Signals Inbox·August 22, 2026·Longevity

Oura's lawsuit: what's happening?

Oura is facing a new proposed class action over how it markets sleep-stage accuracy. The science behind the ring is stronger than the lawsuit sometimes suggests, but the gap between Oura's headline 95% claim, its older 79% figure and much weaker results in some independent studies gives the case something real to fight over.

Q1What Oura lawsuit was just filed?

Oura is currently facing a proposed class action accusing it of overselling the accuracy of its sleep-stage tracking. So far, the case consists of allegations, and Oura says it will fight them.

California resident Madison Surber filed the case in federal court in San Francisco against Oura Inc. and Oura Health Oy. She says she bought an Oura Ring 4 Gold for $513.68 in May 2025 after seeing Oura marketing at Best Buy, on Oura's website and on Best Buy's sales page.

Surber says that marketing led her to expect accurate stage-by-stage sleep tracking. Her argument is simple: had Oura explained the limitations more clearly, she either would have skipped the purchase or paid less.

She wants to represent a nationwide group of buyers, with a separate California subclass, and brings claims including fraud, false advertising, breach of warranty and violations of California consumer-protection laws.

The lawsuit is extremely new. No class has been certified, no judge has decided whether Oura's claims are misleading and no damages have been awarded. Oura's public response so far has been equally short: it disputes the allegations and intends to defend itself.

Q2What exactly is Oura accused of misleading customers about?

The lawsuit says Oura makes algorithmic estimates of sleep stages sound much closer to clinical measurement than they really are.

Oura's marketing has used phrases such as "Built for accuracy" and "Unparalleled Accuracy." More importantly, the current Ring 5 sales page advertises "95% Sleep Staging Accuracy" compared with a clinical sleep lab.

That sounds straightforward to a shopper: wear the ring, and it can tell with very high accuracy whether you are awake or in light, deep or REM sleep.

The plaintiff attacks the gap between that impression and how sleep stages are actually determined. Clinical polysomnography, or PSG, uses signals including electrical brain activity, eye movements and muscle activity. Oura instead reads movement, skin temperature, heart rate, heart-rate variability and breathing, then runs those signals through an algorithm.

Oura itself explains this process on its support pages. It also acknowledges in its educational material that PSG uses measurements its ring does not collect.

So the real argument is about the leap from "our algorithm can infer sleep stages" to a consumer-facing percentage that looks very close to clinical certainty. That is a much more serious question than whether a ring physically contains the same equipment as a sleep lab.

Q3Does Oura really claim 95% sleep staging accuracy today?

Yes. Oura is currently advertising Ring 5 as having 95% sleep-staging accuracy compared with a clinical sleep lab.

The claim is still visible on Oura's Ring 5 sales page under "Rebuilt for greater accuracy." It sits beside two other large statistics: 99% heart-rate accuracy compared with ECG and a statement that 90% of members feel healthier.

The placement makes the sleep figure unusually strong. Someone shopping for the ring sees three simple percentages presented as headline evidence of product performance.

There is also an interesting difference in how Oura supports them. The 99% heart-rate figure carries a visible footnote tied to the ECG comparison. The sleep-staging number does not have an equally obvious nearby explanation of the study population, algorithm version or calculation behind it.

That does not make the figure false. Ring 5 uses redesigned hardware, and Oura may have newer validation data behind the claim. But once a company puts "95%" in large type next to "clinical sleep lab," the methodology becomes important. Right now, the public product page gives shoppers the conclusion much more clearly than it gives them the denominator.

Q4How can Oura say both 79% and 95% sleep-stage accuracy?

Right now, Oura itself publishes two very different sleep-stage numbers, and its public material does not clearly explain how to reconcile them.

Oura's recently updated Member Care page still says its sleep-staging algorithm achieved 79% agreement with PSG sleep-lab tests. That figure comes from its established Sleep Staging Algorithm 2.0 and refers to four-stage classification across wake, light, deep and REM sleep.

As seen above, Oura now puts 95% on the Ring 5 sales page.

A jump of 16 percentage points could be perfectly legitimate if Ring 5 has better sensors, a new model or a different validation protocol. The problem is that we cannot simply assume those two percentages measure the same thing.

Accuracy can mean several things in sleep research. Researchers can measure sleep versus wake, classify all four stages, report sensitivity for a particular stage, average several metrics or aggregate results at the nightly level. Those calculations can produce very different percentages from the same underlying data.

For Oura, this is probably the single most important number to explain. The older 79% figure is understandable because the task behind it is clearly described. The newer headline figure is much stronger while its public methodology is much less obvious.

Q5Can an Oura Ring track sleep stages without measuring brain waves?

Yes. An Oura Ring can infer sleep stages from signals measured on the finger, and published research shows that those predictions contain real information.

Sleep changes more than brain activity. Heart rate, heart-rate variability, breathing and movement also shift during the night. Deep sleep tends to come with little movement and relatively stable cardiovascular activity. REM usually produces a different autonomic pattern. Machine-learning models can learn those combinations and predict which PSG-defined stage is most likely.

The limitation is precision.

A sleep lab observes several of the signals doctors actually use to define sleep stages. Oura reconstructs those stages indirectly from physiological clues available at the finger. In practice, this works well enough to produce useful predictions, but errors are inevitable when different stages create overlapping peripheral patterns.

This is why the plaintiff's broad claim that sleep "happens in the brain, not on one's finger" only gets us halfway. Sleep staging originates from neurological criteria, but a model can still predict those states from correlated signals elsewhere in the body. The useful question is how often that prediction is right.

Q6How accurate is Oura sleep tracking in healthy adults?

Oura performs much better than guesswork in healthy adults, with several PSG studies putting four-stage performance around the high 70s and simpler sleep-versus-wake detection above 90%.

A National University of Singapore validation used 58 participants across 157 nights. Oura reached roughly 76% to 78% accuracy when distinguishing wake, light, deep and REM sleep, while the easier sleep-versus-wake task reached about 92% to 93%.

A larger University of Tokyo study followed 96 generally healthy adults across up to three nights each, producing 421,045 thirty-second epochs. Oura Gen3 with Sleep Staging Algorithm 2.0 again performed strongly on sleep-versus-wake detection and showed stage-specific accuracy ranging from roughly 76% for light sleep to above 90% for REM.

Brigham and Women's Hospital then compared Oura Gen3, Fitbit Sense 2 and Apple Watch Series 8 against PSG in 35 healthy adults. Oura's sensitivity across the three sleep stages came in between 76% and 79.5%, with similar precision. At the nightly level, researchers found no statistically significant difference between Oura and PSG for the estimated amount of wake, light, deep or REM sleep.

Those studies give us a fairly consistent picture: Oura can extract a lot of sleep information from the finger, especially in healthy adults. They also make a universal 95% interpretation difficult to accept without knowing exactly how that newer number was calculated.

What independent Oura sleep studies show

Q7Why did another Oura study find only 53.18% sleep-stage accuracy?

The 53.18% result is genuine, and it shows how sharply Oura's performance can fall in a tougher clinical population.

Researchers at Charité University Medicine Berlin tested Oura, SleepOn and Circul rings against PSG in 45 sleep-lab patients. These people were being evaluated in a clinical setting and included patients with sleep disorders and other medical conditions, a much harder population than the healthy volunteers used in several earlier Oura validations.

Oura reached 85.03% when researchers only asked whether someone was asleep or awake. With four possible labels — wake, light, deep and REM — accuracy fell to 53.18%.

SleepOn reached 50.48%, while Circul came in at 35.06%, so Oura still performed best of the three rings. Yet the individual errors were large enough for the researchers to warn against using these devices for clinical diagnosis or interpreting single nights as though they were PSG.

A healthy person watching long-term sleep patterns and a sleep-clinic patient whose exact sleep architecture may affect diagnosis are asking very different things of the device.

The study also used an older Oura generation and algorithm than the current Ring 5, which limits how directly we can apply its percentage to today's product. Sleep researchers have specifically warned about this version-lag problem because commercial wearables often change faster than validation studies can be published.

Q8Does Oura become less accurate for older adults?

Probably, and one recent PSG study found errors large enough that older users should be cautious with exact sleep-stage numbers.

Researchers compared several consumer sleep trackers in 13 younger adults aged 19 to 24 and 19 older adults aged 56 to 80.

For the older group, Oura underestimated total sleep time by an average of 75.5 minutes. More strikingly, it overestimated deep sleep by 71.5 minutes. The study concluded that consumer devices generally showed wider errors in older adults and struggled particularly with individual sleep stages.

This was a small study based on one PSG night, so we would not use it to declare that every older Oura user gets an extra hour of fake deep sleep. But the direction fits what we saw in the clinical-patient study: performance becomes less reliable once the user's sleep differs from the healthy populations on which wearables often validate best.

For Oura's marketing, that creates a legitimate question. A single universal accuracy percentage is easier to advertise than a result that changes according to age, sleep quality, health status and the metric being measured.

Q9What Oura sleep data can you trust the most?

Oura currently looks much stronger for total sleep and long-term patterns than for treating every light, deep or REM segment as exact.

The difference comes straight out of the validation data.

In the Charité clinical study, Oura could be fairly close on total sleep across a whole night while still misclassifying many individual 30-second periods. A chunk of PSG-defined deep sleep might become light sleep in the app, while another period gets misclassified in the opposite direction. The nightly total can end up looking reasonable even when the detailed hypnogram contains mistakes.

The healthy-adult studies tell a similar story from the other side. Oura tends to do very well when separating broad sleep from wake and gives useful estimates of overall sleep duration. Four-way classification is harder.

That changes how we would use the product. "I have been sleeping about 40 minutes longer over the past six weeks" is the kind of pattern a wearable can be useful for. "I received exactly 58 minutes of deep sleep last night" deserves more skepticism.

Even Oura's current app smooths its consumer hypnogram into five-minute intervals for readability, while sleep research typically scores PSG in 30-second epochs. The product is built to help people see patterns, and that is where the evidence is strongest.

Q10Can a Ring 4 buyer really sue Oura over Ring 5's accuracy claim?

This is one of Oura's best legal angles because the named plaintiff bought Ring 4 before the newer Ring 5 claim existed.

Surber says she bought a Ring 4 Gold in May 2025 after seeing Oura's accuracy marketing. Ring 5 arrived later with redesigned sensing hardware and the stronger headline claim now at the center of the controversy.

She therefore could not have relied on that later statement when deciding whether to make her purchase.

Her own case can still focus on the Ring 4 marketing she actually saw. The complaint describes a wider advertising campaign around accurate stage-by-stage sleep tracking, and California law can sometimes allow plaintiffs to rely on a broader campaign rather than identifying one single sentence that caused the purchase.

The harder question is whether Surber can also represent people who bought Ring 5.

Her lawyers argue that Ring 4, Ring 4 Ceramic and Ring 5 are substantially similar products making the same basic sleep-accuracy promise. Oura can push in the opposite direction: Ring 5 has changed hardware, changed marketing and apparently changed performance claims.

That creates a slightly awkward tradeoff for Oura. The more the company argues that Ring 5 is technically better and materially different, the easier it becomes to question whether one Ring 4 buyer should speak for Ring 5 customers. On this part of the case, the product upgrade may actually help Oura.

Q11How big could the Oura class action get?

The potential class is huge on paper, but current sales numbers tell us almost nothing about what Oura could actually owe.

Oura said it had sold more than 5.5 million rings by late 2025, with more than half of those sales occurring during the preceding year. Revenue exceeded $500 million in 2024 and then reached roughly $1 billion in 2025.

The complaint tries to cover a nationwide group of relevant buyers, plus a California subclass. It also alleges more than $5 million in controversy, which allows the case to proceed under the federal Class Action Fairness Act.

None of that means millions of customers are about to receive refunds.

Surber first has to keep the claims alive. Then she has to persuade a judge that the proposed class is coherent enough to certify. Oura can argue that buyers saw different advertisements, purchased different ring generations and may have valued completely different features such as heart-rate tracking, readiness scores, fertility features or activity monitoring.

Damages would create another fight. If the ring still provides substantial value, the more realistic theory may be that customers paid an extra premium because of overstated sleep accuracy, rather than that every ring was worth zero.

What the proposed class action seeks

Q12Could the Oura lawsuit hurt its IPO?

For now, the lawsuit looks more like an IPO disclosure and reputation problem than something capable of derailing Oura by itself.

Oura is heading toward the public markets from a position of considerable strength. It raised more than $900 million at roughly an $11 billion valuation, doubled annual revenue from about $500 million to around $1 billion and sold more than 5.5 million rings. The company has also confidentially submitted paperwork for a U.S. IPO.

That scale makes one newly filed consumer case financially manageable unless the facts become much worse.

The awkward part is what the lawsuit attacks. Sleep helped build Oura's brand. The company differentiated itself from wrist-based fitness trackers partly by making recovery and sleep central to the product, and those measurements feed into the readiness experience users see every day.

An IPO investor will therefore care less about Surber's $513.68 purchase than about what discovery might uncover. If Oura can show strong internal validation supporting its current marketing, this could remain a manageable legal footnote. If internal data reveal a big gap between what the scientists measured and what the marketing team told consumers, the story gets much more uncomfortable.

Currently, there is no public evidence of that second scenario.

Q13What happens next in the Oura lawsuit?

The next useful evidence should come from Oura's formal response and, if the case survives, from discovery into how the company substantiated its accuracy claims.

Oura can challenge the complaint before getting anywhere near a trial. Expect arguments over what Surber actually saw, whether her Ring 4 purchase gives her standing to pursue claims involving Ring 5, whether the different ring generations are sufficiently similar and whether the alleged statements would really mislead a reasonable consumer.

If the case gets past those fights, the internal accuracy work becomes much more interesting.

We would want to see the full validation behind the current Ring 5 claim: which people were tested, how many nights were recorded, which hardware and software versions were used, whether the test involved healthy sleepers or clinical patients, and exactly how "accuracy" was calculated.

The other key documents would be internal conversations between Oura's researchers, product teams and marketers. They could show whether everyone inside the company used the accuracy percentage in the same way before it reached the sales page.

As of now, none of that evidence is public. The case remains at the beginning, and Oura continues to advertise and sell its rings normally.

Q14So, what's actually happening with Oura's lawsuit?

The Oura lawsuit has found a real weakness in the company's marketing, but its broader scientific attack goes too far.

We have enough independent and peer-reviewed evidence to say that Oura can meaningfully estimate sleep stages from finger-based signals. Across healthy-adult studies, results repeatedly land well above chance, with particularly strong performance on total sleep and sleep-versus-wake detection. Calling the technology pure guesswork does not fit the evidence.

Where Oura is much more exposed is the way accuracy gets compressed into one easy number.

Its current Ring 5 page gives shoppers an extremely strong sleep-staging percentage. Its own support material still describes materially lower agreement for the established four-stage algorithm. Clinical patients and older adults have also produced much weaker results than healthy validation groups. Those findings can all be scientifically compatible because they involve different devices, populations and metrics. A normal shopper would have no realistic way to work that out from the headline marketing.

The plaintiff also has a meaningful weakness: she bought Ring 4 before Ring 5 and its newer claim existed. That gives Oura room to narrow the case even if the broader advertising allegations survive.

Our judgment today is that this is a credible false-advertising case, especially if Oura cannot clearly show how its newest headline accuracy figure was derived. The evidence currently gives much less support to the lawsuit's suggestion that Oura fundamentally cannot track sleep stages because it sits on a finger.

What happens next depends heavily on one question that Oura should be able to answer with data: what exactly did the company measure before it decided that its newest ring was accurate enough to market this way?