All right, good morning. My name's Bill Plavonic. Welcome to the Canaccord Medical Device Technology and Services Conference. This is our 19th annual. Appreciate your attendance today. With us next, we're going to have CeriBell and from CeriBell, Scott Blumberg, who's CFO. This is one of our favorite stocks. We're in publication on that. And CeriBell is in the EEG monitoring space. The format of this presentation, it's actually going to be about a five to ten minute presentation, and then we'll sit down for a 15-20 minute fireside chat. With that, I'll have Scott come on up.
Thanks, Bill. Good morning, everyone. I'm Scott Blumberg. I'm the CFO of CeriBell. What we've done is developed a novel form factor for EEG, which is brainwave monitoring. Primarily serve the acute care setting, the emergency room and EEG, and we've made EEG monitoring very easy and allow for continuous monitoring. We currently serve the $2 billion US market for action in the acute care setting. We also have a pipeline targeted at other indications at the same call point, including delirium and stroke. We believe these indications expand our TAM significantly. Commercially, we just finished a $22.6 million quarter, so we're run rating at around $90 million. We're active in 615 hospitals in the US, which is about 10% of the US target market. Within those 615 hospitals, we have quite a bit of room to grow.
The business is growing rapidly, around 31% year over year growth in Q3, and we're operating at very high gross margins, 88%. The problem we're trying to solve are inherent limitations with what we call conventional EEG. Conventional EEG has been around for quite literally 100 years. I think 2024 was the 100 year anniversary of conventional EEG. It's a very good technology. It was built for a specific purpose. It was built for diagnosis of epilepsy in the outpatient setting. That technology has been applied, and we think misapplied, to our call point, which is the acute care setting in the ICU and ED, where the needs of the patient are dramatically different than in the outpatient setting. In the outpatient setting, precision diagnosis is very important. The functional area of the brain that's not performing correctly is very important. Speed matters a lot less.
In the acute care setting, every second matters. Patients in the acute care setting tend to have longer seizures that can result in what's called status epilepticus, which is a seizure lasting five minutes or more, and that can create pretty profound impact on patients. The longer a patient seizes, the higher the rate of death and disability, and the less effective the medication is. This technology was not built for that purpose, and it's got inherent limitations to prevent diagnosis and monitoring in line with the guidelines, which suggest that monitoring should occur within less than an hour in most cases. It's a very large piece of equipment. You have to wheel around the hospital, but more importantly, it requires manual application by specially trained EEG technicians, where they need to manually prep the skin and glue on dozens of electrodes.
It takes 30 to 45 minutes to set up. Even more profoundly than that, EEG technicians are in short supply in the US. The majority of hospitals do not have 24/7 EEG coverage. They'll typically staff their techs nine to five, nine to eight. There are significant delays in getting signal acquisition. We did a study called the Decide Study a few years back, where we took the top of the top academic centers with great EEG coverage, and the average time between ordering EEG and getting it applied was four hours. That's just signal acquisition. Once you have the signal, you need to interpret it. That requires, with conventional EEG, a specially trained neurologist. They're not always immediately available to read, and when they do read, they most often review at a point in time and do not continuously monitor.
Whereas patients need, and the guidelines suggest monitoring quickly in an hour or less, the reality is substantially different. It can take hours and hours to get an EEG, and if you happen to have a seizure during off hours, say on a Friday, it can take days. That's where we're changing. We've made EEG very simple. We can get the device set up in about five minutes. It allows for early detection, allows for action at the bedside, and then we continuously monitor the patient. How we do that is a combination of hardware and software. The hardware is shown on the screen here. We've got a single patient disposable headband. It's very, very easy to set up. You can be trained. Anybody in this room can be trained to apply it in less than an hour.
We typically, in most of our hospitals, have nurses putting it on, so you do not rely on that very weak EEG technician infrastructure. It plugs into a piece of capital, which is a recorder. It is about twice the size of a smartphone. All that data then gets sent to a portal through the cloud where the data can be interpreted by neurologists. Those three pieces of the device make signal acquisition very, very simple and streamless. That is not all. We have overlaid the whole system with what we call Clarity, which is the brand name for our seizure detection algorithm. That converts the very esoteric waveform of EEG that requires a trained neurologist read into something that can be interpreted at the bedside.
You get bedside alerts when there are long seizures consistent with status epilepticus, and it converts the EEG waveforms into a seizure burden trend line, which allows you to see the % of time the patient's been in seizure as well as how the patient's trending over time. The product is resonating in the market. As I mentioned, we're in over 600 hospitals and did $22.6 million. Inherent in our business model is a great deal of predictability, as you can see from the consistency of the chart here. Around a quarter of our revenue comes from SaaS subscriptions. Around three quarters comes from the raise-to-raise related model driven by our headband. Usage patterns are very predictable over time, and this high degree of projectability is one of the things that gave us the confidence to become a public company about a year ago.
I'll briefly touch on the pipeline. I focused on the first pillar of our pipeline here, which is our mission to make CeriBell the standard of care for seizure detection in the acute care setting. We're going beyond seizure, and we've already made tangible steps to do that. In the second pillar, we are focusing on other indications at the same call point using our platform technology. We have disclosed that two of those indications are delirium and stroke. These affect many of the same patients as seizure, but also additional patients. It has the benefit of synergistic impact in that by making EEG a vital sign where the equation flips from, "I'm looking for seizure," to, "I'm trying to figure out what's going on in the patient," it allows for better information to the physician and also brings more patients into the mix.
Over the long term, we're also looking at brain biomarkers and other sites of care outside the acute care setting. We've got a whole lot to tackle in the first two pillars and multi-billion dollar opportunity there. I think I'll leave it there and hand it to Bill.
All right. Great, Scott.
Please join me for a little chat here. Let's get that one out of the way. Just for the audience, I think it's key to point out, rarely do we see in the world of med devices where we have a guideline set up of how to treat patients. We have a treatment available. There really was a diagnostic solution, just wasn't right, right? You've come in and you've provided that diagnostic solution to get to what the patient needs so the physician can administer. That's really the key benefit of your technology, getting that answer faster.
That's accurate. The conventional EEG is a great, great technology for the outpatient setting. We just think that ours is much better for our call point.
Let's, as a public company, I have to start out with the financial questions. As CFO, it's probably the easiest for you. Yeah, good quarter. You did raise guidance a bit. Can you talk about on the new guidance, what gets you to the higher end of the range and just remind us about the process of onboarding accounts like in Q4 and kind of timing of different things?
Sure. Our current guidance is $87 million-$89 million. Our philosophy in guidance, and this is something as we, the many, many quarters we spent preparing for an IPO, that was taught to myself and our CEO, Jane Chao, more than anything, which is you got to put out numbers you can hit. So our guidance is always colored by that lens, and we make sure that we're conservative in that. Q4 is a bit of a funny quarter for us. It's one of our higher seasonality quarters. Q4 and Q1 tend to be higher for us because the ICU census is higher, and Q2 and Q3 tend to be lower. We're pretty late in the year. We've got a really projectable business model. The range appreciates that although we clearly believe in seasonality, we've seen it multiple years in a row, we still could benefit from more data.
The business on the short term doesn't have big surprises. The reality is that the difference in the range is just how much seasonality we see this quarter.
Gotcha. Okay. So you've guided on the gross margins to mid 80s. You've working through all this pre-tariff inventory in Q4, and then you have some higher tariff products. As we think about, and then you're shifting to Vietnam, just walk us through the tariff impact, the mitigation strategy. When do you get delivery of the Vietnam products? Like Q4, I think is coming down a bit because you've got some mix of pre-tariff and post-tariff things, and then Vietnam rolling in. Help us understand, and then cadence as we get into next year and kind of the Vietnamese product set comes in. Do we see big variations? What's that cadence going to look like?
Sure. We're already accepting commercial products from Vietnam. The reason it's not hitting our books yet is because we do first in, first out accounting. As you mentioned, this quarter, Q4, we transitioned to inventory acquired after April, after tariffs. We still had some tariffs from the 2018 tariffs on our books before that. We pulled no inventory in from China when tariffs spiked for a couple of months to 150% plus. Of course, they did settle around 55%. As we move into this quarter, as you mentioned, it's pre-tariff inventory mixed with that 55% tariff inventory. As you move towards the middle of next year, that then settles to what we see as kind of the long-term mix, of course, subject to changes. The facts change from the macro, which is a mix between China and Vietnam.
We've set up our manufacturing where we're manufacturing in both countries. Both countries, with changing the scheduling and number of lines, can service our entire need. We intend to pull in a mix of products both from Vietnam and China. We think there's advantages to supply chain diversity, but we'll, of course, skew the mix towards the more attractive geography, which currently is Vietnam.
I think as we think about it, the guidance is mid 80s, so we'll kind of be bouncing around that as we go forward.
That's right. We have opportunities to further mitigate the cost through cost reductions that we're working on. We're always working on that regardless of tariffs, and those could serve as potential upside.
All right. I'm going to flip over to commercial here. As I mentioned in the introduction was that we always have a couple of names that we're really focused on, our kind of favorite names of ours going in. You're on that list as one of the four that we're focused on. Part of the thesis is really the accelerated hiring post the IPO. That bolus of reps is going to begin accelerating growth at some point. They'll start contributing to the overall because it was a pretty massive expansion of the sales force. Can you speak about the cadence of that new hiring post the IPO? What are you seeing from these new classes in terms of productivity ramp relative to prior classes?
Yeah. We've expanded both sides of our sales org, for reference, the account acquisition team we refer to as territory managers, and the utilization team we refer to as clinical account managers. I think you're referring more to the territory managers where we've expanded the number of territories from mid 30s to around mid 50s. That took about a year post IPO. Typically, our sales cycle for a brand new rep and a brand new opportunity is about a year. The sales cycle itself is much less, but for a new rep, it takes a few months to train. Sales cycle is four to six months typically. After we get a purchase order, it takes a few months to launch. As you said, that investment really translates into impact starting in 2026.
Before we see the impact and the street sees the impact, and our disclosure on account base is live, active, using accounts with a subscription that they're paying for and active. We've got a number of precursors to that. The most compelling, of course, is a purchase order, which we get three or four months before we get the launch, as we call it. Before that, we track in our CRM system, of course, the various stages of acquisition. As we look at these new reps and try to build confidence that they're going to deliver as we expect, we, of course, track how they've compared to prior reps, and they're tracking quite well.
Excellent. Can you remind me in terms of the territory managers? I think the IPO was September, October last year.
October, yeah.
October. And then you got the proceeds in, it kind of takes a little while to make some decisions about hiring and all that. Were you able to get a lot of them in the door by the end of the fourth quarter, or was it really more of a 2025 when they started bringing them on?
It's been pretty consistent in terms of our hiring cycle.
Okay. Good. You talked about the pipeline of accounts with purchase orders already. Any way to quantify that? Is that pipeline above or below norms, and is that a reflection of that commercial effort?
We said on our last call that our pipeline is the most robust it's ever been. Beyond that, I can't quantify. We've certainly debated that internally, but it's not too far down the road where you'll see it in the numbers. I think a little patience there.
Okay. Like a finance person, I'm going to ask the same question 10 different ways. What's the usual timing between you mentioned the purchase order, you got the training, and then eventually you'll book revenue? Is that a pretty long window typically?
It's a distribution. The average is three to four months. Why does it take so long? It requires integrating with the hospital's Wi-Fi, which requires mobilizing their IT department as well as various agreements around how we handle patient health information. It requires training. And although the training is really simple, there's dozens of doctors and sometimes hundreds of nurses at these hospitals. It requires designing the workflows, making sure that they know how to use it, making sure that they know how to order it within the EHR. Lining up that many schedules is challenging. The flip side of the amount of effort we put into acquiring accounts is once it's on, it's on. We see usage happen without a great deal of intervention from our team. We have absolutely no case coverage, as I'm sure you've seen in prior companies.
We put quite a bit more effort on the front end, but then I believe that once we're in, it's much less effort to maintain. The vast majority of our commercial infrastructure is focused on growth, not maintenance.
All right. Good. On the third quarter call, you noted that you're only 30% penetrated into your 615 active accounts. How do you balance going deep versus new account adoption for the commercial organization?
We do both. The way we manage it internally is actually by bifurcating the sales force. When I said that there's two separate functions, they're truly separate. The territory manager runs independently from the first approach to the purchase order. From purchase order to those pre-launch activities we discussed to 90 days post-launch, they work together. They work together for two reasons. One is that's a really important time. Habits are hard to break. Good habits are hard to break. Bad habits are hard to break. It is very important to set good habits. Also during that time period, relationship transfer happens. Ninety days after launch, the territory manager completely hands off the account. They do not have ongoing maintenance responsibilities. The functions are, other than that launch period where they're working together, singularly focused on either acquisition or utilization.
As you can imagine, we're able to flex our investment philosophy based on where we're seeing the greatest ROI. We evaluate that on an ongoing basis.
One thing I really found interesting was that hospitals often annually set protocols. You're trying to get those protocols to reflect the EEG guidelines. How much of the account base has adopted or is truly following the guidelines? How do you influence this for 2026?
It really depends on the patient population. I think if you kind of go down the cascade of the types of patients we treat, and there are dozens of underlying conditions that can lead to seizure, the ones that are probably the most understood are patients who had prior seizure. And if you had prior seizure, you are very likely to have a seizure in the future. You see really good usage when someone is worried about seizure and they had a prior seizure. It is like clockwork. You move kind of to the next phase, you are talking about things like cardiac arrest where there is a guideline from the AHA suggesting that you should use EEG if a patient survives cardiac arrest and is comatose. Adoption there is improving, but still certainly not 100%.
You move down the cascade to something like sepsis where there's clear clinical evidence that there are risks of seizure and that people should be monitored. There are no specific guidelines for monitoring. In that area, as you can imagine, it's a little less formalized. We work on all those fronts. We have a huge advantage in that our data from the recorder, the clinician or the nurse puts in the patient's underlying conditions. We actually have real-time analytics on how our customers are using our device. We go back to them and explain to them, "Hey, we see you're using it a lot here. You're not using it a lot here. And by the way, here's the clinical evidence." We've got a very clear playbook, but it's tailored to the individual hospital and how they're using the product.
Gotcha. The VA, it sounds like the pilot was successful. How do we think about contribution opportunity here? Will the system, how quickly do you get into all 200 hospitals? Is it all 200 hospitals? Is it system-wide? Is it class med tech adoption? How does that adoption look?
The pilot was very successful. It was a small number of hospitals. With the successful pilot, the VA committed to expansion. It's not all 200 hospitals. It will be a number that's noticeable. From there, we're approaching it two ways. We both are looking bottom up and top down. Of course, we have bottom up support, but this last kind of bolus of accounts was really top down. We continue to drive top down. To some extent, that's dictated by the annual budgeting cycle of the VA. We're also looking bottom up. We've kind of had 1Z, 2Z acquisitions of other VAs. That's a little easier to do off-cycle because it's easier to find the money in a budget to support one hospital versus a lot more. I expect that we'll continue penetrating over this cohort will be over the next few quarters.
Over the next few years, hopefully we get closer to that 200.
Okay. Similarly, can you give us the update on discussions with large IDNs, purchasing groups? I mean, you're saving these people a lot of money. I would expect that there is a big drive in those types of groups.
Yeah. We're really excited about that. We think that's our biggest growth opportunity for next year. We've seen time and time again that our most successful reps embrace this strategy. That's more for a rep that's more akin to a small regional IDN, 5, 10, 15 hospitals. What they'll do is they will go to the hospital, get some buy-in at the ground level, then go up and build a plan to expand, whether it's let's turn on all eight at one time or let's do these two. In three months, we'll show you the data that we're telling you, and then we'll go further. Those are the reps who really accelerate.
As we move into next year, we've talked about what we believe to be our next phase of investment in the commercial org, which is regional IDN leadership to more systematize that and make it a little less ad hoc. With 615 hospitals, we've got our foot in the door at a significant number of the IDNs and hospital systems in this country. We think it's time to monetize that and really not just talk at the ground level, but talk to administration as well.
I'm going to switch over to competition. We got about three minutes left. One concern that investors have is that with AI becoming more accessible, you could be more susceptible to an upstart disrupting the marketplace. Just talk about why do you believe your algorithm is defensible?
First, I think AI starts with data, and we have the data. When you build something with AI or machine learning, you need a source of truth. You need both the data and you need what the data means. We have well over 200,000 EEGs in our database. I think that's a unique asset. We've also invested in the labeling required. In order to have the source of truth, we have multiple trained neurologists label the EEGs. Of course, they don't label all 200,000. In order to find the percentage that form the outliers to improve algorithm, you need to collect all the data. We've made that investment. We've got the data asset.
My understanding is further the big investment that we've seen in society around large language models is different than the investment we've made through a pretty large, certainly very large for a med device data science team, is different than the data science that leads to signal processing.
I'm going to look at the hardware, the devices themselves. It seems that much of the hardware we're seeing from competitors on the market are similar generally. Is a differentiation in your view in terms of the algorithm, or is it the form factor of the device, or how would you say for the lay person coming in?
It's everything. I think the algorithm is probably the most compelling and strongest. I mean, we've got orders of magnitude more data informing our algorithm. We've published multiple publications on our algorithm. Competitors have none that I'm aware of. The form factor does matter too. We've refined that over time. Ours is really simple. People are used to it. There is also the fact that we've got 100 clinical publications and we've got 600 hospitals using it. We believe that we've got advantages over the other systems, but even if they were equal, which we don't believe to be the case, would you put your career on the line to try something else when you've got this thing that's proven and being used elsewhere? We are absolutely winning in the market.
Okay. Since the IPO, your net losses have actually increased. What do you think or when do you think we could see an inflection point in terms of operating losses kind of peaking and starting to decrease? What will drive that, and when could we expect to see that?
Yeah. The flip side of the big commercial investment that we talked about earlier is, of course, that. With the delay in impact from revenue related to the sales cycle and time to train reps, you get the full cost before you get any of the impact. With an 88% gross margin and our entire sales efforts are focused on growth and not maintenance, we have a great deal of flexibility to calibrate between growth and profitability. We're still very much a growth-first company, but we always got our eye on profitability and remain committed to becoming profitable without additional capital.
Excellent. I think with that, our time's up.