privacydeep-dive

Smart Mattress Privacy: What Your Bed Knows About You

Eight Sleep, Sleep Number, and Withings collect granular nightly biometrics. Here's what that data is, who can access it, and how to protect it.

Every night you climb into bed and your smart mattress gets to work. It measures your heart rate, breathing pattern, skin temperature, toss-and-turn count, and deep sleep duration. By morning, a cloud server somewhere holds an intimate, uninterrupted biometric record of eight hours of your life.

Smart mattresses are sold as sleep optimization tools. The privacy implications are sold as footnotes.

This post covers what smart mattresses actually collect, who holds that data, what they can infer from it, and what you can do to limit your exposure.


The Data Your Smart Mattress Collects

Smart mattress platforms vary in their sensors and capabilities, but the data categories are consistent across the major players.

Vital signs during sleep. Heart rate and heart rate variability (HRV) measured via ballistocardiography — a technique that detects the mechanical vibrations your heart creates through the mattress surface. Respiratory rate measured from the rise and fall of your chest. These measurements are continuous throughout the night, not spot-checks.

Movement and sleep staging. Accelerometers detect body movement. Combined with heart rate and respiratory data, sleep-tracking algorithms infer sleep stages: awake, light sleep, deep sleep, REM. These inferences are imprecise — consumer sleep staging doesn’t match clinical polysomnography — but they generate detailed behavioral profiles.

Temperature. Bed temperature preferences and body temperature fluctuations through the night. Eight Sleep’s system actively adjusts mattress surface temperature based on your sleep stage and individual thermal profile, which means it’s continuously building a model of how your body responds to temperature at different sleep stages.

Presence and occupancy. When you get into bed, when you get up, how often you leave during the night. For dual-zone mattresses, which side of the bed each person sleeps on, with separate biometric profiles.

Derived health inferences. This is where it gets consequential. From the raw sensor data, these platforms produce derived metrics: sleep debt accumulation, cardiovascular readiness scores, overnight HRV trends, and respiratory disturbance indicators. Some have added functionality that flags patterns consistent with sleep apnea symptoms.

Sleep apnea screening from a consumer mattress raises its own privacy questions. A mattress flagging potential apnea symptoms is generating a health inference that could theoretically be relevant to health insurance underwriting, disability determinations, or other contexts. Whether or not that concern is immediately realistic, the data is created.


Who Holds This Data

Eight Sleep

Eight Sleep is the dominant player in the premium smart mattress market. Their platform, Pod, requires a “Pod Membership” subscription for access to the AI-powered features — temperature scheduling, sleep coaching, and health insights. The mattress itself continues to function without the subscription, but the data collection continues.

Eight Sleep’s privacy policy states that sleep data is stored on their servers. The company has partnered with health systems and corporate wellness programs, though they state that individual data is not shared with these partners in identifiable form.

What Eight Sleep specifically does with heart rate and HRV data for product improvement, model training, or research is described in terms that are more permissive than specific. Their policy permits use of data for “improving our Services” and “developing new products and features” — language that is standard across the industry but that provides limited concrete information about how your nightly biometrics are actually used.

Sleep Number

Sleep Number’s Smart Beds include their SleepIQ platform, which tracks sleep quality scores, breathing rate, heart rate, and movement patterns. The data is tied to a Sleep Number account and stored on their servers.

Sleep Number has been more explicit than some competitors about using aggregate sleep data for research partnerships. They have published research papers using anonymized data from their user base, which represents one of the largest longitudinal sleep datasets in existence.

Their privacy policy reserves the right to share de-identified, aggregated data with research partners and to use data for machine learning model improvement. The practical reachability of your specific sleep records through de-identified research data depends on aggregation level and re-identification risk, which varies.

Withings Sleep Analyzer

Withings takes a different approach: their Sleep Analyzer is a thin pad placed under the mattress rather than an integrated system. It measures heart rate, breathing, snoring, and sleep cycles and syncs to the Withings Health Mate app.

Withings is a French company, which means their data handling is subject to GDPR and they have to honor European data rights requests. They offer a “Local Data Storage” option in some contexts, though full offline operation isn’t supported. Their privacy policy is more granular than most competitors about data use categories.


What Can Be Inferred

Health data has an inference problem. You don’t have to explicitly share a health condition for health-related inferences to be made about you from other data.

Eight hours of continuous cardiovascular monitoring creates a detailed health profile:

Cardiovascular health. Resting heart rate, HRV, and overnight recovery scores are established proxies for cardiovascular fitness and stress load. Longitudinal trends in these metrics are clinically meaningful.

Respiratory conditions. Breathing rate variability and respiratory disturbance index scores are used by these platforms to flag potential sleep apnea. Sleep apnea affects roughly 1 billion people globally and is associated with cardiovascular disease, diabetes, and other conditions. A mattress generating an apnea likelihood score is generating a health inference regardless of whether it’s labeled as a medical device.

Mental health correlates. HRV and sleep architecture are correlated with depression, anxiety, and stress. Eight Sleep’s own marketing describes their system as tracking “stress levels” through overnight biometric data.

Pregnancy. Body temperature changes are among the early physiological signals of pregnancy. Elevated basal body temperature detected through continuous overnight monitoring would be an identifiable pattern — and one that individuals may not want their mattress manufacturer to log.

None of these inferences are explicitly offered to you as features. They’re latent in the data. Whether they’re actively computed, stored, or used depends on each company’s systems and evolves with each product update.


The Subscription Dependency Problem

Smart mattresses introduce a privacy risk that conventional mattresses don’t: the subscription trap.

When you purchase an Eight Sleep Pod, the hardware is yours. The features — temperature scheduling, health insights, smart alarm — require an ongoing Pod Membership subscription. If you cancel the subscription, you lose access to the app’s features, but the mattress continues to collect data. Whether data collection continues without subscription depends on connectivity settings that are not always clearly documented.

Sleep Number’s SleepIQ features similarly tie core functionality to a connected account. When Sleep Number announced in early 2025 that some Smart Bed features would require a subscription after a free period ended, existing owners discovered that furniture they owned had become a subscription service.

This creates a specific privacy concern: the company that sold you the hardware continues to have a data relationship with you even after the commercial relationship changes. Your sleep data from years of use remains on their servers. Account deletion procedures vary in their completeness and permanence.


The Business Model Question

Hardware companies increasingly fund themselves through data and services rather than hardware margins alone. Eight Sleep’s valuation has been driven partly by the health data platform potential of a large sleep biometric dataset.

Insurance companies have expressed interest in sleep data. Whether smart mattress data will eventually be incorporated into health insurance underwriting models — as fitness data from wearables has been in some products — is unknown. The data exists. It is specific, longitudinal, and health-relevant. The commercial pressure to monetize it exists.

For now, you’re making a bet that the company’s current privacy policy reflects how they’ll actually use your data over the decade your mattress is likely to remain in service.


Practical Steps

Before buying: Read the full privacy policy, not just the feature sheet. Specifically look for: what data is retained after account deletion, whether data is shared with research or commercial partners, and whether you can opt out of non-essential data collection while retaining core functionality.

After buying: Audit your connected devices list periodically. Revoke third-party app integrations (Health app, Alexa, etc.) you don’t actively use. Understand what your account deletion procedure covers.

Account deletion: When you want to stop using the service, submit a formal data deletion request in writing — not just an account cancellation. These are different things in most privacy policies. GDPR (if you’re in Europe) and CCPA (if you’re in California) give you the right to request deletion of your personal data.

The offline option: If you own a smart mattress and want to limit ongoing data collection without replacing it, check whether your specific model has a Wi-Fi disable option. Some do; most make it inconvenient.

Alternative approach: Traditional mattresses, combined with a standalone sleep tracker that stores data locally or in a privacy-respecting app, can provide much of the sleep data benefit without the continuous bedroom biometric feed to a commercial server.


What This Means for Personal Records

Sleep health data is increasingly meaningful to employers, insurers, and healthcare providers. It’s also deeply intimate — a continuous record of your body’s nightly state.

Storing your own sleep health records privately — separate from a mattress manufacturer’s servers — is worth considering. If you export your sleep data from Eight Sleep or Sleep Number (both offer export tools), you can maintain your own record while retaining the ability to audit and delete what the manufacturer holds.

The principle is the same one that applies to medical records, financial documents, and personal photos: data that matters about your health and your life belongs in storage you control.


Smart mattresses are a compelling technology. The privacy tradeoffs are real and ongoing. Knowing what your bed knows about you is the first step to deciding how much you’re comfortable with that arrangement.

Your memories deserve better than an ad platform.

Try daftei free →
← All posts