Data confidence per node
Verified — funding / M&A confirmed by primary sources
Estimated — revenue / users from 3rd-party trackers
Modeled — gross margins (never disclosed by these firms) Ring around a marker = that year's financials are estimates.
Read before citing. Funding rounds & acquisitions are verified. All gross-margin figures are
modeled assumptions — none of these private companies disclose margins. Revenue & user counts are
third-party estimates. Hover any point for its per-metric confidence.
Beyond the landscape: what the numbers actually say
Cross-company structural analysis — monetization depth, capital efficiency, and business-model evolution.
ARPU = revenue ÷ users/lives; capital efficiency = revenue ÷ total capital raised. Directional — gross margins are modeled, several revenue/user figures are third-party estimates.
Company
Model
ARPU / relationship / yr
Rev per $1 raised
GM arc
Who actually pays
Athelas + Commure
Enterprise health OS
~$400–600/life · ~$20k+/seat
~$0.31 (bought via rollup)
45% → 74%
Health system (seats + RCM)
Medvi
DTC cash-pay + AI automation
~$1,604/patient (~$134/mo)
~$20,050 (near-zero capital)
25% → 32%
Patient, upfront, monthly
Whoop
DTC / enterprise wearable
~$245/subscriber
~$0.64
30% → 76%
Consumer (subscription)
HealthSnap
EHR-native RPM/CCM
~$237/life
~$0.79 (efficiency leader)
35% → 71%
Provider (CMS CPT billing)
TytoCare
B2B2C virtual clinic
~$127/life
~$0.40
38% → 63%
Payer / health plan
AliveCor
DTC + enterprise cardiac
~$38/user
~$0.53
42% → 63%
Consumer + self-insured employer
Withings
Connected hardware + B2B
~$35/user
~$0.54
35% → 51%
Consumer + RPM program
Sort by ARPU and the story is in one column: device makers (AliveCor $38, Withings $35) sit on the floor; stack and demand owners (Athelas, Medvi) sit at the ceiling. The sensor is the cheapest thing in the value chain.
Seven derived observations
Athelas monetizes the same patient twice.
Once as a monitored life (device telemetry), again as a provider seat (RCM + AI scribe the clinic pays for). One encounter, two revenue lines — which is why its per-patient economics dwarf single-sided players. The device is just the wedge into the clinic's billing system.
→ You don't win by owning the patient's body; you win by owning the clinic's invoice.
Hardware is a customer-acquisition cost, not a profit center.
Every company that stayed device-first is stuck at the bottom of ARPU and margin (AliveCor, Withings). Every company that climbed the margin axis did it by subtracting hardware from the revenue mix — embedding it (Athelas' seats), abstracting it (HealthSnap's EHR layer), or giving it away (Whoop's subscription).
→ The 2018–2024 margin climb isn't a technology story. It's a subtraction story.
Capital efficiency runs opposite to ARPU.
The highest-ARPU incumbent (Athelas, ~$400–600) is the least capital-efficient ($0.31) — it bought its stack with a $6B rollup and $670M raised. The most efficient real company (HealthSnap, $0.79) engineered mid-ARPU with software leverage.
→ Capital raised measures how much of the stack you had to buy instead of build — not quality.
Margin and durability are not the same axis.
Medvi has the lowest margin (25–32%) yet the highest cash return and ARPU — and is the most fragile (regulatory cliff). Whoop has near-software margin (76%) but burned $425M buying retention. Durability lives in the middle — TytoCare and HealthSnap, anchored to reimbursement.
→ High margin bought with capital and high cash bought with arbitrage are both rented, not owned.
The reimbursement rail is the moat, not the margin number.
TytoCare could raise again in 2026 into chronic care despite unreliable revenue — because payers fund proven ED-visit reduction (–24.1%). Anyone paid by the system they save money for (payer) outlasts anyone paid by discretionary spend (Whoop) or a regulatory window (Medvi).
→ Be paid by the entity you save money for, not the one you sell to.
Everyone is converging on the same corner from different sides.
TytoCare / AliveCor / Withings climb up from hardware; Whoop climbs up from consumer subscription; Athelas descends from enterprise software into hardware+workflow; Medvi teleports in through marketing. The plot converges on high-revenue × high-margin — but the capital axis exposes who earned the ticket and who bought it.
→ Same destination, and the Z-axis is the honesty check.
The next margin shock is labor, not hardware.
The CY2027 staffing-ban proposal punishes players whose ARPU is built on selling monitoring labor (HealthSnap's turnkey nursing) and rewards software leverage (one nurse, 100+ panels). The frontier that moved from cost-of-hardware (2018–24) moves next to cost-of-humans (2026–30).
→ "Kill the device" made the last decade of margin. "Kill the manual touch" makes the next.
The Landscape — where the market has not been built
Players plotted by layers of the stack owned (→) × audit- vs. marketing-grade compliance (↑). Nearly everyone owns 1–2 layers or wins on marketing; the top-right — full rail × audit-grade — is empty. That whitespace is the thesis: Ausa owns all six layers, sensor to escrow, audit-grade.
Bubble size ≈ market presence / funding. Hover any bubble for what it does and why it's here.