The Ultrasound Machine Wants to Become a Stethoscope

A clinician presses a probe against a patient's chest. On the phone in her other hand, a gray fan of anatomy comes alive. The machine looks almost too slight for the job: a white object weighing 300 grams, closer to an electric toothbrush than the wheeled ultrasound systems that once announced their arrival with cables, knobs and a pilgrimage from another floor.

This is Exo Iris, the flagship device from Santa Clara-based Exo. It can look deep into the abdomen, trace blood flow, guide a needle and examine the heart without asking its user to change probes. The hardware contains 4,096 individually controlled silicon-based transducers. The companion software can recognize when a useful view has appeared, perform specified measurements on the device and move the exam into the documentation machinery of a hospital.

That last step is the part casual demonstrations tend to skip. A picture at the bedside is not yet a durable clinical act. Someone has to attach the right patient, record the indication, review image quality, store the study, send it to the health record and, where appropriate, capture a charge. Exo's larger proposition is that portable ultrasound will spread only when all of those chores become as portable as the probe.

A pocket window with three views

Traditional ultrasound probes are shaped for particular jobs. A linear probe is useful near the surface; a curved one sees a wider abdominal field; a phased array fits between ribs to image the heart. Iris combines those imaging geometries in one device. Its specifications list depths up to 30 centimeters, fields of view as wide as 150 degrees, Doppler modes for blood flow and 22 exam types and procedures.

The core engineering choice is Exo Silicon, the company's piezoelectric micromachined ultrasonic transducer architecture. Conventional systems rely on carefully assembled piezoelectric material. Exo's approach brings ultrasound elements onto semiconductor manufacturing methods, pursuing the repeatability and economics that made cameras and motion sensors cheap enough to disappear into phones.

The analogy has limits. A medical imager is regulated, and its output still requires a qualified, trained professional. Iris received an initial FDA 510(k) decision in 2021 and launched commercially in September 2023. Today the standalone device is sold without a required login or subscription; Exo's site listed it at $5,499 when this profile was prepared. Larger Connect and Enterprise configurations add software, integration and service through direct sales.

“The scan is the visible moment. Adoption is everything that happens before and after it.”The product logic behind Exo's three-layer stack

The machine that watches the picture

Ultrasound is unusually dependent on the hand holding the probe. Tilt a few degrees, press too hard or stop at the wrong frame, and the organ can disappear into visual weather. That makes training an obstacle when the intended user is an emergency physician, nurse or field clinician rather than a full-time sonographer.

SweepAI is Exo's answer. As a caregiver moves across an organ, the software evaluates image quality, flags useful frames and, for cleared applications, produces measurements or indicators. Commercialized functions run on Iris rather than waiting for a cloud round trip. That matters in an ambulance, a rural clinic, a disaster response or simply a hospital corner with unreliable Wi-Fi.

Exo said in June 2025 that Iris carried 14 FDA-cleared AI indicators after adding detection and localization for pleural effusion and consolidation or atelectasis. Earlier additions covered functions involving the heart, lungs, inferior vena cava and bladder. Its Global Longitudinal Strain analysis, cleared in 2025, is designed to help evaluate heart muscle function at the bedside.

There is a thoughtful educational detail inside Exo U: short, in-app scan lessons, some under 30 seconds, demonstrate probe placement and standard views. The format borrows the rhythm of social video, but the timing is clinical. The lesson appears when a learner is holding the instrument, not months after a lecture.

The unglamorous moat

Exo Works begins where the dramatic gray image ends. Its tiers can create worksheets, connect to DICOM-enabled scanners, send studies to PACS or a vendor-neutral archive, manage review and quality assurance, track proficiency and connect patient encounters to EHR and billing systems. An individual can buy the probe. A department buys a working program.

That split explains the business. Hardware generates a one-time sale. Exo's terms describe recurring subscription fees for Exo Works, while installation and integration can add service revenue. Connect is aimed at teams that need collaboration, QA, credentialing and fleet management. Enterprise adds multi-facility controls, private-cloud options, order and encounter workflows, EHR integration and billing. The physical device is the invitation; workflow is the long conversation.

The evidence available publicly is mostly Exo's own case-study material, but it shows the problem the company wants buyers to notice. At Wellstar MCG Health, Exo reported that reviewed POCUS exams rose from 10 percent to 100 percent, revenue capture increased 47 percent and the organization broke even on the implementation in three months. At Geisinger, 86 ultrasound machines were connected across six institutes and ten campuses; one department reported that documentation time had been cut in half.

Those numbers do not prove every deployment will behave the same way. They do show why software is strategically important. A cheaper probe saves capital. A documented and reviewable exam changes operations.

A crowded pocket

Exo is not alone at the bedside. Butterfly Network sells a semiconductor-based whole-body probe. Philips has Lumify, GE HealthCare has Vscan Air, and Clarius and EchoNous compete with their own handheld systems and AI features. Hospitals can also keep using cart-based machines, call a sonographer or refer the patient to radiology. The alternative is often a workflow, not another gadget.

Exo's difference is the degree of vertical integration. It controls a distinctive transducer architecture, puts three probe geometries in one unit, runs assistance on the device and sells software for the institutional life of the exam. The combination is more defensible than any isolated feature, though it also asks Exo to be good at semiconductor engineering, regulated AI, consumer-grade industrial design, clinical education, enterprise software and health-system sales at once.

The company has financed that ambition heavily. It raised a $220 million Series C in 2021 led by RA Capital Management, with BlackRock, Sands Capital, Avidity Partners and Pura Vida Investments participating. Exo said that round brought total funding above $320 million. Additional data supplied for this profile records $24.5 million in debt financing in 2024. No reliable public valuation or revenue figure is available.

From one probe to other machines

The most consequential recent development may be less visible than a new AI button. In May 2025, Samsung Medison announced a strategic equity investment and collaboration with Exo to co-develop advanced ultrasound solutions. Omar Ishrak, the former Medtronic chief executive and former leader of GE Healthcare Systems, joined Exo's board as Chair of Strategy.

The language pointed beyond Iris. Exo described handheld and “soon-to-be OEM-integrated” systems, suggesting its silicon and embedded intelligence could appear inside a broader range of ultrasound products. For a company that spent a decade carrying one scanner into a pocket, the next market may be other manufacturers' machines.

Exo's customers include individual clinicians, emergency and critical-care teams, educators and health systems. Named users and partners include Geisinger, Wellstar MCG Health, West Virginia University Medical Center and Henry Ford Health. The use cases stretch from a trainee learning a cardiac view to a hospital auditing thousands of bedside exams, and from a rural clinic to military or disaster medicine where a connection cannot be assumed.

The company describes a fast, hands-on culture and recruits across nanotechnology, chips, consumer electronics, devices, AI, software and clinical work. That mixture fits the product. Iris must feel simple in a hand precisely because the organization behind it is complicated.

The habit is the market

Point-of-care ultrasound occupies a useful middle ground. It cannot replace every formal echocardiogram, radiology study or specialist interpretation, and Exo does not market Iris for untrained consumers. It can answer narrower questions sooner: Is there fluid where it should not be? Is the bladder full? Is the heart pumping poorly? Where is the vessel before the needle enters?

That speed can alter the order of care. The clinician does not have to move the patient to the image; the image comes to the patient. Yet the future depends less on whether a probe fits in a coat pocket than on whether clinicians trust it, institutions govern it, educators teach it and administrators can see the work it produces.

The stethoscope became ordinary because it was close at hand, inexpensive, teachable and embedded in routine. Exo has tackled those qualities as engineering and workflow problems. Its pocket window is open. The question now is how often medicine chooses to look through it.