Butterfly Network, Inc. (BFLY) stock intrinsic value fair value estimate & owner earnings analysis

Butterfly Network makes handheld, point-of-care ultrasound devices built around its own semiconductor ("Ultrasound-on-Chip") technology, paired with software subscriptions that let clinicians use, store and analyze scans. Beyond its own hardware, the company increasingly licenses that chip and software platform to other companies that want to embed ultrasound into their own products ("Butterfly Embedded"), including a November 2025 agreement with Midjourney to explore applications such as brain-computer interfaces and surgical robotics. Revenue has grown for three straight years -- $65.9 million in 2023, $82.1 million in 2024, $97.6 million in 2025 -- and grew faster still, 33%, in the first half of 2026. The company carries no debt and held $150.5 million of cash at the end of 2025, funded mostly by an equity raise completed in January 2025 rather than by operations. The business is not yet cash-flow positive. After stripping out swings in working capital and the cost of stock-based compensation -- both of which can make a given year's cash flow look better or worse than the underlying business really is -- owner earnings were meaningfully negative in every year of the last five, though the loss has narrowed sharply: from roughly $(1.11) per share in 2021 to $(0.22) per share in 2025. Some of 2025's improvement, however, came from a one-time working-capital benefit (a build in deferred, prepaid subscription revenue) rather than the core business getting cheaper to run, and cost of revenue actually rose sharply this year, from about 41% to 53% of sales, because of a $17.4 million charge to write off components made obsolete by a hardware transition -- a type of charge that has recurred periodically as the company refreshes its device generation. The first half of 2026 showed continued strong revenue growth and a narrower net loss, but operating cash flow was weaker than the same period a year earlier, largely reflecting inventory built up to support that growth. Three cases: Bear case: the pace of loss-narrowing seen in 2023-2024 -- much of which came from one-time cost cuts the company itself says are not representative of ongoing operations -- slows well before the business reaches break-even, and losses stabilize at a modest ongoing rate rather than disappearing. Discounted to today, this outweighs the cash on the balance sheet, and the shares would be worth less than nothing: roughly $(1.30) per share. Base case: revenue growth moderates gradually from its recent pace, and the shift toward higher-margin software and licensing revenue -- a trend already visible over the last two years -- continues to narrow losses, crossing to break-even around the middle of the next decade and reaching a modestly profitable steady state by year ten. Combined with today's cash cushion, this is worth roughly $0.97 per share. Bull case: the Embedded licensing program scales faster and losses narrow more quickly, reaching break-even in three to four years and a healthier profit margin by year ten, worth roughly $2.91 per share. For the bear case to be too pessimistic would require steady, structural improvement in the underlying cost base (not just one-off cuts) as the software and licensing mix grows. For the bull case to play out, the Embedded partnerships -- still mostly early-stage co-development deals rather than proven, at-scale royalty streams -- need to convert into meaningful, repeatable revenue; that has not yet been demonstrated. Given how much of the outcome depends on execution that has not yet been proven at scale, and that the downside case is a genuine loss of value rather than just a lower return, the margin-of-safety price sits at a large discount to the base-case estimate; a value closer to that estimate depends on direct evidence -- a couple of quarters where underlying owner earnings resume improving on a like-for-like basis, after this year's inventory charge and the softer first-half 2026 cash flow.