Amazing. Okay, thank you. Welcome, everybody. It's Patrick Wood, obviously, on the UBS MedTech team. Appreciate everybody coming here, and excited to have Joe DeVivo here from Butterfly as CEO. Thank you so much for joining us.
Thank you very much for having us.
Just before we came on, we were having this discussion. A lot has changed at Butterfly in the last five years or so, certainly since I've been looking at it. What might be helpful, I think, to level set people is just to give a sense of very high level, big picture, the last few years, what's changed moving from just point-of-care ultrasound, which is still an amazing market, to the broader opportunity set.
Thank you for that. Really appreciate it. We're on this kind of lifelong journey to help democratize imaging, and our whole vision for Butterfly has been, how do we use technology to reduce the cost, improve access to imaging for people around the world? How do we create the opportunity where every doctor and every nurse can have an imaging device just like they have a stethoscope? We're the stethoscope salespeople 40 or 50 years ago. They used to literally, doctors used to put their ears on the chest, and then they had to be convinced by someone that it made sense to have a stethoscope to be able to hear it properly, and then what were the sounds and what did that mean? Of course, we've learned that imaging inside the body is extremely valuable.
But to be able to have the capability of imaging where the patient is, we are seeing just every day lives being saved because things are found early. We are in an amazing evolution. I think we are seeing it is becoming more standard of care in emergency medicine. It is becoming standard of care in the intensive care unit. As the market embraces the concept of physicians and nurses having the ability to image where the patient is, at the same time, the technology is evolving to make it easier and easier. The image quality is improving because our semiconductors are improving exponentially with Moore's Law. We are now having new AI tools come out that make it easier for novices in ultrasound to be able to at least acquire or capture the image.
A lot of things are converging, and now when we look at the semiconductor technology and its improvements, we look at the ability of doing these large language models and creating automation with AI, and then the compute convergence, it is kind of putting us at the center today of imaging because we are liberating our semiconductor for other companies to use. That is really what has changed. We maintain our focus. We are training doctors every day. We are going in with our salespeople into new facilities, putting in software and devices. We are educating through digital activation. We are on the journey of bringing point-of-care ultrasound the standard of care, but we have also now taken a few other paths along the way to diversify ourselves.
Definitely want to get to those. Maybe starting with, I have a soft spot for radiology, but starting with the point-of-care ultrasound.
You touched on it earlier, the unique thing in some ways about ultrasound is the technical challenge for the operator, right? As a ultrasonographer, it can be a difficult thing.
That's right.
These are emergency first responders and things like that. Could you maybe, for people listening in the room, walk them through some of the initial challenges, what that meant for the initial adoption of point-of-care ultrasound, and then you hinted at it with some AI solutions, things about how you've managed to upskill those first responders to being able to use effectively, clinically, a POCUS solution.
Ultrasound's hard. If anyone's looked at an ultrasound image, it's not intuitive. Black and white, grainy, and it's difficult because first of all, you have to have the ability of knowing, mastering human anatomy. So when you place the ultrasound device on the body, you have to actually find the organ that you wish to find. Then when you get there, you need to know what it is you're doing, what it is you're looking at. So those challenges are very hard. There's only 82,000 ultrasonographers in the United States. So if you think about what it takes in order to go get an image, you have to be inpatient, you have to get a patient, you have to make an appointment and go to the hospital. Whereas, doctors are now, what we're doing is we're starting to train them up a single use case.
Like, look, let's just do a pulmonary scan. If you just do this pulmonary scan and you're looking for pneumonia, pulmonary wetness, then you're starting to reduce the significance of the learning curve, and then they will learn one use case after another. But truly, to expand our market and to get it into everyone else's hands, we can't wait for the evolution of everyone learning. Today, about 80% of the medical schools today teach their kids on Butterfly. So they have them either in the lab or now medical schools, just a month ago, I was on stage at a graduation for the Edward Via College of Osteopathic Medicine, and where the dean of the school, after they received their white coat coming into their education, the dean of the school handed them a stethoscope, and I handed them a Butterfly.
160 kids starting off their education with a Butterfly device. That is happening more and more and more commonly now, the stethoscope and the device. All of these kids who are going to come out in 10 or 20 years are going to change the course of medicine. They are going to make the standard of care. Everyone is going to be required to have ultrasound skills, just like they do a stethoscope. We are a public company. You care about every 90 days worth of performance. You are not going to turn around and give me credit for what is going to happen in 20 years. We have to do things to compress the adoption.
The biggest thing to do to compress the adoption are there are many companies, 50 to 100 different independent companies who are developing AI tools that either help the acquisition of an image or the interpretation of an image. You can imagine now, as you said, in emergency services, when they are in an ambulance and they are triaging a patient on the way who is in critical condition on the way to the hospital, you can actually do a cardiac echocardiogram. You can do a cardiac echo. You could do a pulmonary congestion scan. You could do an abdominal scan. You can do these with AI tools that walk people through, place the probe here.
When you place the probe where the AI tells you to place it, the AI tool actually looks at the image, and then based upon the quality image, will tell you, "Hey, move the probe to the left, move the probe to the right, push down, twist it." When you make it easier to acquire an image, then they will send the imaging to the emergency room before the ambulance gets there. When the patient now is being brought out of the ambulance, instead of being waited to be triaged, they go directly to care. There are thousands of use cases. We are working now where we have actually created a Home business where we are helping at-risk providers with patients in nursing homes, because these patients over time are not titrating or they are not getting a more frequent diagnosis.
If their disease is going the wrong way, they basically will crash and then have to go to the hospital for care. We are teaching nurses who are not classically trained in ultrasound how to acquire a pulmonary congestion image. We send it over to a remote cardiologist who will read that image and then send them a diagnosis back, so their medication changes and they are able to stay in site. We are actually now actively doing it. We are seeing all the scans every day from our partner, and it is marvelous.
The whole concept is power every doctor, every nurse, allow them to have tools that make the image acquisition interpretation easy, be able to treat them or diagnose them at least in where they are in their home or in a clinic or in a remote area, and not have to travel and wait for one of the 82,000 ultrasonographers. We are opening the world up, and it is just very exciting.
In that Home and community care example, you are also lowering hospital-acquired infections for the most vulnerable population. You just keep them out.
100%. We are keeping them out. Also, while we are doing this pilot for pulmonary congestion, we are actually starting to see, even in the first few weeks, this tremendous desire to be able to scan other parts of the body. You mentioned infection. We had just one of the first 35 that were scanned at this nursing home this month as we started up our first commercial entry, is they were worried about someone's bladder, and so they did a bladder scan, which was not on our protocol, and found that the Foley catheter was misplaced, and this patient would have been septic a day or two if they had not found this.
It is just the power of giving people the ability to see and to do it, not when they suspect something and then make an appointment to go do something in a higher cost of care, but have the ability to make immediate decisions is going to make a great impact. That is how, one of the ways you bend the cost curve in healthcare.
Could you take that, because you mentioned obviously Home and community care business. Could you give people a little bit of background of the genesis of how that came about? You explained a bit about what you are doing there, but how do you grow that? Is that about interfacing with more and more centers, word of mouth? How does that grow, that business?
Today, we are working with a nationwide at-risk provider who has a very high population of patients in skilled nursing facilities where they are completely responsible for the cost of their care. Every readmission, let us call a readmission or admission a $10,000 incident. Probably a little bit higher depending upon how you do it. But every time we avoid that incident, they save $10,000, and that goes right to the bottom line. The reason why we built this community care business is we had seen other providers, other large providers who were doing home care. In the past, they would buy probes, and we would find that those probes wound up in the closet, because it is very difficult to run a program.
It is very difficult to train all these people how to do it, very difficult to have the software to manage the data, manage the devices, and then manage the clinical workflow. We said to ourselves, in that whole adoption thing, you are not going to wait for me for 20 years for everyone to figure it out, so I have to do things early to accelerate the experience, accelerate the adoption. So we said, "Okay, we are going to eat our own dog food. We are not going to sell you probes and hope that you can build your own program. We are going to take our own capabilities, we are going to train all your nurses, and then we are going to deploy the technology at no cost to you. We are going to integrate it into your systems.
We are going to develop the clinical workflow and pathway, and then you just pay us per patient, and you pay us per scan." We now have a program where we have a nice revenue stream that is starting for the company, but more importantly, this is an effort of us teaching the marketplace that this can be a significant reduction in care and a significant reduction in cost of care, and a significant amplifier. Now, we are meeting with other very large payer providers, and we think this is going to be a big business for Butterfly.
Would it be fair to say, as per that example, the system as a whole is generally tending to move towards capitated care as a concept, and anything that is upstream in the diagnostic side, like you guys, is a way to take cost out downstream? The better you know, the more you can clinically differentiate and take cost out. Is that fair as a summary?
Yeah, I think, and one of the things that is so exciting about now licensing out our chip is that we are working with so many other companies, but they all have the same objective, which is, first of all, how do we intervene earlier? So, how do we create diagnosis earlier so people understand progression of their health and how does that work towards their health journey? And also, not having to rely in facility, which is the high-cost environment. How do we provide care or diagnosis? We sell to ambulance companies, we are in cruise lines, we are in all these different places where people are, even helicopter transport where they will literally do a lung ultrasound or a cardiac ultrasound while they are transporting in a helicopter.
But in all of these locations, how do we push out the capability of getting an image where these patients are so they can get care faster? And it is all amplified by a low-cost device, a low-cost device that can image the entire body, and image the entire body with the tools around AI that can help acquire the image or even diagnose the image rapidly at a low cost. These are all accelerators, they are all amplifiers to being one of the It is all about pushing care outside of the facility, getting to the patients earlier, obviating the need for disease progression, and then reduction of costs or increasing costs, et cetera, so.
One thing that struck, or at least jumped out to me, in Q3 was the number of large enterprise deals that you guys had signed recently. Could you give us a sense of the genesis of that, and do those larger deals look different from some of the smaller you are doing? Is there more Compass involved in them? How are they structured?
Well, we have two different aggregate channels. We either sell directly to a doctor, and when we sell to a doctor, they have a device and they have our subscription to our software and our cloud. When a doctor practices in a hospital, the device needs to sync into the EMR. You need to be able to capture the record, you need to be able to capture the image in the DICOM, and you also need to bill for it, so push it in the revenue cycle management. We have had many hospitals where they would call us and a bunch of their doctors own Butterflies, and they would be upset with us because they do not port into their system.
They are like, "Why are you bringing these devices in that is not our protocol?" So, what we did is we built a middleware called Compass that allows all the data from the devices to port into the EMR, the DICOM, and revenue cycle management. We have been selling enterprise for a while now. It has been about three years. It seems like the marketplace has actually woken up, because as doctors want to do point-of-care ultrasound, historically only 30% of those images ever make it into their systems. So they are, A, non-compliant, because there is patient data that is being acquired in their facilities that are not getting into the patient record. So that is really the biggest motivator. Second, of course, 70% of those point-of-care ultrasound images are not only not being documented, but they are not being paid for.
So, the whole concept of our enterprise is to shore up not only the Butterfly usage in the hospitals, but our enterprise middleware works with all of our competitors' devices. We built a platform to solve the problem for our customer. Now, it is a growing awareness across health systems that they need an enterprise middleware to capture all this data. Historically, we would go and talk to the C-suite, and the Chief Medical Officer would not even know what point-of-care ultrasound meant. Now, though, IT, the CMOs, I think they are more aware to it. So we are building out our software team. We are selling software, we are doing the implementations.
On top of that, of course, it works best with our device, so we then bundle our devices in it, and then we pull it through with solving the point-of-care ultrasound data problem for hospitals, gives us a pole position on the hardware. That is our whole strategy.
I was going to say, I think I had underappreciated the degree of software integration from your side. It's almost like the network in your name.
It's a lot. It is a lot. It's hard to get into the workflow of IT, because they're doing so many projects. To get this even be enough of a priority for them to work on is a big deal. It's a whole other selling process. What's really fun about Butterfly is we have this ability to build our digital brand and engage consumers directly, and customers and doctors and nurses directly. We're building a brand, then what happens is that creates this local groundswell. We had a call from one facility who was very upset at us. They were like, "You guys, is this a Trojan horse? Why are all these Butterflies in our hospital?" We looked, and in the one city that this one institution practices, 140 of their doctors had bought Butterflies.
We told them, "Look, you have an enterprise program, you just don't know it yet." That's turned into now a nationwide deployment of our software, and this one health system now is over 800 devices. I haven't seen an order bigger than 20, so they're just buying all the time as this is building across. I do think point-of-care ultrasound is coming of age.
How should we think about some of those, the impact of those larger enterprise deals flowing through into sales? Is this something that is structured over five, 10 years? You go to gateways, how does that work?
It is. Historically, we would sell to a department would put it in, and then we would go to the second department, the third department. Now, the deals are starting to look multi-departmental and larger. When we sign a deal, it is normally either a three or five-year software deal. Our TCV for a lot of these deals are six, seven figures now, and it is really nice to see the size of these deals. From a revenue standpoint, we have to go through the implementation, and then once we get started, each month is one-twelfth of that revenue. It takes time. It is not like you get this capital purchase and it just hits, but we are building up this annuity now of larger software deals that should prove very good for our revenue.
I would love to pivot to embedded. This might be the single largest change that I have somewhat been sleeping on, but it is kind of interesting. Maybe again, just to level set everybody, give a little bit of background. I know you touched at the start, but what is happening there with your platform, sort of essentially being licensed out in a way?
With our semiconductor, the legacy ultrasound industry uses crystals, and crystals have to be cut to a certain geometry to be able to deliver a sound array. Our semiconductor is infinitely programmable. Like a TV screen, you have a bunch of pixels, see an image. You do not have a TV just for a country western. You can see anything on the TV, right? That is what our chip is like. We can program our chip to be able to deliver sound in any manner, any function. When people are looking to develop new technology, to have to go and solid state it every time you want to make a modification is burdensome. We have learned, we have actually had inbounds.
We had a very large consumer product company four years ago that does $2 billion in fish finders come to us and say, "Hey, can we use your chip for fish finding?" Our brilliant scientists at Butterfly had said, "No, we are in the point-of-care ultrasound business." But we learned with having almost 700 patents now, we have learned that the utility of this chip, because sometimes companies solve problems for themselves that everyone else has. Like Amazon Web Services. They never in their business plan to be a web services company, but they needed global cloud compute to do online retailing. So, they built it for themselves, had excess capacity, outsold that capacity, and now it is the biggest profit center for Amazon. NVIDIA, from 1993 to 2005, was a video game company. You will not find anywhere in their business plan that talks about AI, talks about data centers.
It's not there. They launched a development software because they had enough people saying, "Hey, these chips are really powerful, can we use them for other purposes?" They created a development software called CUDA that allowed third parties to come in and change the settings. What we've done is taken a page out of that book. What we realized is that 700 patents, $300 million for our first chip, a relationship with TSMC that can build it at scale, that we have something very special. There's many people who are developing applications for ultrasound in robotics, in orthopedics, in brain-computer interfaces, in whole body imaging, that would, instead of creating a platform based on an analog crystal, if they create a platform based upon a semiconductor, it's infinitely programmable.
Meaning that you can get the sound, you can view it, you can run it through your models, and then if you want to modify it, you modify it. You have this closed loop of rapid improvement. Then if you have software updates, and this happens for Butterfly all the time. It's like your iPhone, when you have a new operating system that comes out, or new updates to your software every couple months, they push it, and all of a sudden, all the devices in the world are loaded up with that software. For us, we have our people constantly improving our presets. We'll push updates out to all the Butterflies around the world. Next thing you know, they're imaging better because we've done something different, because we can change the settings to it. It's infinitely programmable.
We have over 12 different partners, 12 or 14, I forget, different partners today that are all licensing our chip. They are all licensing our CUDA-like software. I guess we got to give that a brand name. They're licensing our software, they're buying our chips, and some partners are licensing our IP to be able to build their IP on top of ours, so they can stand on the shoulders, not have to reinvent it. Even we have one partner who signed an exclusive deal where they're the only ones who can do whole body ultrasound with our chip.
Those deals you're signing, they scale with the volume of the customer, so you're kind of in it together in a win-win?
Yeah. If you go back and look, here we talk about point-of-care ultrasound, it's a big lift. We have our direct sales force, we have all of our marketing, all of our education tool, all the development to keep up with it. If we now want to go into whole body imaging, you got to replicate all that cost. If we want to go into HIFU, you got to replicate all that cost. If you want to go into brain-computer interfaces, not only you have to replicate all that cost, but now you're doing clinical trials and all this other stuff. If we create a licensing arrangement with a partner who will embed our chip into their product forever, then we have a forever hardware revenue stream. If they use our IP, we have a forever royalty stream.
Of course, if they're using our software, we have that forever. So now, for the investor, we create a lot of leverage because we can participate in a lot of markets without risk and without our having to develop the cost, but then participate in the success of it. So, it's turned into a very big multiplier for us, and we've initially had relationships with earlier stage companies that were very visionary in their ideas, and we're working with them. We're now starting to talk to some companies that already have businesses, and we might be able to move our chip into their existing products. If we sign those types of deals, the commercial revenue will be larger and will happen quicker.
I think over the next couple years, the revenue that we bring into the company will actually start to dwarf our point-of-care ultrasound business. Although that'll continue to grow, I just think our opportunity is semiconductor.
Things like HIFU, I think people have no idea how big that could be eventually. If I think of all those agreements you have, you obviously mentioned those, one on the consumer side, BCI, HIFU, obviously on the healthcare side. Going forward, looking at the potential applications, is there more in industrial or consumer or healthcare? Do you see a preponderance of the kind of partnership deals you're seeing today, and is that going to look different in the future?
Well, everything has been healthcare so far, just because of the proximity and the awareness that we've put into the healthcare part of the business. But, we are close to signing our first non-healthcare deal, and it has to do with material stress. So, you have airplane wings that are under stress, you have bridges that are under stress, and you can actually place, and they do place ultrasound devices, that can see inside the composition of the material, see the material change under stress, and can back the stress off before you have a permanent failure of that material. So, we are in hopefully final discussions with one partner. But then we've also done our market research work, and there are opportunities in underwater work, in underwater drones for oil discovery.
There's opportunities in the military for security, and I think there might be some very big military opportunities in the future. Then there are opportunities in consumer products. So, I think we've just brought on some salespeople who are only selling this, and they're being paid to go and discover. All the partners we have so far have all come to us. We haven't picked up the phone and called someone and said, "Hey, do you want to do this?" They've approached us saying, "Could we have availability of your tech?" So we're now going to be in the next phase over the next year where we're going to start having salespeople call people cold and start creating opportunities. And I think all these markets are You can do it yourself, but I've done a GenAI search, what are all the possible uses for ultrasound?
And that response comes back, and it's like, look at all these industries that use ultrasound today. And all we're doing is we are the digital to the photography's film. So if anyone's using ultrasound today in their products, we go and say, "Well, why don't you do a digital part of it? And here's all the benefits." And so, it's early on, but I think the opportunity is large for us.
It's a potentially pretty awesome mousetrap. What protects you from somebody trying to rip you off and steal the idea?
Yeah, a lot of people ask that, and especially, I think there's, especially those who are very experienced in the semiconductor industry, is that you feel you have this lead, and then someone comes in and copies this and that. I don't think we'll be without competition, and I think as this market gets more and more valuable, I think there will be other entrants. But we will be launching next quarter our fourth-generation semiconductor. That fourth-generation semiconductor will actually, for the first time, have a better image quality than a normal handheld piezo device.
You start asking yourself, well, why would you buy. If I have a digital camera that can produce a better image than film, but now you have to develop that film, you have to go. Why, if we have a better digital image now in ultrasound, why would you use that legacy product? So, I think probably our biggest protection is our evolution. I mean, we've been doing this now for 15 years, and we're on our fourth generation. Someone's first generation will not be as good as our fourth. I think as long as we keep developing, we have the best partner in the world, TSMC. I think our lead. They're going to have to figure out how to do it in a manner that we haven't done it. I'm sure people will. I'm sure people will innovate.
But, I think we have a very big head start, and we're just going to keep our foot down hard on the accelerator.
You guys have pretty good margin structure on the gross margin side, and as this flows through, presumably the cash generation's pretty meaningful. What are you going to do with all the money?
Oh, just continue to grow. Well, we've been using investment capital for many years, so it would be nice to return that capital to shareholders, right? There is so much for us to invest in. We actually just put a shelf in place just to have the optionality in the event that we needed cash. I do not think it is inconceivable that we would continue our organic development in new markets. I also do not think it is inconceivable that, with a better balance sheet, M&A, we can continue to find things that help accelerate our ability to grow on our strategy. So, it is amazing to see, because just three years ago, we were burning about $160 million a year and had less than a year's worth of cash. Now we have gotten our burn down. We have a very strong balance sheet. Our revenues are wonderfully growing.
As you get to each summit, you kind of pause and look around and say, "Okay, what do I do next?" The opportunities are bright.
Yeah. It is always a good question when you are getting that question.
Look, three years ago, it was not very comfortable, let me tell you.
Yeah. No, I totally understand that. If you think of adjacent technologies that could make sense, would it be the kind of thing that first responders, they are already there, they are getting used to Butterfly, other diagnostic-based equipment that would be used in a similar manner? Is it a core point-based tool process or like a technology or any and all of the above?
Anything that helps us service our customers better, anything that helps us accelerate our flywheel. I think on the POCUS side, there are a lot of different digital tools that are used in conjunction for certain use cases. They also all would like to port their data into a single place that can capture it. We definitely see some synergies on that part. We see synergies on AI tools that we might want to not just have a part of our ecosystem, but maybe have inside of our system. There are things that can help us accelerate. It is a build versus buy decision. You can do things organically or if someone has really solved the problem. We have today about 33 Butterfly Garden partners.
So 33 companies are building AI tools that will then link into the Butterfly ecosystem and have availability to the We have sold over 200,000 devices now, have probably 80,000 active users. Anyone who has software would love to have access to the ability of selling software to those people who are already ready to go. One of the beautiful things about the whole Butterfly Garden is if we see something that is really extraordinary, then we could always insource it, so.
You do a chunk of these meetings, you do all your earnings calls, you see the sell side note and you meet with the buy side. From your perspective, what is something that gets a lot of external focus that does not internally or vice versa? What is a big deal internally that you find it is surprising you do not get asked about as much? Or if you like, another way is like, what do you think people may misperceive or underestimate about the business?
That's a good question. I do think we've made a lot of progress on educating investors on the semiconductor part of the business and the opportunity that's arising. I think investors, for each of these, there's this like challenge between short-term and long-term revenue. Each partnership we sign is a whole new TAM that we have access to. Everyone's just focused on, "Well, what does that mean for revenue over the next 90 days?" I don't think people really realize how big. It's difficult because we're in such a hard spot because our partners that we sign and embedded don't want us to tell the world about their business models. They're stealth. When Midjourney, one of our partners found out that we had to file an 8-K, they were mortified because they didn't want the world to know that they were doing this yet.
We're caught between signing up partners that we know are going to be massive opportunities with the partners not wanting us to tell anybody. I think when you truly, I told this story of a friend of mine who runs a biotech. They're doing $5 million in revenue and they have a $2 billion market cap. It's one product.
Right.
They look at it as what can be. I have 10 of those, and people don't fully understand how big these can be. That's a great place to be. I'm fine with it. I'll deal with the quarterly, "How many deals are you going to sign this quarter?" As these partners start rolling out commercially and people can see the revenue, it will still be a show-me story at that point in time. I don't think people truly understand how the revenue can layer and how much opportunity with each new deal we sign.
I like that. Awesome. Joe, thank you so much.
All right, thank you.
Appreciate it.
Appreciate you very much. Thanks for having us here.