All right, well, good morning, everybody, and welcome to the Needham Growth Conference. This morning we are excited to have the CFO of SoundThinking, Alan Stewart, presenting, and we'll give it over, pass the mic here.
Thank you, Josh. Okay, everybody hear me all right? All right, thank you for being here. I really appreciate you taking the time, to hear a little bit more about the company, one of the companies that I think is one of the best in the world, but I am a little biased on that. Just a couple things. I will be, covering some forward-looking statements, talking about the future, talking about what we've done, what we are doing in terms of our, our solutions and our products, so just be aware of that. If you have any concern, please make sure you look at all our SEC filings. So SoundThinking, if I just talk a little bit about it, some of you would know that we used to be called ShotSpotter. So ShotSpotter, we went public in 2017 as a company named ShotSpotter. We changed our name to SoundThinking.
There was a reason for that, and the reason is we introduced new solutions based on acquisitions that we've done, but also some partnerships that we've had as well. What we have right now is called the SafetySmart Platform, and that basically combines all of the different solutions that we have into one platform. They're all still software solutions, so from— if you think about us from a revenue perspective, we are a SaaS company. About 90%-95% of our revenue is subscription or recurring revenue or SaaS-based. The balance, you know, the five about 5% is professional services, but we're definitely a software company. So just going into the actual SafetySmart Platform, I think it's important if I start covering sort of what is there. If you think about it, the first one, which is the ShotSpotter, that's what we went public with.
That is the gunshot detection solution. What that is, is that for those of you who aren't aware of us too much, we have sensors. Like in New York City, we have 90 sq mi , where we have about 25-30 sensors per sq mi , and these sensors are on light posts, they're on the top of buildings, they're on some residences, churches, buildings, things like that, and what we do is we listen for a pop, boom, or bang for a loud noise to determine if that is something like a firecracker, a backfire, a car, or an actual gunshot. We take those, run them through our algorithms, and then send it into an incident review center that also listens to it and looks at it from a screen to determine that it's actually a gunshot.
If it's actually a gunshot, we send out that alert to the police within about 45 seconds. We have to give it to them within one minute, and that happens right now all the time. We do thousands and thousands of those a year. We're in 170 cities right now. We're also in South Africa. We're in Uruguay. We started in Uruguay, t hey've already expanded. We're in the Bahamas. But again, 170 cities right now. That's something that is still growing, but as opposed to in 2017 when we went public, that was 100% of our revenue. It's now only about 65% of our revenue. That's because the other solutions that we either acquired or developed are the remaining 35% of our revenue as we expanded our platform. So it's still a, it's still the largest.
We add, you know, over 20 cities a year. Some of those are new cities. We might expand in 20 cities a year as well. So if you think about us adding another 100-120 miles every single year, that adds a significant amount of revenue and also a significant amount of cash, which we generate with our ShotSpotter solution. Our next product is ResourceRouter. This is one of the first acquisitions that we did. This is a patrol management solution, and why this is incredibly important is because if you think pre-COVID or pre-George Floyd, there were a lot more officers out in the police departments. Now some of them are down about 40%. Most of them are down at least like 10%, so they're trying to find ways to use the officers that they have in a more effective way.
That's what ResourceRouter does. Although we acquired this several years ago, it took us a couple years to get it dialed in to where we would say that it is something that most of the customers that are using it find it incredibly valuable. In fact, we've sold more of those in 2024 than we sold in the prior three years combined. So we finally hit the mark on our ResourceRouter. Very excited about that. It's one of our, although ShotSpotter uses a lot of AI algorithms to determine what the gunshots are. ResourceRouter is also an AI algorithm-based solution.
What we use there is all the past data that the police have on the crimes that have occurred in their cities, but then we add in things like weather, locations, are they close to schools, are they close to bars, when did the bars let out. All those things determine our built into the database to determine when the police should do what we're recommending they should do. Plate readers are a relatively new one. This is something that we just started in September of last year. We are a reseller for a company called Rekor. Rekor builds license plate readers, and license plate readers, for those of you who know, are used in many, many cities, but they're also used in cities that know they need to expand them.
They might have like a five, a seven, or a 10-year license plate reader in there now, so they're getting ready to replace those or expand them. They might have it in only 20% of their city, but they want it in 40%. That's why we decided to team with one of the best providers of license plate readers, Rekor, and now we're reselling them because they're no longer selling to municipalities. They're focused more on transportation. So this allows us to take the customers that we have with the other solutions and potentially upsell them to a license plate reader and then maybe tie that into a ShotSpotter solution.
So if you have a ShotSpotter alert going off, you might have a PlateRanger LPR in the actual street, so you can look at the ingress and egress areas and figure out who was going in and who was going out. So we also have done that for a long time with like video management systems where they would pan, tilt, and zoom the cameras. So we're just adding an additional thing in terms of the PlateRanger for LPRs. CaseBuilder is a solution that we actually acquired in 2020. What we acquired was a software solution that was an on-prem based, used here at NYPD. They've been using it for about a decade. It's sort of like their ERP-type system, so it has many different modules within it. It helps them manage cases.
For NYPD, it could be like, it could be parking, it could be personnel management, it could be, any number of different things. We took that on-premises-based software and built it into a cloud-based solution called CaseBuilder. So we are now selling that solution to not just police departments, but one of the largest ones we did near the end of 2023. We sold to a Department of Corrections. You all know who, where it is. It's here in Rikers Island. So we were sole sourced this solution for a six-year contract, about $18 million, so about $3 million a year. $2 million of that is subscription, a million of that is professional services because we're using that to help build new modules for things that they might need.
So in other words, instead of just doing personnel, they might need one for prisoners, they might need for people out on parole, they might need for their internal facility for their restaurants or whatever. These are the different modules that we're building. The good news about that is as we continue to have new customers, we're expanding the software capability and then use those modules for other ones that we're selling. In fact, we've already sold it to another prison. We also just sold it last year to a Department of Correction. I'm sorry, to a DOJ in California, and they're using it for insurance, Medicare fraud. So it has nothing to do with like crime other than fraud. Obviously stealing Medicare is a crime itself, but it's a completely different one than someone else is using. So there's other verticals that this can expand into.
It's one of the things I'm most excited about, was the CaseBuilder. CrimeTracer, we kind of call that like Google for cops, and it's Google for cops because it has information in there. Like it says, "Alan Stewart, he's aged, 40, been married for 36 years, has two kids." I'm not 40, obviously. He has two kids. The kids live here. He has a, a heart tattoo on his left shoulder, which I don't, but, I could. This is the kind of information that's included in the CrimeTracer, and that's information that you can get if you have CJIS certified, which is Criminal Justice Information Services, certification that allows you to get that information, which police do, FBI agents do, but the normal just someone out on the street does not. So we sell that. That's something that's really incredibly important, and also growing as well.
Lastly, SafePointe is probably the largest, new thing that we've done in terms of a TAM expander. It's a weapons detection platform, but it's covert. So it's like they don't look like this, but they have them called bollards, and we put them in front of buildings where you're walking through them, and I'll show you a picture of them in a little bit where we're determining, again, using AI algorithms based on the magnetic movement of this metal that's going through there, what kind, whether someone is carrying a gun versus a laptop, what kind of a gun they're carrying, and then also like where the location of that, of that gun is. And then we notify the, the security force there that that's, what's going on. There's another company that's, called Evolv that's doing this right now.
They're the leader in the market right now, but this market is between $20 and 30 billion. So, we're cheering them on and cheering anyone else on that's doing this because it's building the market because if someone, let's give you an example. We focus mainly on hospitals and casinos. If you're doing a hospital and you see the next hospital next to you doing it and you're not, you're like, "Well, I need to then do something." So it just expands the market. So, it's something that only gave us about $1 million in revenue last year. I'm sorry, in 2023. 2024 will be a multiple of that, and 2025 will be a multiple of that. So, it's one of the fastest growing areas for us as well. So, these are our solutions that we have.
It's taken us a long time to get them, but as you can see, there's some focus areas down the bottom in red, but it's, they're all pretty exciting. There's one kind of. I'll give you an example of something that basically is a combination of one. There was a chief of police, former chief of police, in Oakland who was basically robbed at a station he was filling up with gas, and three people came down and shot him. I think he ended up having like 22 holes in him. He had a gun with him. He shot one of them, and so there are people around there taking pictures. No one's calling 911. So, the only reason he was alive was because they had ShotSpotter there.
ShotSpotter got the police there right away, and then once they found out, he shot one of the guys, the other two got away. Well, what do you think they used to figure out where the other two guys were? CrimeTracer. They used that information based on the one that was shot, who his associates were, where were they, where did they go, where could they possibly be. And I think they ultimately found one in Berkeley. I think they found one in Texas using information like this. So it's an example of using a couple different solutions put together. You could then solve the case using the CaseBuilder solution as well. So we're starting to see several customers purchase multiple solutions that we have. In fact, Newport News just acquired four of them, actually. They acquired ShotSpotter, ResourceRouter, CaseBuilder, and CrimeTracer.
Can you say exactly what that information, what CrimeTracer is to help get those individuals?
I can. In fact, when I get to a demo, I'll show you the actual product itself to kind of show you exactly how that works. Let me move on to that, because I think that's important. It's a great, great question. So if you think about this, so we went public down here this fall, number one, small acquisition, about $1.5 billion in terms of the TAM. Then we expanded into search, case management, patrol management. That expanded a little bit more. And you can see the three, the LPR, the license plate readers and the concealed weapons detection is the largest TAM of all. But this is something all within the public safety and security market landscape that our company is doing, focused on and expanding in each one of those.
Some things, basically in terms of, and we will actually report our results of 25 February , so you can look at that. We've given guidance, north of $100 million. Last year was only $92 million, so that's growing. adjusted EBITDA, you know, approaching 20%. Last year was 15%, 23%. We have a very high net revenue retention rate and also a very, very low attrition rate. Four years, we haven't yet announced 24, but the four years before that, our attrition rate was less than 1%. Now think about that. If you're a SaaS company and you keep your attrition at 5% or less, you're doing a great job. We had it at 1% or less for four years. We don't spend a lot of money, only 52 cents per dollar for sales and marketing, also very small, versus other SaaS companies.
We're a big Net Promoter focused company, which is really challenging when you're dealing with like police agencies who in most cases don't like most of their vendors, but we're in like world class having a 66 score. And then just some of the other ones that I talked about in terms of the actual TAM that we have, some coverage area in terms of ShotSpotter over 1,100 miles. CrimeTracer has over a billion data points within that, and we have a lot of intellectual property, which I'll cover a little bit in terms of 34 patents. So if you think about that competitive advantage, I don't know what happened to that slide, but, you know, so we talk about patents, in terms of a high NPS score. We have 170 customers in terms of actual law enforcement agents. We've done this for a long time though too.
We don't really have any real competitors in the gunshot detection, which is the ShotSpotter. There are others that have tried and provided more proximity-based sensors out there. It's never really been effective. Although there are some that are trying and will continue to still try to compete against us, we're willing to go head to head against any one of them because we know that our technology, what it does and how it does, is better than anything else that's out there. So in terms of what we're doing and how we're growing, prior to going public, we were about $15 million in revenue, very small to go public with. Now over $100 million. I mean, you can do the numbers yourself of 600% growth, and we're continuing to grow.
And, we do that by cross-selling products, expanding what we have, adding new solutions, expanding internationally, which we're actually doing quite nicely now, having gone in, Cape Town, South Africa, Nelson Mandela Bay, Uruguay. We're also going to be going live in Brazil. We've already talked a little bit about that, and Bahamas as well. Strategic M&A, when I say strategic, what we will not do is we will not buy companies just to stay on revenue. Okay? There's a lot of companies that are out there. They say, "Well, we got to grow top line." Well, yes, you have to grow top line, but you also want to be careful. So, we'll buy strategic M&A, and we've done that about every two years. We find something that makes sense. That's all gone into our SafetySmart Platform and has been very effective.
Of course, customer retention is incredibly important for us. In terms of the gunshot detection, this is just a picture of kind of what it does, how it looks like. You have some of a shooter that's right there and you've got sensors that are out there. What they do is they triangulate or multilaterate. In most cases, we have six or seven sensors that go off, for each one. And then the time of arrival from that gunshot to the different sensor goes into our algorithms, which then gets reviewed to determine whether it's an actual gunshot or like a firecracker. If you have a gunshot, it's generally going to go sensors that are out that way. If you have a firecracker, it's more around, it's more circular.
So it's a lot of things that we've built over time, using a lot of AI algorithms to continue to upgrade our solutions because we don't have a lot of competition. We compete against ourselves. So that's what we do. It's very, it's incredibly effective. This is kind of what it looks like on the police agency. You can have it on your phone. So like on the right, you can look. Someone has in that particular case on the far right, there's like seven different, you know, gunshots that went off in this particular patrol area. They wanted to figure out the 10. So, there's one where, they actually went to the number that was 10. They click on that information, and when you get that information, you can listen and hear the actual gunshots.
We'll also give them additional information like there were multiple shooters there or it was a vehicle going three miles an hour to the east. This is the information that we provide when we actually give them an alert. Incredibly important and also very helpful and safe for the police as well. Yes, sir.
I'm new to the story, but I noticed you lost Chicago. Could you comment on that and the guidance that you gave that included the person on that customer?
Yeah. So, the question was, we lost Chicago. That is absolutely true. I would say Mayor Brandon Johnson lost Chicago in terms of putting ShotSpotter out of Chicago. We had the support of the superintendent, we had the support of the city council, we had the support of the community that was there, but he ran his campaign on that. So, he ultimately did shut that off in November. So, basically, in terms of this year's numbers, yes, they absolutely included the revenue from Chicago. Now, but let me answer your question even more importantly. The guidance we gave for 2025, Chicago's no longer there. Did our revenue go down $8.5 million that Chicago's paying us? No. Our revenue's going up. Our adjusted EBITDA's going up.
So even though we're losing and lost Chicago, which is one of our larger customers, we are still growing. We're growing as a company, top line revenue. We're also growing from adjusted EBITDA. So that's because even though we've lost Chicago's miles, we've added more than two times the miles he's taking away over the last two years.
What was his rationale?
He ran his campaign on that. He had three big things on his campaign that he was going to do: defund the police, get rid of some gang database, I believe, and don't quote me on this, but also cancel ShotSpotter. He couldn't defund the police. I don't think he did very good on the gang database. So, what he had left was he had to do something that he committed to his, you know, proponents that were out there, and that's what he did. Now, to be very, very clear, Chicago needs gunshot detection. They really do. There's been a lot of people since they shut it off that have been shot and/or killed that were in the areas that we were deployed, and no one called 911.
They died or they got injured and things that otherwise at least they would've had an opportunity to basically live, so we're hoping that at some point, you know, we get back in there 'cause they really need us. But from a revenue perspective, it's not going to hurt us. It'll our revenue growth is going to grow smaller as a percentage without that. But we're still going to grow revenue, still going to grow adjusted EBITDA. Okay. The ResourceRouter, this is the patrol management. This is the one where we basically tell the police officers, "Okay, you have less people. What do you need to do?
You have to use some kind of technology in order to use those resources in a better way." If you take a look at just some of the quotes, you know, "Use AI and machine algorithms to provide risk assessment for where and when crimes are likely to occur. The system directs police these areas to suggest non-enforcement actions to prevent those crimes from happening." Okay. And then the next one basically says it's exactly what it does. It allows them to meet expectations of modern-day policing, which where you have less police officers by using data-driven evidence-based accountability. So these are things that, it's also why we sold more of this in 2024 than the prior three years combined. CaseBuilder solution, I think I've talked a little bit about this already.
This is the one that the Department of Corrections is using, but it's also a way to manage cases. If you're a detective, okay, you're out there, you have to solve cases. You get rated on how you perform on how many things you solve. So a lot of them are using like paper things. They're using Excel spreadsheets. They're using other things to manage their cases instead of an actual software solution, which we've built. The reason it's so powerful is because it took us about a decade to figure out how to do it with NYPD. If you can make NYPD successful using managing cases, which is one of the most sophisticated police agencies out there, that's why what we built into a cloud-based solution is so effective.
So this is the one where you asked the question about how did we, actually find the information. And I'm sorry you can't read a lot of this, but this would be like that one person would be over here. So this is the one where you figure out exactly all the information that you can about that person. Okay? That's the information where, again, if it said Alan Stewart there, it would have my one son, Adam, lives in Tacoma on this street. My son, Alex, lives in San Diego on this street. Okay? I have three friends that I hang out with, and they're named Bill, Mark, and David. I mean, that's the kind of information that is actually in there back there in the database.
So in terms of when Ersie Joyner got shot, that was the information that they had based on the one person and the data that they had about the one person that they knew 'cause he got shot and killed about the other people that he was tied to. So if you can see here in this particular case, it might be like the gun that he was using was used by someone else in another location. It might be that that person got arrested for doing a burglary with the other person. That's the kind of information that's in the database. That's why it's so important and why it's so, so interesting. There's not a lot of other solutions out there. We do have some competition in this area.
The FBI has a database, although it's what I've been told is it's not very effective and it's a little older. LexisNexis tries to do this as well. But we have the largest database and we, we believe we have the most effective one. Yes.
Okay. So, just an apology. So, a lot of people reacted after the fact. Do you have the opportunity to develop something effective to prevent it? So, I think that is one of the ones for Chicago moving away from that.
That's a great question. The question was a lot of our stuff is reactive. Do we have anything to do that's preventative? Well, the short answer is yes. And I would say the newest one would be the PlateRanger, which is a license plate readers where you can determine if someone's going to a place. Okay? Now you may not know if they're going to rob a bank. Okay? Then you're going to go back and look what it did. Okay? But more importantly, it's the last one. It's the SafePointe, which is the weapons detection solution. That's where I'm walking into a building or a school or a hospital with a weapon. I haven't shot anyone yet. Okay? But we're going to give that information to the security team within less than 20 seconds.
They have a camera so they know that it's the bald short guy. Okay? That's me, you know, that has the gun. And they're going to figure out what they're going to do about that. Okay? So that's preventative as well. Most of the stuff though is reactive, but it doesn't mean it's not helpful or incredibly important. So if you think about a ShotSpotter, if I'm shooting and I shoot him, okay, I'm gone. Okay? But he's there bleeding out. He saw me do the shooting. Okay? The shell casings are still here. So even though it's after it happened, if no one calls 911, which in our communities happens 80% of the time, no one calls 911, he's going to bleed out and die. And he's going to be gone. And I'm definitely gone.
So even though it's after the fact, it's still incredibly helpful to save someone's life or maybe use those shell casings to figure out, okay, I was using those in three other, you know, gunshots and then they find out where I'm at. So it doesn't have to be preventative to be incredibly impactful and important. Yes, sir.
Can PlateRanger tie into ResourceRouter? So can a plate be read, you know, the background of who's potentially driving the vehicle and then resources be allocated to say, "Hey, these people with this sort of background are driving towards a bank, let's say, and let's roll a couple of cars that way." Or is it that dynamic?
Yeah. So the question is, can you use PlateRanger to, maybe prevent a crime based on the information that you have, from the actual license plates that you're seeing? The short answer is yes, but I would also say that you have to be pretty actively looking at what every single camera is doing in order to do that.
Now you can have something where, okay, and this is, I can't say it's necessarily in our product roadmap, but if you have a license plate and that's in the database, it's in CrimeTracer, okay, and all of a sudden you get a little red light that goes off, okay, and to be very clear, this is not something we developed yet, but exactly what you're talking about could be possible because you know that that particular car using that license plate was seen in several other bank robberies or, you know, car thefts or things like that. Okay? So that is possible for sure. That's a great question. All right. So here's a SafePointe. This is something if you look at the far, the top right one, this is basically what it looks like. Actually I'll use this one. This is even easier.
So if you're going into somewhere, you're walking through these things, they really just look like either planters or ashtrays or something like that. It's a covert weapons detection system. Okay? When I say covert, that's really important because you don't want people to know they're walking through. You might have to put up signs and say, "Hey, cameras and weapons detection system are in use." Whatever the law requires you have to do. But what is even more importantly is you're going through that and they see a camera and then they figure out, "I'm carrying a you know a Glock 9 on my right hip." And then so you could figure that out, take a look at the picture of it, and then make sure that you go after me. That's something that this does. This is something that we're really excited about.
You can see just a couple of examples like down on the right. See the box around that's yellow? It's yellow because that person has a weapon. It's yellow though because we also know that that's the security guard for that particular business. So it's not red. The one on the left of that is a red box. The guy's carrying a rifle, rifle. They don't know who it is. That's not a police officer. That's someone that they need to be really, really concerned about. This is something that, again, very small revenue for 2023, revenue multiple times that 2024, and this can be significant revenue growth for us in the future. Yes, sir.
I'm sorry. I think if I got this right, you said that the target market were hospitals, casinos.
For us right now, it's hospitals, casinos. We get a lot of calls from schools right now, which you would expect. But here's the challenge. We would love to sell to every school, but we also want to make sure that we're setting expectations appropriately. There are customers that, or there's companies out there that have not done that very well. I won't name them, but they say it'll do one thing and then it won't do that or do something different. We're not a company that does that. We're a company that tries to set expectations appropriately. So for us, we don't believe that the schools, frankly, have the right security team in most cases to be able to respond appropriately, fast enough in something that, you know, you'd see going through there.
So we've chosen at this point at least to not focus on that. But we still have deployed in a couple of schools 'cause they basically begged us to, but it's not a focus area for us for that reason. We want them to be able to respond appropriately instead of having something, you know, go in there and then that someone gets shot up because they didn't have the right people to look at the camera and figure out, you know, where to go. So that's why we're focused in the areas we are. So it's hospitals, it's casinos, it's some financial institutions, some big ones that you would know who they were, and other areas like that. PlateRanger, this is the license plate reader. And this is also kind of what you said in terms of the question.
You know, this is the kind of type, look at the, just the red check boxes. Okay? The make, model, color, the type, class, category, direction of travel, plate, number, state of origin. These are things that we can have. It's a very high quality image that we have that allows us to then take that information and potentially tie it into something else. How am I doing on time? I still got a little more time. Okay. So that really we do believe is a perfect solution for all agencies. Collect the data from multiple sources, organize all of that together, and then deliver something, which is exactly kind of what you're talking about. So our growth strategy is pretty basic. You know, we want to make sure that we have a tight product market fit back to the schools. Okay?
It's a great product that some schools need. They need to figure out where, you know, weapons are coming in, but they need to be able to do something with that. Develop high quality solutions, figure out what our go-to-market strategy is. And then we're huge focus on customer success. We have a customer success team that five years ago had two people, now have 25. Because what we do is we want to make sure that every single customer we have understands what best practices are. They aren't selling to them. So, they're not there trying to increase revenue. They're trying to make sure that, okay, here's what the customers that we have that are using it effectively are doing. Okay? And they're doing it really well. You might want to try to do this. That's what our customer success team does.
All of that drives Net Promoter Score, which is basically we have a very high Net Promoter Score. It's in the world class range of 66. But we want that to be very high because if you're, let's just focus now just on like a police agency. If you're a police agency and you like what SoundThinking does, you're going to talk to the other police officers in other cities and say, "Hey, if you want to do something with, case management or you want to figure out where to put your officers, you need to talk to SoundThinking." We don't have to have our sales team do that because the chief of police is doing that. What kind of commissions are they getting from us? Zero. That's why our cost of sales and marketing is incredibly low because we have our customers selling it for us.
All of those things increase the long-term value and reduces the cost of capital. So that gives us a sustainable long-term profit growth. Just a couple of things here real quick. And again, we haven't given the actuals, but you can see we've continued to grow. This is something that we're expecting to continue to do. If you just highlight, just really kind of go down to the bottom, you can see from our adjusted EBITDA percentage of Q2, it rolls around a little bit depending on where we do certain things. Like Q1, it was a little lower 'cause we had our all-hands meeting. Our all-hands meeting costs like $1 million to bring 300 people together from across the country that gets hits one particular quarter and then it goes back up into Q2.
So these are things that I think are important to understand, but also our revenue continues to grow, and it's something that I think is incredibly important. And so our net revenue retention rate is solidly positive. We had a couple of years over 120%, always over 107%. It will always be over a hundred as far as we can tell. And then very low sales efficiency. I'm sorry. Our sales efficiency is very effective because our cost is very low. Sorry about that. And then one thing that I think is incredibly important because right now, last year we were at 15% adjusted EBITDA. We can get this company to 40% adjusted EBITDA. Now think about that.
If you think about a company that's even if we're only growing 10%, but we had 40% adjusted EBITDA, for those of you who follow SaaS companies, like a rule of 40, we'd be a rule of 50. Okay? And if we happen to hit 15%, a rule of 55. At that time, at that value, the market cap goes significantly up. It's something that we know we can get to in the next three to four years. And we can get there because the last two solutions that we acquired, they haven't quite hit that break-even point yet. They don't need to grow very much in order for that cost of goods sold to be a smaller percentage and gross margins go up. As gross margins go up, adjusted EBITDA goes up.
Because we've added a lot of cost already in sales and marketing and R&D in the last couple of years and don't need to add a lot in G&A. As gross margin goes up and OpEx goes up less than revenue, what happens? adjusted EBITDA goes up. That's why we know where we're going. And if we continue to do and execute as we say we're going to, we'll get to that 40% adjusted EBITDA and a gross margin that's about 70%. So with that, I think I have a couple of minutes left for additional questions. I've already had a couple of really good questions. Thank you for those. Any other ones? Yes, sir.
Can you say the customer or can you be independent, and what does the cashless conversion look like?
Yeah. So the question is, basically in terms of CapEx. Okay? So we spend about $6 million a year in CapEx. That was primarily related to the ShotSpotter in terms of the actual sensors 'cause we own them. All we're doing is providing managed services, subscription to the customers. We have to take care of all those sensors. So that was about $6 million. As SafePointe continues to grow, those bollards that you saw, I'm gonna need to purchase those. So it is likely that our CapEx will continue to go up, year over year. Two years ago, we had a CapEx of about $10 million, but that's because we had to replace all the sensors from 3G into LTE sensors. That was a one-time, one-time deal. We won't have to do that again.
So, $6 million, you can see it go up a little bit as SafePointe grows, which would be a good thing 'cause that means revenue is growing significantly there. In terms of free cash flow, we generate cash. Okay? So, for example, let me give you specific examples. The beginning of 2024, we had about $5 million in cash. The beginning of 2024, this is last year. Okay? We then paid down about $3 million of our debt. We only have $4 million in debt, by the way. Paid that down. We paid out a couple million in bonuses. We repurchased $6 million of our shares in the open market. So, if you just added those alone, that's like $12 million. At that point, the five would've gone to minus seven. The last time we reported our cash, we had $15 million in cash.
So, we generate a lot of cash is basically the answer to your question. And most of that is free cash flow. This year I would expect in 2025 it'll be very similar. In other words, north of $15 million a year in terms of generation of free cash flow. Yes, sir.
I believe New York is up for renewal. Do you provide some color around funding size currently and perhaps what can be CapEx in the past?
Sure. Yeah. So, the question is New York is also up for renewal. And this was, if you look at some of the analyst reports, they're like, okay, well, Chicago canceled. Does that mean other big cities are gonna cancel? I'll tell you this. I have absolutely zero concern that New York is not gonna renew. So, it was due at the end of December, but it always takes a couple months, frankly, for them to do that. So, we would expect to get that in the next month or so. No, no concerns at all that that's not gonna happen. Great question though. Yes, sir.
From a competitive perspective, I've seen at least one venture-backed company do some growing in Vegas. Maybe you could talk about the competitive landscape and that seemed like an interesting solution. It's still very early, but are those competitive than the complementary?
Yeah. So that's a great question. So, in Las Vegas, in terms of a drone company, we think drones would be absolutely amazing to deal with ShotSpotter. Because if you have that, and by the way, we have been in some discussions with some cities about this. Okay? If you have drones that are maybe on the top of, you know, fire stations. Okay? So, every fire station has a drone. And then you got a ShotSpotter that goes out over here, boom, drone comes up, gets there in 40 seconds. Maybe 10 seconds out, 30 seconds there. Depends how close it is. That's faster even than a police officer that happens to be in the same area. You can get there, get the little camera, figure out exactly who that is.
So that's something that, as drones become more capable and as they deal with some of the FAA issues with the 400-foot limits, we think that most of our customers, or many of them, will be trying to implement drones alongside with some of our other solutions, specifically the ShotSpotter. It's a great question, and it is something that I think most cities are gonna go to ultimately. I think I'm out of time. All right. Thank you very much. Great questions. Appreciate it and have a great rest of the day.