Suki
B2B enterprise SaaS on a fixed per-clinician seat licence: reported list $299/month (Suki Compose, notes only) and $399/month (full Suki Assistant with deep EHR write-back). Third-party/reseller reported — Suki publishes no price list and enterprise deals carry undisclosed volume discounts, so the rungs are upper bounds, not contracted ASPs. A second B2B2B motion, 'Suki for Partners', licenses the speech/AI engine to vendors who embed it: athenahealth's Ambient Notes, HealthEdge GuidingCare, AvaSure, WellSky.
Earnings, margins, COGS & capex
Mid-scale private growth company, clear #4 in a category whose top two are capitalised 2-6x heavier. The distinguishing fact is PRICE: a reported $299-399/clinician/month ($3,588-4,788/yr) is 1.79-2.38x the $2,008/yr point estimate of measured marginal E/M revenue and 1.20-4.61x across that measurement's 95% interval ($1,038-$2,979/yr). Suki and Ambience's full suite are the ONLY rungs above 1.0x at every point; Abridge (0.84-2.41x), Ambience's base tier (0.94-3.08x), Nabla (0.48-1.38x) and Freed (0.40-1.14x) straddle parity. That is a wide-interval observation about a thin surplus, not demonstrated overpricing — the JAMA authors call $167.37/clinician/month a 'conservative lower bound of the financial benefits' and say the analysis 'cannot generalize to cost-benefit considerations', and every list price is a trade-press estimate whose error runs against the vendor. Suki's own evidence lands on both sides: Rush's estimated $202/user/month sits inside the JAMA 95% CI; its marketing headline of $1,688 per user per month is ~10x the point estimate.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~25¢ is cost of goods and ~0¢ operating expense, leaving ~75¢ of operating profit.
Revenue trend
Margins
compressed by per-encounter inference and four EHR integrations; falling token prices are the tailwind
spending to hold ~10% share
equity-funded
improves only if list falls or the measured benefit is revised up
COGS structure
Two variable inputs per encounter — speech recognition and LLM note/code generation. Suki mixes internally developed speech models with partner models under a Google Cloud partnership, lowering per-minute ASR cost but converting the saving into R&D headcount. The Suki-specific drag is integration depth, the thing it markets as its edge: real-time write-back into Epic (Haiku and Hyperspace via Suki INSIDE / Epic Toolbox), Oracle Health, athenaOne and MEDITECH Expanse is four certification and upgrade surfaces on four release cadences — cost that does not fall with token prices. Epic's marketplace toll is trivial (~$1,700-1,900 per vendor per year); staying integrated is the expense. Coding output adds clinical-review and audit cost a note generator never carries.
Capex
Immaterial. No owned data centres; compute is cloud opex. Capital goes to engineering headcount, clinical/coding review and enterprise GTM.
Latest earnings
n/a
None. Directional only: usage tripled in 2025, 400+ organizations, 150+ features shipped, expansion into coding, nursing and payer care management.
- Reported list price
- $299 (Compose) - $399 (Assistant) /clinician/month = $3,588-4,788/yr; reported list, no primary source
- Price / measured value
- 1.79-2.38x the $2,008 point estimate; 1.20-4.61x across the 95% CI
- US category share
- ~10% of ~$600M 2025 US ambient revenue — #4 behind Microsoft/Nuance 33%, Abridge 30%, Ambience 13%
- Funding / marks
- $168M cumulative (company-stated, Jan 2025); $70M Series D Oct 2024, no post-money; secondary ~$400-500M
- Attach: measured vs marketed
- Rush est. $202/user/month (inside the JAMA 95% CI of $86.52-$248.21) vs a marketed $1,688/user/month
- FDA status
- No 510(k) clearance, no PMA approval, no De Novo — marketed outside device regulation, so no regulator has evaluated accuracy
Growth drivers
- Four-EHR real-time write-back — Epic via Toolbox, Oracle Health, athenaOne, and the first fully integrated ambient solution on MEDITECH Expanse
- The 'Suki for Partners' engine — athenahealth's Ambient Notes, HealthEdge GuidingCare, AvaSure and WellSky embed Suki
- athenahealth Preferred Solution Partner status (26 Jan 2026) — 450+ practices, 3,400 MAU, +50% in six months, 82% adoption
- Point-of-care coding from the ambient capture (Fall 2025), with a stated 48% reduction in amended encounters
- Personas no ambient vendor has monetised at scale: the nursing consortium and payer-side care managers
- Land-and-expand at named systems — Rush across 28 specialties, MedStar, Austin Regional Clinic, 12+ MEDITECH systems
Bull & bear
Suki is the only ambient vendor selling the engine as well as the seat — the right shape for a category being absorbed into the system of record. As Epic, athenahealth, MEDITECH, payers and virtual-care vendors each decide they need ambient inside their own product, the vendor with four real-time EHR write-backs and a licensable platform gets paid on the bundling instead of killed by it, at roughly one-sixth the leader's revenue multiple.
- The platform motion is proven in the wild: athenahealth's own Ambient Notes runs on Suki, HealthEdge put it inside GuidingCare, and AvaSure and WellSky embed it
- Integration breadth is the durable asset — real-time write-back across four EHRs reaches install bases Epic-concentrated rivals do not contest
- Rush went pilot to enterprise-wide across 28 specialties, 74% of clinicians reporting reduced burnout and 95% wanting to continue; KLAS validated ROI across three systems
- Coding is the honest escape from the price/value trap, and the JAMA authors call $167.37/month a conservative LOWER bound that 'cannot generalize to cost-benefit considerations'
- Suki's own named-customer number is the credible one: Rush's est. $202/user/month sits inside the JAMA 95% CI and above the point estimate, from a buyer that then expanded — and PHTI, whose March 2025 assessment found 'gaps in evidence' on ambient ROI, reported in April 2026 that one health system turned the same ~5% Level-5 shift into 'over $1,000 per provider per month', ~6x the JAMA point estimate. Single-site, self-reported and uncontrolled, so $202 is the conversion to underwrite
- Cheapest expression of the theme: ~$400-500M on ~$60M canon-implied revenue is ~7-8x against ~45x on Abridge and ~40x on Ambience
Suki holds the worst position on the board's only quantified axis — the most expensive independent seat, defended by the least capital, in the one category where the system of record has already shipped the same function inside a contract the buyer holds. It has never disclosed a revenue figure, has no fresh mark since October 2024, and its marketing claims a per-user benefit roughly ten times what the only controlled study measured.
- The most exposed rung: 1.79-2.38x the point estimate, 1.20-4.61x across the interval, above 1.0x everywhere — while Abridge, Ambience's base tier, Nabla and Freed straddle parity and can meet a CFO in the middle
- Substitutes are free or bundled and arrived on schedule: Epic AI Charting GA 4 Feb 2026 with Microsoft supplying the transcription; athenaAmbient at no additional cost in the network that just named Suki a preferred partner; Doximity Scribe at $0
- Capital asymmetry decides commoditising categories: $168M against ~$1.07B (~$757.5M excluding a press-reported, never-announced April-2026 extension) and $345M, no post-money since Oct 2024, and the next round struck against a $0 rung
- The marketing gap repeats the ~6x spread between Dragon Copilot's marketing and the controlled measurement: $1,688/user/month against a controlled $167.37 (95% CI $86.52-$248.21), while its own Rush figure was $202 and PHTI's April 2026 counter-datum of 'over $1,000 per provider per month' is single-site, self-reported and uncontrolled
- The measured benefit is small and unevenly consumed: 16.0 fewer documentation minutes and 13.4 fewer total EHR minutes per 8 scheduled patient hours, no change after hours, +0.49 weekly visits, only ~32% of adopters using it on >=50% of visits — the threshold associated with roughly 2x the EHR-time and 3x the documentation-time reduction, which is why the authors call $167.37 conservative and equally why an enterprise buyer paying for every seat realises the population average
- The E/M channel is worth ~$0 where the clinician is not paid fee-for-service — Kaiser, the VA, staff-model plans, full-risk MA/ACO — removing the coding argument from the largest centrally procured accounts
What it is worth
No market price exists. Anchor is the last primary round (Oct 2024 Series D, $70M led by Hedosophia, post-money NOT disclosed) plus conflicting secondary marks (~$500M per Sacra's 2025 figure, ~$400M elsewhere on the same round), sanity-checked against canon-implied revenue (~10% of a ~$600M 2025 US category, ~$60M) and peer private marks (Abridge ~$5.3B on ~$117M contracted ARR, ~45x; Ambience ~$1.25B on ~$30M ARR, ~40x). No DCF is possible — no revenue, margin or cash-flow disclosure exists.
<$250M
or a structural outcome rather than a mark: Epic AI Charting and athenaAmbient absorb the ambulatory seat, Doximity's $0 and the cheap paid rungs reset what a note is worth, share slips from ~10%, and $168M cannot fund the response — resolving as a flat-to-down round, an acqui-hire of the speech/AI platform, or an asset sale of the partner book.
~$400-600M
the existing mark carried forward. Suki holds ~10% share, grows with the category rather than through it, discounts enterprise seats quietly against Epic AI Charting and athenaAmbient, and the multiple stays low because no revenue or retention figure is disclosed to argue against.
~$1.0-1.5B at the next primary
requires the platform motion to become the story (disclosed revenue from the athenahealth, HealthEdge, AvaSure and WellSky embeds plus nursing and payer seats), independent validation of the coding attach, and list held. At ~$150M+ of revenue that is 7-10x, still a discount to today's leader marks.
Both sides of the fraction are soft, so the honest output is a range. The ~$400-500M figure is a NEGOTIATED MARK from a round that closed in October 2024 — ~22 months stale, struck before Epic AI Charting reached GA — and secondary trackers disagree on it; it is not a clearing price and must not be inferred from the $70M raise. The denominator is a share-implied derivation from a VC estimate relayed by trade press. On those two soft numbers Suki screens at ~7-8x revenue against ~40-45x for Ambience and Abridge, and that gap is the question: either the market is correctly pricing the #4 in a commoditising category that also holds the most exposed price/value position on the board, or the leaders' marks are the stale ones and the engine-licensing motion is the shape that survives bundling. A low multiple in a category whose next round must be struck against a $0 rung is as likely a warning as a discount. Not investment advice; no price target is expressed or implied.
SWOT
Strengths
- Broadest real-time EHR write-back of any independent ambient vendor — Epic, Oracle Health, athenaOne, MEDITECH Expanse
- 'Suki for Partners' licenses the engine, so athenahealth, HealthEdge, AvaSure and WellSky distribute Suki rather than compete
- Third-party outcome evidence — a KLAS ROI validation across FMOL Health, McLeod Health and Rush on longitudinal Epic Signal data
- Product surface beyond the note — order staging, patient summaries, clinical Q&A, voice editing, point-of-care coding
- Persona expansion nobody has commercialised at scale — a nursing consortium with named systems, plus payer-side care managers
Weaknesses
- Top of the independent ladder with the least capital to defend it: above 1.0x the measured benefit at every point in the 95% interval
- Capital asymmetry is structural — $168M against Abridge ~$1.07B — ~$757.5M of announced rounds plus a press-reported, never-announced $316M April-2026 extension — and Ambience $345M, inside ~$1.71B across four vendors, or ~$1.39B excluding that extension
- No disclosed post-money since Oct 2024 — ~22 months without a fresh mark
- No revenue, ARR, gross-margin or retention figure has ever been disclosed
- No randomised evidence for its own product — the category's one RCT tested Nabla and DAX Copilot and did not replicate across both arms
Opportunities
- Point-of-care coding is the only route to defending a seat on hard dollars — validated, it argues price against total reimbursement, not the E/M slice alone
- The payer seat via HealthEdge reaches a utilisation-management buyer — the one channel immune to the capitation carve-out that zeroes the E/M attach
- Nursing is the largest unmonetised ambient persona, built consortium-first rather than retrofitted
- Being the embedded engine inside other vendors' products turns bundling — the category's central threat — into a tailwind
- MEDITECH and Oracle Health install bases are under-served by Epic-concentrated rivals
Threats
- Epic AI Charting reached general availability 4 Feb 2026, licensed inside the EHR relationship at an undisclosed price — not free, but it needs no second contract, BAA or renewal; Epic held 43.7% of US acute-care hospitals in 2025
- athenahealth ships athenaAmbient at no additional cost inside the core fee — Suki's newest channel hosts a $0 substitute
- Doximity Scribe is $0 to every verified US clinician — it need only be good enough to make a $3,588-4,788/yr seat need justifying
- Menlo Ventures found large systems 'just as likely to switch vendors as to stay' and buyers calling scribing 'becoming commoditized'
- ~60 ambient products at PHTI's March 2025 count, against a bottom-up US terminal pool of ~$1.0-1.2B/yr — what the category would bill at full penetration at a price justified by the measured value. The ~$600M booked in 2025 was billed at today's list prices, a different basis: on the $3-6B/yr those list prices imply at full penetration it is ~10-20% penetration, not half (canon PB-060 — both figures are live, neither is interchangeable)
Moats, dependencies & bottlenecks
Moats
Oracle Health, athenaOne, MEDITECH Expanse) Strong on breadth, medium on depth re-earned every EHR release The one axis where Suki leads, but integration is permission the EHR grants; Toolbox and preferred-partner status are revocable.
structurally differentiated Medium-high once embedded athenahealth, HealthEdge, AvaSure and WellSky already embed it — the only moat here that gets stronger as the category bundles.
Rush's 95% continue-use is real, but Menlo found large systems 'just as likely to switch vendors as to stay'.
the best candidate for a future moat 48% fewer amended encounters is company-stated; validated it re-bases the price argument off the $2,008 E/M slice, but today it is a claim.
Weak relative to the field #4 at ~10% share on $168M, against a 33%-share incumbent inside Microsoft and a 30%-share leader with ~$1.07B raised, ~$757.5M of it announced.
Dependencies
Integration channel, gatekeeper AND competitor Suki INSIDE runs in Haiku and Hyperspace via Epic Toolbox; Epic held 43.7% of US acute-care hospitals in 2025 on $6.7B of 2025 revenue (company-stated to Becker's 28 Apr 2026, unaudited — Epic files nothing, no margin knowable) and its AI Charting reached GA 4 Feb 2026. The fee is trivial; the dependency is the designation.
Preferred-partner channel, OEM customer AND competitor Preferred Solution Partner (26 Jan 2026) and Suki powers athenahealth's own Ambient Notes — while the same network ships athenaAmbient at no additional cost.
Channels into non-Epic install bases MEDITECH Expanse is the clearest whitespace (12+ systems); Oracle Health is building native clinical AI on a base that lost 56 hospitals in 2025.
model and cloud providers (Google Cloud/GOOGL; GPUs via NVDA) Core technology and per-encounter COGS Internal speech models mixed with partner models; falling token prices are a margin tailwind, a rival's model step-change the risk.
Funding and the ROI argument itself No post-money since Oct 2024, so the next primary is the repricing event; and the coding attach is worth ~$0 in capitated settings.
Advantages
- Broadest real-time EHR write-back of any independent vendor, including the only full MEDITECH Expanse integration
- A licensable engine as well as a product — the moat that strengthens as bundling accelerates
- Third-party KLAS ROI validation across three named systems on longitudinal Epic Signal data
- A named-customer attach ($202/user/month at Rush) inside the only controlled measurement's confidence interval
- Cheapest revenue multiple among the top four private names (~7-8x vs ~40-45x)
Weaknesses
- The most exposed rung on the price/value ladder — above 1.0x at every point in the 95% interval
- Least-capitalised of the big four at $168M, no disclosed post-money since Oct 2024
- Zero disclosed revenue, ARR, gross margin or retention — every financial view is derived
- A marketing revenue claim (~$1,688/user/month) roughly 10x the only controlled measurement
- Its newest and best channel, athenahealth, simultaneously ships a no-additional-cost substitute
Bottlenecks
- Four deep EHR integrations are four certification and upgrade treadmills — cost that does not fall as inference gets cheaper
- A per-seat licence bills for adoption the system may not achieve — only ~32% of JAMA-cohort adopters used their scribe on >=50% of visits — the threshold associated with roughly 2x the EHR-time and 3x the documentation-time reduction, so the buyer pays for every seat and realises the population average
- Coding output raises the audit bar — a wrong CPT/E&M suggestion is a compliance exposure, not a typo
- Long procurement and BAA cycles, and a stronger second-contract objection now the EHR ships the function
- No published randomised or independently controlled evidence for Suki's own product
- Capital: $168M cannot fund a price war against a $0 rung and a 33%-share incumbent inside Microsoft
Top signals & trends
Top signals
Dated, quantified traction validating the engine-licensing motion; discount for athenaAmbient shipping at no additional cost in the same network.
First disclosed payer-side seat — the one channel the capitation carve-out does not zero, because the buyer is not billing E/M.
The most consequential dated event against every independent seat licence, hardest at the top of the ladder.
An ~8x internal spread and ~10x the $167.37 point estimate (95% CI $86.52-$248.21) — the shape of Dragon Copilot's ~6x marketing gap. Rush is the defensible figure; PHTI's April 2026 report of 'over $1,000 per provider per month' at one health system is single-site and uncontrolled.
Absence of a mark is not a down mark, but it removes the only external check on value.
Third-party-collected evidence beats a vendor case study — still three organisations, non-randomised and commissioned.
Category-level, landing hardest on the highest-priced rung; nine to ten months old at 6 August 2026.
Trends
Suki's partial hedge is to be the engine those systems license rather than the app they displace.
Squeezes the top hardest. Dragon Copilot is the exception at a reported $369-604/provider/month — above Suki, and increasingly an OEM component inside Epic.
The authors call it a conservative lower bound that 'cannot generalize to cost-benefit', so it disciplines pricing without settling it.
The only path to a defensible enterprise seat price; Suki shipped it from Fall 2025.
New seats at systems that already bought ambient once; Suki is ahead of most rivals here.
The coding attach that justifies a premium seat is the same attach payers are most likely to claw back.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Cloud and speech/NLP partner; Suki also runs internally developed speech models, so ASR is build-and-buy.
Anthropic, Google) LLM inference for notes, coding and Q&A — per-encounter token cost is the main variable COGS.
Cloud + GPU infrastructure (AWS / Azure / GCP; NVIDIA) Hosting, HIPAA-eligible services and GPU capacity; Azure is MSFT, parent of Suki's largest competitor.
Pilot to enterprise-wide across 28 specialties (3 Mar 2025): +10% encounter volumes, ~5% rise in Level-5 coding, est. $202/user/month, 74% reporting reduced burnout.
Marquee health-system partnership, expanded alongside the $70M Series D (Oct 2024).
The other two KLAS ROI-validation sites (2026); McLeod is also a nursing-consortium member.
3,400 monthly active users, ~500k notes/month, 82% adoption as of 26 Jan 2026 — the highest-velocity disclosed channel.
Austin Regional Clinic, and the nursing consortium (Boone Health, Citizens Memorial, Fisher-Titus) The least-contested install base, an extended ambulatory partnership, and design partners for Suki for Nurses — not disclosed paid seats.
Category #1 at a reported ~33% share and 100,000+ clinicians daily, reported list $369-604/provider/month (trade-press, not a Microsoft disclosure). Also supplies the transcription inside Epic's AI Charting. One of the two NEJM AI RCT arms where the effect did NOT reach significance (-1.7%, 95% CI -9.4 to +5.9).
Private, ~30% share, marked ~$5.3B on ~$1.07B raised ($757.5M announced plus a reported $316M April-2026 extension) — leader by logos, brand and Epic depth, at an estimated ~$2,500/clinician/yr (Abridge publishes no pricing; 1.24x point, 0.84-2.41x interval), so it can undercut Suki's list and still hold a premium.
Private, ~13% share, ~$1.25B mark on $345M raised. The only other rung above 1.0x across the whole interval (full suite $4,000-5,000/yr = 1.99-2.49x point, 1.34-4.82x interval), and the closest rival on coding.
Private EHR incumbent — 43.7% of US acute-care hospitals in 2025, $6.7B of 2025 revenue (company-stated to Becker's 28 Apr 2026, unaudited — Epic files nothing, no margin knowable). AI Charting reached GA 4 Feb 2026, licensed inside the EHR relationship at an undisclosed price. Simultaneously Suki's channel and its most existential competitor.
Private. Named Suki a Preferred Solution Partner (26 Jan 2026) and powers Ambient Notes with Suki's engine — while shipping athenaAmbient at no additional cost inside the core fee.
Private, ~4% share, $120M raised, ~$1,428/yr (0.71x point, 0.48-1.38x interval). Holds the category's only positive randomised result — -9.5% time-in-note (95% CI -17.2 to -1.8, P=0.02) — at a third of Suki's list.
$0 to every verified US clinician — the free rung. It need not beat Suki, only be good enough to make a $3,588-4,788/yr seat need justifying.
Private, ~4% share, $99/month = $1,188/yr (0.59x point, 0.40-1.14x interval) — PLG into small practices, expanding upmarket; the cheapest paid rung.
#2 EHR building native clinical AI, and the largest net acute-care share loser three years running (-56 hospitals, -14,676 beds in 2025) — both a channel and a bundling threat.
Private. Commure sits beside Suki in Epic Toolbox's ambient category; DeepScribe is specialty-tuned, Heidi self-serve — the long tail among ~60 ambient products at PHTI's March 2025 count.