From 5.5M to 100M. The architecture the milestone needs.
95M new users in 24 months. Hardware can't move that volume of physical units. The only motion that delivers users in millions per deal is concentrated B2B — payers, employers, pharma. The 18–24 month window is set by Apple's HealthKit rollout and Samsung's sub-$200 ring: move first or move never.
Today
5.5M
Hardware-led DTC
Hardware leads · Continuum absent
10M users
10M
Inference layer · B2B wedge opens
Hardware partners · Continuum is the contract
50M users
50M
Inference protocol · network effects compound
Hardware peer · vendors compete on UX, not data
100M users
100M
The biological inference layer · default infrastructure
Hardware peripheral · interchangeable
10M at 6 months. 50M at 15 months. 100M at 24 months. Compressed because the window is.
B2B revenue exceeds B2C at 10M (the architecture pays for itself). Network effects compound at 50M (cross-vendor agreement sharpens with every new vendor). Protocol-default at 100M (Continuum is the substrate every health-adjacent product runs on). The pace is set by Apple — not by us.
Three inflections in 24 months: B2B revenue exceeds B2C at 10M (~6 months in). Network effects compound at 50M (~15 months). Protocol default at 100M (~24 months). Each milestone is gated by a B2B contract or a vendor-integration unlock — none are hardware-led.
What unlocks at each milestone.
Each phase is gated by a structural unlock. Without the unlock, the milestone doesn't materialize — and the next milestone after it doesn't exist.
5.5M
Hardware-led DTC
arr · ~$500M
Direct ring sales + subscription unlock. Continuum doesn't exist yet inside Oura — readiness, sleep, recovery scores all live in the ring app. Hardware is the moat; the inference layer is the unbuilt opportunity.
unlocks
- ·Best-in-class hardware
- ·Premium subscription model
- ·Strong consumer brand + post-acquisition retention
revenue model
Hardware + B2C subscription
comparable
Pre-platform Stripe (2011) — payment processor before ecosystem
risk if missed
Samsung Galaxy Ring at sub-$200 commoditizes the ring within 12 months. Apple Vitals integrates HealthKit into iOS at scale. Without an inference layer, hardware moat erodes faster than the next subscription tier can replace it.
10M
Inference layer · B2B wedge opens
arr · $1B ARR at ~10M lives
The Phase-2 unlock isn't user growth — it's 5 concentrated payer / employer / pharma deals. Each closes a 9-12 month sales cycle. Reaching $1B ARR doesn't require 100M users; it requires 5 right rooms.
unlocks
- ·1 Aetna deal: ~3M lives × $40/yr = $120M ARR
- ·1 Elevance / Anthem deal: ~4M lives × $40/yr = $160M ARR
- ·2 Fortune 500 self-insured employers · 1M lives × $50/yr = $100M ARR combined
- ·1 pharma attribution deal (GLP-1 cohort, supplement validation): $100M+ floor
- ·Long-tail · 50 mid-market employer deals × ~$10M each = $500M ARR
- ·Total at ~10M concentrated lives: ~$1B ARR
revenue model
Concentrated B2B contracts. Employer-wellness PEPM at $30-50/life/yr; pharma per-protocol pricing. 5 lighthouse deals carry the first $1B.
comparable
Castlight Health early B2B trajectory · ~$8-12 PEPM at scale, but Continuum is platform-only (no nav services), so blended PEPM lands $30-50.
risk if missed
Without B2B wedge live in this window, Apple HealthKit becomes the de facto aggregator. Once Apple owns the schema, Oura is back to selling rings against a free aggregator built into every iPhone.
50M
Inference protocol · network effects compound
arr · ~$2.5B
Continuum is the cross-vendor inference protocol. Apple Watch, Whoop, Garmin, Fitbit, Withings, Eight Sleep — all integrated via the same standardized 23-field schema. Pharma uses Continuum for outcome attribution at scale (GLP-1 efficacy, sleep-aid trials, supplement cohorts). Insurance underwrites continuous risk scoring. Each new vendor sharpens cross-vendor agreement, which sharpens inference quality, which attracts more vendors — the protocol flywheel.
unlocks
- ·Pharma outcome-attribution contracts (~$2B addressable)
- ·Continuous-risk-scoring integration with health + life insurers
- ·Apple/Samsung schema interoperability — forced by reimbursement pressure on hardware vendors
- ·Lifetime baseline (cross-vendor data per user) becomes the structural moat
revenue model
Per-seat + variable (pharma, insurance) + developer API
comparable
Stripe 2018 compressed — developer ecosystem + enterprise contracts compounding
risk if missed
If Continuum isn't the default inference layer at 50M, the aggregation moat closes for everyone except whoever wins the protocol race. Closed ecosystem (Apple) wins by default. Oura captures rings; everyone else captures the platform.
100M
The biological inference layer · default infrastructure
arr · $2-5B+ ARR (range depends on premium-tier scale)
At 100M users, the blended ARR comes from three cohorts, not one. Consumer end-users (free / freemium / partner-bundled) at $1-5/yr. Employer + payer wellness members at $30-50/yr. Premium clinical / pharma-attribution / longevity-clinic members at $80-200/yr. Plus pharma per-protocol revenue. Blended ARPU $18-38, ARR $2-5B+ depending on how aggressively the premium-tier cohorts scale.
unlocks
- ·60M consumer end-users · $1-5/yr · $60M-300M ARR
- ·30M employer/payer members · $30-50/yr · $900M-1.5B ARR
- ·10M premium clinical/pharma members · $80-200/yr · $800M-2B ARR
- ·Pharma per-protocol attribution · $0.5-1.5B incremental
- ·Conservative blended: $2B · Mid-case: $3-4B · Upper bound: $5B+ if premium cohorts scale
revenue model
Three-tier cohort mix · consumer freemium + employer/payer PEPM + premium clinical-managed. Pharma per-protocol incremental. Blended ARPU $18-38.
comparable
Stripe at maturity captures ~$5-15/user-equivalent ARPU on $1.4T GTV. Continuum captures more per-user because biological inference is more concentrated value than payment routing. Closer analog for the premium tier: Omada / Livongo per-engaged-user economics applied to a platform layer.
risk if missed
The aggregation layer always consolidates to one player. In 18–24 months that player exists. The only question is whether Oura ships it from inside or watches Apple, a startup, or both ship it from outside. There is no third option.
Why hardware-only doesn't scale to 100M.
Every assumption that gets us to 100M depends on a non-hardware growth path. Here's the structural argument — four reasons the ring alone doesn't reach the milestone Michael set.
Hardware can't add 95M users in 24 months
From 5.5M to 100M in two years means adding ~4M users per month. Even the fastest-shipping wearable companies in history couldn't move that volume of physical units through retail and DTC. The math forces a B2B distribution channel — there is no hardware-led path to this milestone in this timeline.
Apple is the gravity well — and the timer
Vitals is rolling into iPhone right now. Apple ships hardware-adjacent inference for free at iOS scale. Every month Oura waits is a month Apple owns more of the schema. The 18–24 month window exists because that's how long Apple's rollout takes to consolidate. Move first or move never.
Samsung commoditizes the ring
Galaxy Ring at sub-$200 collapses the price moat by 2027. Hardware differentiation becomes battery life and form factor — undurable. The data layer is where investment compounds across the same 24-month window.
B2B is the only motion that delivers users in millions
One Aetna deal = ~3M members. One UnitedHealth deal = ~5M. One major self-insured employer (Walmart, Amazon, Target) = ~1–2M employees. Three-to-five concentrated B2B contracts deliver tens of millions of users in the same time it takes to ship a ring redesign. The aggregator math only works through concentrated buyers.
At 100M, Continuum is the protocol. Oura is the brand on top.
Stripe is to payments what Continuum becomes for biological inference. Invisible infrastructure every consumer-health and clinical product runs on. Hardware vendors compete on form factor and battery life — the inference layer is where the value compounds.
Oura's brand survives — and thrives — in this future. The same way Shopify runs on Stripe and captures the consumer moat, Oura captures the consumer wearable moat while Continuum captures the platform moat. Two layers, same parent. Hardware revenue stays; platform revenue is added on top.
$2–5B+ ARR at 100M users — a range, because the upper bound depends on how aggressively the premium-tier cohorts scale. Conservative: $2B (consumer + employer/payer only). Mid-case: $3–4B (with measured premium clinical). Upper bound: $5B+ if pharma attribution and longevity-clinic cohorts scale per Omada / Livongo per-engaged-user economics. Even the conservative case is category-defining.
The window is 18–24 months. Apple is rolling HealthKit into iOS at scale right now. Samsung is shipping a sub-$200 ring. The aggregation layer always consolidates to one player — and in two years that player exists. The only question is whether Oura ships it from inside, or watches Apple, a startup, or both ship it from outside. This artifact is the answer, built in 4 days.
Same architecture. Two doors.
Run from inside Oura: you become the platform aggregator before Apple does. Run as a startup: this is fundable as a Plaid-shape B2B infrastructure play. The question Michael asked Friday has a 4-day answer. The question Tom needs answered is which door you walk through.
// the math
phase today → 100M
{
"users": 2.5M → 100M,
"arr": $200M → $5B+,
"timeframe": 18–24 months,
"wedge": B2B → protocol,
"hardware": lead → peripheral
}
// architecture compounds; hardware doesn't