Connect — are positioned to provide the platform-neutral synthesis that individual device makers cannot.

Section 5: AI in Fitness Apps — The GenAI Wave

The integration of large language models and generative AI into consumer fitness applications accelerated dramatically in 2023–2025. Prior to 2023, "AI" in fitness apps largely meant rule-based recommendation engines, statistical anomaly detection, or narrow ML classifiers (e.g., workout type recognition from accelerometer data). The release of GPT-4 and comparable models triggered a wave of conversational AI features across the category.

Major GenAI product releases, 2023–2025

• Fitbod AILaunched adaptive AI workout adjustments using session feedback and Apple Health data. Targets intermediate lifters looking for personalized strength programming.
• FutureCombines human coaches with AI workout generation. Uses biometric data to inform coach recommendations. Raised significant funding citing AI differentiation.
• FreeleticsReleased a conversational AI coach interface in 2024, integrating with its existing adaptive training engine.
• Whoop CoachLaunched in 2023, powered by GPT-4. Allows users to ask natural language questions about their biometric data. One of the first mainstream wearable companies to ship a GenAI interface.
• CoraAI fitness coach that integrates Apple Health data (HRV, sleep, workouts, nutrition) to generate personalized daily training recommendations and conversational coaching. Focuses specifically on recovery-aware planning.

Does AI coaching actually work?

The evidence base for digital health coaching — the category that GenAI fitness apps belong to — is substantially stronger than the hype cycle might suggest. A 2023 systematic review in npj Digital Medicine (Linardon et al.)^[17] covering 58 randomized controlled trials found that smartphone app-based interventions increased physical activity by a mean of 1,850 steps per day versus control conditions — a clinically meaningful increase associated with reduced all-cause mortality risk.

A separate 2024 meta-analysis covering conversational AI health interventions^[18] found that chatbot-style interfaces improved adherence to health behaviors by an average of 27% versus static app experiences. The mechanisms cited: personalization, immediate feedback, and reduced friction in reporting.

What is less established is whether AI coaches that incorporate real-time biometric data (HRV, sleep quality) outperform AI coaches that use only self-reported data. This is an active research question as of 2026. Early signals from Whoop, Oura, and Cora suggesting that users engage more with recommendations grounded in their own data are plausible but have not yet been subject to independent RCT-level validation.

Section 6: The Platform Wars — Data Moats and Ecosystem Lock-In

Apple HealthKit — the dominant aggregation layer

Apple HealthKit, launched in 2014 with iOS 8, is the most important health data platform in consumer fitness technology. It acts as a centralized data store on the user's device, aggregating data from the Apple Watch, third-party apps, and manual entries. As of 2025, Apple has not published the total number of HealthKit-connected apps, but the developer ecosystem numbers in the thousands.

HealthKit's significance for fitness tracking is structural: it allows any app with user permission to read HRV, sleep data, workout history, and dozens of other health metrics that a user has accumulated across all their Apple devices and apps. This creates network effects — a user who has 3 years of Apple Health data is locked into the Apple ecosystem not by cost but by data portability friction. Applications like Cora that build on top of HealthKit inherit this longitudinal data without requiring users to re-enter historical information.

Google Health Connect — the Android play

Google Health Connect, launched in stable form in 2023 with Android 14 integration, is Google's answer to HealthKit^[19]. It provides a standardized API for Android apps to share health data without going through individual app-to-app integrations. In principle, it enables the same aggregation benefits on Android as HealthKit on iOS. In practice, as of 2026, Health Connect is earlier in its adoption curve: fewer third-party apps support it, data types are more limited, and historical data portability is constrained.

The two platforms remain siloed. No cross-platform health data standard exists that allows a user to move their HealthKit data to Google Health Connect or vice versa. FHIR (Fast Healthcare Interoperability Resources) is a standard in clinical healthcare but has not been widely adopted as a consumer fitness data exchange format.

Garmin's data moat

Garmin occupies a distinctive competitive position. Its devices are disproportionately popular among serious endurance athletes — the segment that generates the richest training data and has the highest willingness to pay. Garmin Connect, the platform that aggregates this data, has accumulated years of longitudinal training load, performance, and HRV data from a highly engaged user base.

Garmin has historically been reluctant to export data to third-party platforms (though basic HealthKit sync is available). This creates a tension: users who want AI coaching on top of their Garmin data must either use Garmin's own coaching features or accept that third-party AI coaches will have incomplete data.

Whoop's subscription model shift

Whoop pioneered the subscription-only wearable model: the hardware is cheap or free, and users pay a recurring monthly fee for data access and features. This model has been commercially successful and has influenced others (Oura Ring added a subscription tier in 2022).

The model creates a dynamic where Whoop's revenue depends on continued engagement — users who feel they are benefiting from the data will subscribe; those who don't will churn. This incentivizes Whoop to make its recovery score genuinely actionable and to build features (like Whoop Coach) that convert raw data into behavior change. The commercial interest and the user's interest are, in this model, more aligned than in hardware-only businesses.

Section 7: What's Next in 2027 — Cora's Predictions

Caveats: These are informed predictions based on current trajectories. They are hedged appropriately. Technology markets are notoriously difficult to forecast. Treat these as directional signals, not certainties.

Frequently Asked Questions

Citations

Cite This Report

Related reading from Cora