Consumer healthtech · unit economics

Ōura: the customer pays for their own acquisition

A hardware-plus-subscription model that works precisely where a service-only model doesn't — and the two structural attacks now aimed at it, four months after a confidential IPO filing. Public sources only. August 2026.

The thesis in one sentence. Ōura sells a ring at roughly $273 realised against about $131 of manufacturing cost, so every customer arrives having already paid a gross profit that covers the cost of acquiring them — and the $71.88-a-year membership that follows is a high-margin annuity on top. That inversion is why Ōura is profitable at scale while service-only wellness businesses are not. The risk is inverted too: both revenue lines are now under attack at once, hardware from subscription-free rings at $99–349, and the subscription from competitors who have made "no subscription" their primary marketing claim.
Reported Estimated Assumption
Ōura has never published audited financials, a gross margin, an attach rate or a CAC. Every input is tagged: Reported = published by the company, a regulator, a court or a filing. Estimated = triangulated from adjacent public data. Assumption = mine, no public anchor. A public S-1 will convert most of this into hard numbers.
Section 1

What happened at Ōura

Twelve months in which the company quadrupled revenue, won a patent case, lost the exclusivity it bought with it, and filed to go public.

Aug 6, 2026
Buys its San Francisco headquarters
Closing on the roughly 50,000 sq ft building at 500 Pine Street, occupied since January, up from under 10,000 sq ft. Headcount is now above 1,200 across six offices. The building last traded at $73M in 2018; Ōura's price is undisclosed.
Jun 20, 2026
Ultrahuman returns to the US market
The Ring Pro ships in the United States at an early-bird $349, with a heart-rate sensing architecture explicitly engineered around Ōura's patents, a 15-day battery, and no subscription for core insights. Ōura's exclusion order bought roughly eight months of exclusivity against its closest competitor, not a permanent exclusion.
Jun 4–16, 2026
Health Radar and the clinical layer
Ring 5 ships alongside Health Radar — blood-pressure signal detection from nighttime dipping patterns, nighttime breathing trends, symptom detection — built with 40+ in-house MDs and PhDs. Counsel Health integration adds in-app messaging with licensed physicians in 43 states, and Lab Uploads lets members import outside lab results.
May 28, 2026
Ring 5 launches — and raises prices
40% smaller, titanium, at $399 for base finishes and $499 for premium. That is a $50 increase on the base and $100 on several finishes. The cadence is notable: 18 months after Ring 4, an acceleration trade press tied directly to pressure from subscription-free rivals.
May 21, 2026
Confidential S-1 submitted
Ōura confirms it has filed a draft registration statement with the SEC. Share count, price range, underwriters and venue are all undisclosed. European coverage framed the listing above €9 billion.
May 20, 2026
Five million paid members
On pace to pass 5M paid members in Q2, membership up more than 4× in two years, and the only retention figure the company has ever released: above 80% renewal after the first year. Also disclosed: 11% of members are healthcare providers and 51% manage at least one chronic condition.
Jan 6, 2026
FDA opens the door Ring 5 walks through
Updated general-wellness guidance clarifies that non-invasive estimation of blood pressure, oxygen saturation and glucose does not automatically make a product a regulated device. Ōura ships nighttime blood pressure five months later with no clearance, no clearance cost and no clearance timeline — while running a separate IRB-approved study for the version it will submit.
Oct 14–21, 2025
$900M Series E at $11B, and the ITC win lands
Fidelity leads, with Iconiq joining. Days later the exclusion order against Ultrahuman and RingConn becomes enforceable — and RingConn settles into a multi-year royalty-bearing licence rather than exiting. Ōura converted its patent into a revenue stream and a legitimised competitor rather than an exclusion.

Not found: audited financials of any kind, gross margin, EBITDA, attach rate, CAC, marketing spend, multi-year retention, current contract manufacturers, enterprise revenue share, the RingConn royalty rate, and the outcome of Ultrahuman's Federal Circuit appeal or the USPTO challenge to the underlying patent.

Section 2

What happened around it

The category reorganised itself around a single competitive claim: no subscription.

Aug 3, 2026
The subscription critique goes mainstream
The Verge publishes "The Screenless Wearable Boom Has a Subscription Problem," arguing that Whoop and Ōura invented the screenless category to escape data overload and then recreated it, and that subscription-first models erode as hardware commoditises. Its framing — "everything useful is paywalled" — is the argument every competitor is now running.
Jul 21, 2026
Garmin enters at $199 with no subscription
CIRQA, a screenless band with sleep staging, HRV, body battery and 80+ activities, explicitly subscription-free. Reported to have sold out immediately. Garmin has distribution and a brand Ōura cannot dismiss.
May 26, 2026
Fitbit Air lands at $99
Google's screenless tracker with skin temperature, 24/7 heart rate and irregular-rhythm notifications, positioned explicitly against Whoop. The AI coaching layer sits behind a $10/month Google Health Premium tier — the same architecture as Ōura, at a quarter of the hardware price.
Jun 23, 2026
FDA closes the Whoop warning letter
Whoop had been cited in July 2025 for its Blood Pressure Insights feature as an unapproved medical device. The letter closed after modification — but it had already produced a putative class action in the Northern District of California, which is the risk the FDA's blessing does not cover.
Feb 5, 2026
Apple retreats — reportedly because of Ōura and Whoop
Apple scales back its AI health coach, and the reported reason is that the planned service was judged not competitive against Ōura and Whoop, which "offer more compelling and useful features." Apple's smart ring has separately been reported as abandoned entirely. The most dangerous potential entrant has stepped back — though nothing makes that permanent.
Mar 2026
Whoop raises $575M at $10.1B
The closest business-model comparison, at roughly 2.5M members and $1.1B run-rate. On those figures Whoop trades near 9× revenue while Ōura's last private mark is closer to 5.5× forward revenue — the larger business carries the cheaper multiple.
Section 3

Market forces

The demand signal is real but the category is small, and the two research houses that track it disagree about whether it is still accelerating.

Tailwinds

  • The category is growing off a small base. IDC puts 2026 smart ring shipments at 4.9M units, up 12.8%. CCS Insight puts it at 6M and up 53%, with a 32% CAGR to 2030.
  • Ōura is the category. It held roughly 80% share as recently as 2023 and remains the revenue majority, which is why category ASPs rose 9% in 2026 — Ōura raised prices.
  • The FDA loosened the wellness boundary in January 2026, letting Ōura ship blood-pressure estimation with no clearance cost and no submission timeline.
  • HSA and FSA eligibility covers the ring, the membership, Health Panels and the Stelo integration — converting discretionary spend into pre-tax spend, worth an effective 25–40% depending on bracket.
  • Enterprise and insurer channels are the real moat. 1,200+ organisational partnerships, 4,600+ retail locations, and a Medicare Advantage deal with Essence Healthcare that now feeds nighttime breathing data into physician workflow for sleep apnoea screening. A $199 ring cannot replicate procurement relationships and clinical integration.
  • Apple has stepped back from both the ring and the subscription health-coaching layer.
  • Consumers do pay for wearable services — 32% of wearable owners have an attached subscription.

Headwinds

  • Ōura is the most expensive hardware in the category and one of only two products requiring a subscription. Every other vendor undercuts on at least one axis; most undercut on both.
  • The patent moat monetises but does not exclude. Ultrahuman designed around it and returned in eight months; RingConn is licensed and selling at $199–314.
  • The underlying patent is under USPTO challenge by Samsung and others. Invalidation would remove the royalty stream and the litigation threat together.
  • Smart rings are under 1% of wearables by units — 4.9M against roughly 626M total. This is a niche with a premium price, not a mass category.
  • Hardware cadence is accelerating into replacement fatigue. Ring 5 came 18 months after Ring 4, which raises R&D and inventory intensity to defend a subscription base.
  • Rings have an attrition mode no other wearable has: fingers change size with weight, pregnancy, temperature and age, and a ring that no longer fits is a full repurchase.
  • The two houses disagree on growth — 12.8% versus 53% for the same year. That gap is the central unresolved question for anyone underwriting this.
  • Privacy is a live reputational exposure. See Section 4.
Section 4

The evidence and privacy critique

Ōura's measurement accuracy holds up better than sceptics assume. The harder problems are what people do with the data, whether anyone's health actually improves, and who else can see it.

On accuracy, the evidence is genuinely good

A 2025 systematic review and meta-analysis in OTO Open pooled six validation studies against polysomnography, n=388, and found no statistically significant difference on any of seven sleep parameters — total sleep time, efficiency, wake after onset, onset latency, light, deep and REM. That is a stronger result than the usual dismissal of consumer sleep tracking allows for, and it should be conceded plainly.

The caveats matter, though. The sample is small, it pooled earlier ring generations rather than Ring 4 or 5, and null results with wide confidence intervals are not demonstrated equivalence — the deep-sleep interval spans −10 to +13 minutes, a large fraction of a typical night. Agreement on nightly totals is also a much weaker test than epoch-by-epoch staging agreement, which the meta-analysis does not report.

The sensor has a documented bias problem

A 2024 JMIR meta-analysis of 23 pulse-oximetry studies (n=59,684) found SpO2 accuracy breached the FDA's 3% threshold across every skin-tone group — light 3.96%, medium 4.71%, dark 4.15% — and wearable pulse-rate error ranging from ±13.5 to ±33.7 bpm against a ±5 bpm industry standard. The finding is broader and worse than the familiar "PPG fails on dark skin" framing. Ring 5's marketing claim of "more powerful LEDs to improve accuracy across varied finger sizes and skin tones" reads as an implicit acknowledgement.

The number that should worry a subscription underwriter

A December 2025 systematic review in Biomimetics covering 107 studies and roughly 100,000 participants found adherence falling from 80% at three months to 43% at twelve months. Ōura reports above 80% membership renewal after year one. Both can be true at once only if paying consumers behave very differently from study participants — or if renewal is measuring billing continuation rather than active use. A subscription that keeps billing while the ring sits in a drawer is churn that has not yet expressed itself, and it is the single largest risk to the lifetime value in the model below.

The same review found 65% of studies carried moderate-to-high risk of bias, 89% relied on undisclosed proprietary algorithms, only 35% documented participant race or ethnicity, and Ōura alone accounted for 72% of all studies — a degree of vendor concentration that makes the evidence base hard to read as independent.

Orthosomnia is now measured, not anecdotal

A 2026 national survey published in SLEEP (n=1,280) found 32.4% of US adults use digital sleep tracking, and 30.9% of those screened positive for orthosomnia risk — anxiety about their own tracker data — roughly twice prior estimates. A validated instrument for it now exists. For a company whose core product is a nightly score, one in three users showing anxiety about that score cuts both ways: it drives engagement in the short run and abandonment in the long run.

Whether any of it improves health is unresolved

The best available evidence — a Lancet Digital Health review of 39 systematic reviews covering 163,992 participants — found activity trackers add roughly 1,800 steps and 40 minutes of walking a day, and about a kilogram of weight loss. Effects on blood pressure, cholesterol and HbA1c were "typically small and often non-significant." And 34 of the 39 reviews were rated critically low confidence. That literature is about wrist-worn step trackers; there is essentially no equivalent outcomes base for sleep-and-recovery-score rings, whose mechanism is quite different.

The privacy exposure is specific and self-inflicted

The Department of Defense has been Ōura's largest enterprise customer since 2019, with tens of thousands of rings deployed, a Fort Worth facility built to serve it, and an enterprise platform hosted on Palantir's federal compliance infrastructure. Ōura describes the Palantir arrangement as a small inherited security layer and says it does not share or sell user data. The backlash in late 2025 was nonetheless real and named — cancellations covered in national press, a petition, and academic write-ups.

The acute version is reproductive data. Ōura markets aggressively to women — cycle, fertility, menopause and pregnancy tracking, with 26M cycles and 350,000 pregnancies logged — and women in their early twenties are its fastest-growing segment. Wearables are not covered by HIPAA. Building the most trusted women's-health data product and being the Pentagon's largest wearable vendor are strategies in tension with the same customer, and 2025 showed the tension is not theoretical. It belongs in an S-1 risk-factors section, which means it will shortly be public.

Section 5

The competitive set

Eleven priced competitors, and Ōura and Whoop are the only two that require a subscription for core functionality. That is a structural alignment, not a coincidence.

ProductHardwareSubscription3-year costScalePosition
Ōura Ring 5$399–499$5.99/mo required$615–7155M+ members
~8M rings cumulative
Highest hardware price in the category and one of two products gating features behind a fee
Whoop One / Peak / Life
the model comparison
included$199 / $239 / $359 per yr$597–1,077~2.5M members
~$1.1B run-rate
Hardware bundled into subscription. Converges on the same three-year value as Ōura from the opposite direction
Ultrahuman Ring Pro$349none (optional add-ons)$349~$150M run-rate
#2 by share
Designed around Ōura's patent and back in the US since June. A 43% discount on three-year cost of ownership
RingConn Gen 2 / Air$314 / $199none$199–314not disclosedLicensed by Ōura and paying royalties — a permanently sanctioned price anchor at half the cost
Samsung Galaxy Ring$399 listnone$399not disclosedPersistent deep discounting suggests weak sell-through; successor not expected before 2027
Circular Ring 2$304freemium$304not foundECG and AFib, HSA/FSA eligible, "140+ biometrics free forever"
Garmin CIRQA$199.99none$200sold out at launchIncumbent distribution, explicitly subscription-free, July 2026
Amazfit Helio$149.99none$150not disclosedHelio Ring 2 confirmed for H2 2026 with ECG and body composition in patent filings
Fitbit Air (Google)$99$10/mo optional$99–459not disclosedSame architecture as Ōura at a quarter of the hardware price, with the AI layer paywalled
Movano EvieMEDnot foundnonecompany in distressThe only FDA-cleared smart-ring pulse oximeter, held by a company facing Nasdaq deficiency notices

Three-year cost assumes one device and continuous membership at list, before HSA/FSA treatment or discounting. Ōura's realised hardware price is materially below list — see the assumption register.

The moat monetised instead of excluding

Ōura won at the ITC in August 2025 and the orders became enforceable in October. Eight months later Ultrahuman was back, selling a designed-around product at $349 with no subscription, and RingConn had converted from defendant to licensee — legitimised, paying royalties, and selling at $199. Ōura now collects an undisclosed royalty from a sanctioned competitor that undercuts it by half. The patent bought time and a revenue line; it did not buy the market. And the patent itself is under USPTO challenge.

Everyone else sells the absence of a subscription

Ultrahuman, RingConn, Circular, Amazfit, Renpho, Leep, Garmin and Samsung all market subscription-free as their headline differentiator, and they name Ōura when they do it. Survey data says 92% of prospective buyers would pay extra for health features — but that is stated preference, and the revealed preference visible across 2026 is that the fastest-moving products in the category are the ones that removed the fee.

Section 6

Ōura versus Whoop, over three years

These are the same business with the revenue recognised in opposite orders. Ōura charges for hardware and rents the software; Whoop gives away hardware and charges more for the software. Holding the customer's holding period constant shows how close the two land — and how differently they get there. These update live with the model in Section 8.

Whoop hardware cost is a modelled assumption — the company does not disclose it, and the device is given away, so it sits in cost of revenue rather than as a price. Ōura figures use the derived realised selling price, not list.

What the comparison shows. Whoop collects more revenue per member but hands over a device for free, so its gross profit is deferred and its cash is negative at the moment of acquisition. Ōura collects less recurring revenue but banks a hardware margin on day one. The practical difference is who funds the customer acquisition. Ōura's hardware gross profit arrives before any marketing cost is amortised, which is why it can be profitable at $1–2B of revenue while carrying a $71.88 subscription; Whoop has to keep a member for roughly a year before the relationship pays for its own hardware. Ōura's exposure is that the hardware margin depends on holding a price premium the category is attacking; Whoop's is that a cancelled member takes a device with them.
Section 7

Assumptions

This comes before the model deliberately. Ōura has published a revenue trajectory, a member count and one retention figure. Everything else below is derived or mine.

InputBasisWhere it comes from
Ring price $399–499 (Ring 5), $349 (Ring 4)ReportedCompany store, verified August 2026. Ring 5 raised base finishes $50 and several premium finishes $100
Membership $5.99/mo or $69.99/yrReportedCompany membership page. Unchanged through 2026. Required on Gen3 and newer for anything beyond three daily scores
Realised hardware ASP $273EstimatedDerived by an analyst from Finnish statutory filings. Against a $349–499 list this implies 22–45% of realised give-up to discounting and retail channel margin. The pivotal and least-corroborated number in the model
Hardware cost $131 per ringEstimatedTwo unrelated methods converge: a TechInsights teardown put Ring 4 bill of materials below 39% of the $349 price (≈$136) at 50,000 units, and the Finnish filings imply $130.55. At ~3M units a year realised cost should be lower
Hardware gross margin ~52%EstimatedFollows from the two lines above. No company-disclosed gross margin of any kind exists
Subscription gross margin 88%AssumptionNot disclosed. Described only as "high" by third parties. Covers app infrastructure, clinical content and support
First-year renewal 80%ReportedThe only retention figure Ōura has ever published, stated twice in May 2026. Note it measures renewal, not use
Steady-state renewal 65%AssumptionMine. First-year renewal typically overstates steady state for hardware-tethered subscriptions because device attrition compounds. Set it to 80% to model the company's own figure persisting
Subscription attach 60%AssumptionNever disclosed; the CEO declined to give it in February 2026. Cumulative rings sold is not an installed base, so the 5M members against ~8M rings is not an attach rate. Calibrate using the implied-members tile in the model
CAC $90AssumptionNever disclosed. No marketing spend figure exists either. The only adjacent datapoint is that roughly 22% of employees sit in sales and marketing
Units 3M per yearReportedCompany cites ~3M rings in 2025; an independent read of Finnish filings gives 2.5M. Treat 2.5–3.0M as the range
Revenue trajectoryReported~$250M (2023, implied), >$500M (2024), ~$1B (2025), "close to $2B" 2026 guidance. All CEO statements; none audited
Hardware / subscription split 80 / 20ReportedAttributed to the CEO; implies roughly $400M of subscription revenue on ~$2B in 2026
Fixed operating cost $700M/yrAssumptionMine, scaled to 1,200+ employees and a company guiding to ~$2B. Used only for the company-level bridge
Competitor pricesReportedEach vendor's own store or press release, verified August 2026

Not found, and therefore not estimated: audited financials, gross margin, EBITDA, net income, attach rate, CAC, marketing spend, multi-year retention or cohort curves, enterprise and defence revenue, the RingConn royalty rate, Ring 5 bill of materials, current contract manufacturers, and tariff exposure. Note also that the $96M Department of Defense contract widely cited in 2024 was cancelled in March 2025 and should not be modelled as revenue.

Section 8

The model

Two revenue lines with very different margins, and a customer who pays the acquisition cost on the way in. Toggle any line off to remove it, or drag to change it. The Whoop comparison in Section 6 updates with every change.

Hardware, per ring sold
$
$
$
Subscription
%
$
%
%
Acquisition & scale
$
k
$M
Why the attach rate is a guess, and how to calibrate it

Ōura has never published one, and the CEO declined to give it in February 2026. Dividing 5M members by ~8M cumulative rings gives about 63%, but cumulative shipments are not an installed base — that figure ignores returns, dead and retired devices, replacements, gifts and enterprise units. Use the implied paid members tile instead: adjust attach and renewal until it lands near the 5M the company reported for Q2 2026, and you have a pair of assumptions that is at least internally consistent with the one number they disclosed.

Ōura's three public numbers do not reconcile, and that is the most useful thing in this model. The company has told us roughly 3M rings a year, above 80% first-year renewal, and 5M+ paid members. Hold those together: members equal annual units × attach × membership life, so attach × life must be about 1.67. If renewal really is 80%, membership life is five years and the attach rate can only be about 33% — far lower than anyone assumes for a product where the subscription is mandatory for most features. If instead attach is the 60% modelled here, then steady-state renewal is about 64%, not 80%. Both cannot be true. The "80% renewal branch" preset shows the first reading; the base case shows the second. Which one the S-1 confirms matters more to the valuation than anything else on this page.
Two honest limits. This is a steady-state model, so for a company growing this fast the implied-members tile runs ahead of reality — a base still ramping has not yet accumulated its full membership life. And implied revenue sits below the ~$2B 2026 guidance because unit volume is guided to grow well past the 3M reported for 2025; at a $273 realised price, roughly 5.5–6M units are needed to reach it.

From one ring to the P&L

Read down the ladder. Each line is live.

The line that makes this business different. Look at contribution at the point of sale. In a service-only subscription business that line is deeply negative — you pay to acquire a customer and recover it over years. Here the hardware gross profit lands the moment the ring ships, so the customer effectively funds their own acquisition and everything the subscription earns afterwards is incremental. That is the whole structural argument for selling hardware at a margin rather than giving it away.

What it costs. The same structure makes revenue dependent on selling new units every year rather than on a base that compounds. Subscription revenue is roughly 20% of the total, so growth still has to come from moving boxes — into a category where every competitor is cheaper and most charge nothing recurring. Drag the realised price down toward $200 and watch how quickly the advantage disappears.

Two modelling caveats. CAC is charged entirely at the point of sale, which is the conservative treatment. And the model assumes one ring per customer per lifetime — no repurchase on the roughly 18-month hardware cadence — so lifetime value is understated for loyal upgraders and overstated for anyone whose ring stops fitting.

Lifetime gross profit per customer
Hardware margin arrives at once; subscription margin accrues over the membership life.
RevenueCostGross profit
Every bar carries its value directly — nothing is encoded by colour alone.
Three-year cost of ownership across the category
Hardware plus any required subscription over three years. Ōura's bar moves with the model.
Lifetime value against renewal rate
Hardware profit is fixed; only the subscription tail moves. Your CAC is the break-even line.
Lifetime gross profitYour CAC
Show the full P&L as a table
Section 9

Strategic cases

Judge each against the structural fact established above: hardware margin funds acquisition, and subscription is only a fifth of revenue. Click "Model this" to load any case.

1 · Hold the premium and out-feature the field

Working so far, narrowing

The current strategy, and it has been vindicated: Ring 5 raised prices into a market full of cheaper rings and revenue is still guided to roughly double. The defence is feature cadence — 14 new features in a year, Health Radar, physician messaging, lab uploads, an in-house clinical team of 40+. What makes it fragile is that the premium is being defended against products that are 50–75% cheaper and free of a recurring fee, and the realised selling price is already 22–45% below list. Drag the realised price toward $220 and the hardware margin that funds acquisition compresses fast.

2 · Drop or restructure the subscription

The CEO has ruled it out

Every competitor markets subscription-free against Ōura by name, and the CEO has publicly refused twice, arguing the fee funds ongoing development. The model shows why he is right on arithmetic: the subscription is high-margin and removing it costs roughly a fifth of revenue at close to 90% margin, which is most of the profit. It also shows the trap — the fee is the single most attacked feature of the product, and defending it forces the hardware cadence that raises R&D and inventory intensity. The preset models a partial capitulation: attach falls and realised price drops to meet Ultrahuman.

3 · Lean into enterprise, insurers and clinical channels

The real moat

This is the part a $199 Shenzhen ring genuinely cannot copy. Ōura already has 1,200+ organisational partnerships, the Department of Defense as its largest enterprise customer, and an Essence Healthcare Medicare Advantage deal that as of June 2026 feeds nighttime breathing data into physician workflow for sleep apnoea screening. Procurement relationships, clinical integration, SOC 2 and HITRUST certification and a 40-person clinical team are the assets that compound. It is not a preset because the unit is a negotiated contract rather than a retail sale — and note that enterprise revenue share has never been disclosed, so its current contribution is unknown.

4 · Push diagnostics and adjacencies

Convergent, not differentiating

Health Panels at $99 for 50 biomarkers, the Dexcom Stelo integration at $99, GLP-1 journey tools, lab uploads. The strategic problem is that Whoop launched Advanced Labs through the same lab partner — both companies resell Quest — so this adjacency is convergent rather than differentiating. It is also membership-gated and US-only, which makes it an ARPU lever on the existing base rather than an acquisition channel. Model it by raising the membership price rather than the attach rate.

5 · The IPO itself

Timing is the whole question

The confidential S-1 went in on 21 May 2026, and Ōura's last private mark of $11B on roughly $2B of guided revenue is about 5.5× forward revenue — cheaper than Whoop at roughly 9× on $1.1B, despite Ōura being the larger business on both revenue and members. The case for going now is that revenue has quadrupled in two years, the company says it is profitable, Apple has retreated and the FDA has just widened what can ship without clearance. The case against is that the S-1 will have to disclose the attach rate, the real retention curve, the gross margin and the enterprise concentration — every number in the register above marked "not found" — into a market where the competitive set reorganised around undercutting the subscription eight weeks before filing.

6 · Defend the patent, or lose the royalty and the shield together

Binary, and outside the company's control

The patent behind the exclusion order is under post-grant challenge at the USPTO by Samsung and others. If it survives, Ōura keeps a royalty from RingConn and a credible deterrent against anyone else. If it falls, the royalty stops, the deterrent goes, and the eight months of exclusivity it bought against Ultrahuman turn out to be the entire return on the litigation. The outcome is not disclosed as of this writing and is not something Ōura controls — which is exactly why it belongs in a risk section rather than a model.

Synthesis. Ōura has the better structure of the two consumer-health models examined in this series: the customer funds their own acquisition, which is why it is profitable at a scale where service-only businesses are not. But that structure ties growth to unit sales in a category under 1% of wearables by volume, where it is the most expensive product and one of only two charging a recurring fee. The durable assets are the enterprise and clinical channels, which are the least visible part of the business and the least discussed in its own marketing. The S-1 will make the argument either way, and it will do so with numbers nobody outside the company has yet seen.