Butterfly Network shrank ultrasound. The hard part came next.

An ultrasound scan has two lives. In the first, a clinician sees something useful: a bladder filling, a heart moving, a needle approaching its target. In the second, the image becomes a record that colleagues can review, a supervisor can check, and a hospital can account for. Butterfly Network made the first life portable. Much of its business now concerns the second.

Consider the University of Rochester Medical Center. Its 2024 case study described 862 Butterfly probes, with 73 non-Butterfly devices also connected to Compass, the company’s workflow software. The reported result was a 116% increase in point-of-care ultrasound charge capture. A seductive number, provided one reads the noun: charges. It does not mean profit doubled, or that every patient’s care improved by the same percentage.

The less photogenic achievement was organizational. Rochester established a governance committee, developed credentialing and quality-review arrangements, and rolled out the program in phases. Pocket ultrasound had acquired a place in the institution. The machine was small; the change surrounding it was considerable.

A chip with bedside manners

Jonathan Rothberg founded Butterfly in 2011. The original iQ reached the market in 2018. Seven years is a useful corrective to the apparent simplicity of the finished object: a probe, a cable, a phone. Medical hardware does not arrive merely because somebody has drawn a handsome rectangle.

Traditional ultrasound transducers commonly use piezoelectric crystals to send and receive sound. Butterfly built its approach around microscopic mechanical elements on a semiconductor chip. Its Ultrasound-on-Chip technology lets a single probe serve applications that traditionally call for different transducer types. The operator changes a preset rather than reaching for another probe.

That is the company’s distinctive engineering bet. A software-controlled semiconductor platform can support new imaging functions and successive hardware generations. It also gives Butterfly a technology to license. Its expertise therefore spans chip design, acoustic imaging, mobile software, and the clinical work needed to turn those components into a regulated device.

The iQ+ followed in 2020; iQ3 launched in the United States in February 2024. The third generation brought a new chip platform, improved imaging, and faster data transfer than its predecessor. It subsequently received the 2024 Prix Galien USA award for Best Medical Technology. Useful recognition, though a clinician still needs to judge a device against the examinations they actually perform.

The $3,899 question has several answers

Butterfly competes in handheld point-of-care ultrasound, often shortened to POCUS. GE HealthCare’s Vscan Air uses a wireless, dual-headed probe. Philips Lumify offers separate transducers for different applications. Butterfly’s proposition is a single semiconductor-based probe coupled with software, education, and institutional workflow tools. Portability is shared territory; the architecture and surrounding offer differ.

The company’s U.S. individual pricing lists iQ3 at $3,899 and iQ+ at $2,699, before software. Core software is $299 annually; Advanced is $420 annually. A $1,500 one-time Advanced option guarantees software access for five years. The probe price is the opening line of the budget, rather than the last.

Hospitals purchase a broader arrangement. Compass AI works with Butterfly and other ultrasound devices, connecting documentation, quality review, proficiency tracking, and fleet management. It supports integration with electronic health records and image archives. Enterprise prices require a sales discussion. Someone must also own implementation, provide compatible mobile devices, and arrange training.

There are other buyers. Medical schools can put probes into students’ hands early. Veterinarians have their own product line, now including iQ3 Vet. In 2022, Petco agreed to deploy iQ+ Vet across its network of nearly 200 full-service veterinary hospitals. Dogs, it turns out, also have a stake in semiconductor design.

The machine does not teach itself

In October 2023, chief executive Joseph DeVivo said customers pointed to “education as the biggest barrier to mass adoption.” That admission explains an important product decision. Butterfly expanded its education offer with expert-led Butterfly Certified training and ScanLab, an AI-guided practice application.

“education as the biggest barrier to mass adoption”

Joseph DeVivo / Butterfly education announcement, October 2023

ScanLab supplies image-quality feedback, anatomy labels, and guided practice. Butterfly Academy provides courses, while Certified offers virtual and in-person teaching developed with the Global Ultrasound Institute. These products address different stages of learning. ScanLab is explicitly educational, not a diagnostic tool. A practice app and a clinical AI feature should never be mistaken for one another.

The distinction becomes especially important away from well-equipped hospitals. Butterfly’s Gates-funded Kenya program deployed 500 iQ+ devices and trained 514 practitioners across 224 facilities in 2022. The $5 million grant supported a larger, two-country plan for 1,000 probes. South Africa’s phase launched in April 2024. The funded package combined equipment with instruction; mailing boxes alone would have been a rather incomplete intervention.

A newer clinical tool addresses a more specific task. In March 2026, Butterfly announced FDA clearance for its blind-sweep Gestational Age Tool, based on deep-learning models developed at the University of North Carolina at Chapel Hill. Guided sweeps support an estimate of pregnancy age. It is a defined application, not permission to treat every image as an automated diagnosis.

Silicon advances. Inventory sends a bill.

The business has had less elegant moments. Butterfly’s 2023 results described a reorganization and substantial cost reductions. Its subsequent annual report recorded $21.1 million of excess-inventory losses in 2023 and a further $17.4 million charge for excess and obsolete inventory in 2025. Technological advances and changes to the product portfolio were among the stated reasons for the later charge.

There is a practical tension here. A chip company wants its next generation to improve quickly. A hardware company must sell the generation already sitting in inventory. Butterfly has to be both. Its costly stock of devices is a reminder that medical technology has warehouses as well as patents.

Funding helped pay for the long development and commercial effort. Butterfly raised a $250 million Series D in 2018. Its 2021 Longview business combination delivered approximately $589 million before transaction fees. By the second quarter of 2026, revenue was $32.6 million, up 39% year over year, alongside a $12.9 million net loss. Adoption was growing; profitability remained unfinished business.

Someone else’s device, Butterfly’s chip

Butterfly also sells the possibility of building beyond its own probe. Its Embedded program licenses the ultrasound platform and supports co-development. A 2023 agreement with Mendaera brought the technology into development of an image-guided robotic system, with revenue sharing upon commercialization. Butterfly Garden provides a developer ecosystem for applications around the platform.

This adds a second route to market: partners can turn the underlying technology into products Butterfly does not have to invent and distribute alone. It also adds dependencies. Licensing agreements, clinical development, regulatory clearance, and commercial sales happen on different clocks. A promising partner project is not the same thing as a finished, revenue-producing device.

Copy the rollout, not just the shopping list

Rochester offers the most transferable lesson. Begin with a defined clinical use, appoint people to govern the program, train and credential operators, review image quality, and make documentation part of the workflow. Expand in phases. A hospital can borrow that sequence even if it buys a different brand of probe.

The conditions matter. A compatible phone or tablet, appropriate training, and reliable access to charging are part of the system. Butterfly says scanning can work offline after setup, but its app must reconnect to the internet at least every 30 days. Whole-body capability means versatility across applications; it does not make every specialist examination unnecessary.

Butterfly’s achievement is to move a useful piece of imaging closer to the patient. Its remaining task is wonderfully ordinary: help people use it well, keep the record, and make the economics work. The semiconductor deserves attention. So does the committee that decides where the scan goes next.