What is the five-pillar scorecard?
Weak-to-moderate fit: 1 of 5 pillars passes, 3 partial, 1 fail. Score 52/100, Neutral.
| PASS | Sunrise industry |
| PARTIAL | Leadership, Moat, Iron Fortress |
| FAIL | Free Cash Flow |
| 52/100 | Overall DalalBytes score |
| Pillar | Grade | One-line evidence |
|---|---|---|
| 1. Sunrise industry | PASS | Point-of-care ultrasound is a small market on an adoption S-curve; embedded ultrasound opens categories beyond medicine |
| 2. Leadership | PARTIAL | The current team delivers on guidance, but the SPAC record, the caveat-heavy control structure, and insulation by 71% founder voting power cap the grade |
| 3. Moat | PARTIAL | The only CMOS ultrasound chip at commercial scale, a software layer, and a patent estate; durability unproven against Exo plus Samsung and a subscription model rivals skip |
| 4. Iron Fortress | PARTIAL | No debt and $124.7M of cash, but about two years of runway, a $905M accumulated deficit, and a history of selling stock to survive |
| 5. Free Cash Flow | FAIL | No free cash flow yet; H1 2026 operating cash use was about $30M, and a $23.4M SBC line kept the dilution going even in a record year |
Pillar 1, Sunrise industry: why does it pass?
Point-of-care ultrasound is an early, expanding market by any definition found in the research: $4.2B to $6.0B in size today depending on who counts, with the handheld and compact segment forecast to compound at about 8% into the mid-2030s as image guidance becomes standard of care rather than a specialist service. The AI layer accelerates the curve: guidance software (compass-style QA, blind-sweep gestational age) removes the trained-sonographer bottleneck that historically capped the market.
Butterfly's own second sunrise is embedded ultrasound: full-body consumer scanning (Midjourney), brain-computer interfaces (Merge Labs, Bridge Neurotech), robotic needle guidance (Mendaera), each an application that did not exist as an ultrasound market five years ago. A company can ride one sunrise into a niche or two sunrises into a platform. The sunrise test is about the industry, not the company, and the industry is rising on both fronts. The deduction is size and timing: the handheld segment is only about $1.5B today, and the embedded categories will not be material industries until 2028 or later. Rising, but from a small base and on a partner's calendar. Pass.
Pillar 2, Leadership: why only partial?
Two regimes must be graded separately and then combined honestly. The founding regime gave the SPAC projections that a class action still litigates, an FY2022 net loss of $168.7M, a cost base of $232.5M that required two restructurings, and founder control of about 71% of the vote with about 2% economic ownership. That record fails the pillar outright.
The current operating regime is materially better. Joseph DeVivo (CEO since April 2023, ex-InTouch Health and Caption Health) cut costs by about $2M per month, beat FY2024 guidance, landed FY2025 inside the original revenue range with an EBITDA loss better than both the original and the cut guides, raised FY2026 guidance, and built the Embedded business from zero to a third of a record quarter. Heather Getz's resignation as CFO in August 2025 and the move to John Doherty (December 2025) is a churn debit in an otherwise stabilizing finance function, and the insider tape shows management selling into the 2026 rally while the sponsor's fund exited much of its position.
A proven operator working under a founder-control structure with a litigated past and a selling present grades partial. The upgrade path is visible (stated below), but pillars grade the file as it stands, not as it may become.
Pillar 3, Moat: why only partial?
The moat assets are tangible. The CMUT-on-CMOS chip is the only semiconductor ultrasound platform at commercial scale in the West, fabricated by TSMC, with a software-defined beam-steering architecture that turns image modes into software releases. The company claims more than 800 patents and applications (a 2022-era figure, not re-verified in this pass). Compass AI and the Garden developer ecosystem (30 AI partners) create workflow lock-in that a probe vendor cannot easily replicate. The Gestational Age clearance (March-April 2026) is the first FDA-cleared blind-sweep AI of its kind, an example of the software moat producing regulatory assets competitors do not hold.
Why not a pass. First, the moat has a same-architecture rival: Exo's Iris is a silicon single probe with Samsung Medison as an equity and co-development partner since May 2025, and Samsung's distribution and balance sheet are a moat of their own. Second, the subscription requirement ($299-420 per year) is a competitive wound Butterfly inflicts on itself against subscription-free GE, Mindray, and cheaper Chinese devices. Third, the Fujifilm Sonosite patent suit, pending since March 2022 on seven patents covering handheld-ultrasound fundamentals, is an unresolved challenge to the moat's legal walls; no outcome surfaced in the sources reviewed. Fourth, image quality per application is contested territory (third-party comparisons find no universal leader), so the probe itself is not an unbreachable wall. A real moat, narrowing at the edges while litigation and a same-thesis rival with a giant behind it test it: partial.
Pillar 4, Iron Fortress: why only partial?
The fortress frame starts strong: zero debt, $124.7M of cash at June 30, 2026, total liabilities of $93.6M that are mostly leases and deferred revenue, and equity of about $196M at year-end 2025. There is no lender who can force this company's hand, and the burn rate has narrowed every year since 2023.
The walls are thinner than the frame suggests. The accumulated deficit is $904.8M. H1 2026 operating cash use was about $30M, roughly two years of runway at that pace, and H2 2026 guidance (Q3 EBITDA loss $6-9M) does not fix it this year. The survival mechanism, demonstrated twice, is the equity market: the January 2025 offering sold 27.6M shares at $3.15, about 11% of the company, at a 10.3% discount, and share count has climbed from 205M (2023) to about 264M. A fortress that periodically sells its own bricks to stay standing grades partial: no debt to die by, but no self-funding surplus either. If Embedded cash conversion in 2027 pushes the company to sustained cash generation, this pillar upgrades first.
Pillar 5, Free Cash Flow: why does it fail?
The simplest pillar to grade. Butterfly has never produced free cash flow. FY2025 operating cash use narrowed to about $12.7M on one 10-K summary (management cites total cash use of $19.4M excluding offering proceeds), FY2025 adjusted EBITDA was still -$26.5M, H1 2026 used about $30M of operating cash, and the FY2026 guide still calls for a $19-23M adjusted EBITDA loss. One positive-cash-flow quarter (Q4 2025, +$6.3M, a management figure) shows the business can touch the line; it has not held it. Stock-based compensation of $23.4M in FY2025, roughly a quarter of revenue, means that even the EBITDA trajectory flatters the per-share economics. Fail, with a clear path out: the fail converts to partial the first year the company holds cash generation for four consecutive quarters without a raise, and to pass when free cash flow per share turns durably positive.
What would change the grades?
| Pillar | Upgrade trigger | Downgrade trigger |
|---|---|---|
| Sunrise | Handheld and embedded ultrasound categories reach disclosed, material scale (partner scanners shipping in volume) | POCUS adoption stalls on reimbursement; embedded categories slip past 2028 |
| Leadership | A full guidance cycle delivered with the CFO seat stable and insider tape neutral-to-buying | Another guidance cut on "timing," or founder-level interference with the operating team |
| Moat | Sonosite suit resolved favorably; Garden AI revenue becomes a disclosed line; Exo fails to scale | Adverse litigation outcome, or Exo plus Samsung winning system-wide deals Butterfly contests |
| Iron Fortress | Runway extended past three years on operating improvement, not issuance | A raise at a discount before Embedded cash lands |
| Free Cash Flow | Four consecutive cash-generative quarters with no raise | Burn re-widening if milestone recognition outruns cash collection |
The five-pillar read agrees with the report verdict: a sunrise and a partial-moat story priced ahead of its cash proof. 52/100, Neutral. This pillar set upgrades quickly if 2027 delivers commercialized partners and cash, and degrades just as quickly if the licensing line proves to be one partner deep.