SSTI - NASDAQ

Good afternoon, and welcome to SoundThinking's First Quarter 2025 Earnings Conference Call. My name is Kevin, and I'll be your operator for today's call. Joining us are SoundThinking's CEO, Ralph Clark; and CFO, Alan Stewart. Please note that certain information discussed on the call today will include forward-looking statements for our future events and SoundThinking's business strategy and future financial and operating performance. These forward-looking statements are only predictions and are subject to risks, uncertainties and assumptions that are difficult to predict and may cause actual results to differ materially from those stated or implied by those statements. Certain of these risks and assumptions are discussed in SoundThinking's SEC filings, including its registration statement on Form S-1. These forward-looking statements reflect management's beliefs, estimates and predictions as of the date of this live broadcast, May 13, 2025, and SoundThinking undertakes no obligation to revise or update any forward-looking statements to reflect events or circumstances after the date of this call. Finally, I'd like to remind everyone that this call will be recorded and made available for replay via a link available in the Investor Relations section of the company's website at ir.soundthinking.com. With that, I'll turn the call over to Ralph. Ralph, please go ahead.

Good afternoon, and thank you for joining SoundThinking Q1 2025 Earnings Call. I'll start by providing some high-level commentary on our financial results, progress we're making on our key strategic initiatives and then share some thoughts on the headwinds and tailwinds we are seeing in the public safety and security marketplace. I'm pleased to report that we're off to a strong start in 2025, delivering disciplined growth and expanding our platform and data aggregation capabilities to help position SoundThinking as a clear leader in the public safety technology space. Our work matters, and we believe it's resonating in the market. I'm proud of the momentum we've built and even more excited about what's ahead. In the first quarter, we delivered 12% revenue growth year-over-year of $28.3 million, driven by solid new sales and renewal activity. As a reminder, there was approximately $3.5 million of catch-up revenue in the quarter based on the renewal of two delayed contracts with New York City Police Department, including ShotSpotter and Technologic, which totaled $64 million over a 3-year term. Our adjusted EBITDA grew 50% year-over-year to $4.5 million, highlighting our operational leverage and profitable growth strategy. ShotSpotter remains a core part of our SafetySmart portfolio and is trusted by public safety agencies across the country. In Q1, that trust was reaffirmed with NYPD's decision to extend their ShotSpotter contract for an additional three years. We believe that this renewal with one of our largest and longest-standing customers, 12 years and counting, speaks volumes about the sustained value and operational reliability of our gunshot detection platform in America's largest city. We also went live with 4 new cities plus one expansion in the quarter. Our international ShotSpotter pipeline is robust and growing. We're very pleased to see a soon-to-be live ShotSpotter deployment in Niteroi, Brazil, marking our return to this strategically important market. We will now have two strong and important reference customers in both Spanish-speaking and Portuguese-speaking Latin America. We expect to see accelerated traction with ShotSpotter internationally in the second half of this year and early 2026 based on these touchstones. Let me also take a moment to address the recent 5-year Chicago gunshot detection RFP, which we bid on in April. We believe our submission represents a comprehensive and compelling proposal that reflects our deep experience, proven performance and long-standing partnership with the city, which spanned well over a decade. We fully respect the integrity and objectivity of the procurement process, but quietly feel confident in the strength of our offering and the differentiated value that we will bring back to the city of Chicago. And while we await the outcome of the formal bid adjudication process, I must also emphasize that our current outlook does not include any contribution from Chicago. Any potential reengagement with Chicago would represent pure upside. Our CrimeTracer solution is evolving into one of the most powerful AI-enhanced law enforcement tools in the country. We warehouse over 1 billion CJIS compliant documents for more than 1,000 law enforcement agencies and federate to billions more through Index, Navy Links and Thomson Reuters, Clare, giving investigators access to an expansive depth of structured and unstructured data. But what we believe makes CrimeTracer truly unique is how we can apply generative AI and soon-to-be large language models and agents to make that data just not searchable, but contextually insightful. This quarter, we deployed new features that allow investigators to ask natural language queries like show me persons of interest who match known patterns in recent thefts involving white bands in three counties and receive actionable insights that would take hours, if not days, to surface manually. The combination of search, summarization and synthesis is transforming what's possible for digital case work and crime linking. Furthermore, we also made progress integrating CrimeTracer's large data footprint with PlateRanger's ALPR data to unlock powerful response and investigative use cases for our customers. We believe that this integration provides a force multiplier for real-time and retrospective investigations, allowing investigators to seamlessly move back and forth from narrative-based incident search to connected license plate recognition. We believe this ability will enable agencies to help solve cases faster with fewer resources. The early feedback has been extremely positive and has proven to be a powerful differentiator in driving PlateRanger traction. More importantly, we view this as foundational to move into building a truly multimodal investigative intelligence platform that combines people, places, vehicles and incidents in a single pane of glass. We continue to scale ResourceRouter, our proactive patrol planning tool with strong adoption and demonstrable results. ResourceRouter is currently deployed in over 20 agencies, more than double its installed base less than 18 months ago. Agencies are using it to allocate limited officer resources to areas with the highest probability of criminal activity, and we're seeing real-world impact in both community engagement and crime suppression. As public safety agencies face mounting pressure to do more with less, ResourceRouter seems to be an essential capability, not just a nice to have. Now let me turn to one of our most exciting growth areas, SafePointe. Following the passage of California Assembly Bill 2975, hospitals are now required to deploy weapon detection systems at all public entrances by 2027. This is a seismic shift in health care security's policy and plays directly to SafePointe's strength. Our system uses passive low-frequency magnetic field detection, not invasive scanners or disruptive walk-through gates or intimidating checkpoints. That means hospitals can maintain patient dignity, visitor experience and operational throughput, all while meeting the new legal compliance standard for safety. We've already engaged with multiple large health care systems in California, and we see similar legislation advancing in Maryland, New Jersey and beyond. With over 400 hospitals impacted in California alone, each with multiple entrances, the now mandated and addressable opportunity is significant. We believe SafePointe can become a category-defining solution in this space, not only for hospitals but for any environment where high flow, high-stake security is required. As we move further into 2025, we remain bullish and focused on executing to our growth strategy and strategic road map. And while there have been very positive tailwinds to our business, driving strong demand for our capabilities, we remain acutely vigilant around lingering headwinds, especially those related to municipal funding and budgets. We are taking appropriate risk mitigation efforts, including doubling down on our customer success and engagement efforts, adding a grant writing resource and modeling higher levels of attrition than what we have historically experienced to account for that volatility. To that end, we continue to believe that we are well positioned to drive both revenue and ARR growth for 2025. We are reaffirming our full year revenue guidance range of $111 million to $113 million, while slightly reducing our adjusted EBITDA guidance range to 20% to 22% to account for the modest impact of the current tariff regime, along with investments we are making in AI modeling and tools in AWS that are being incorporated in our products as well as for our internal operational use. I will now turn the call over to Alan to discuss our financial results for the first quarter 2025 as well as guidance for the full year 2025 in detail, and then we will be happy to take your questions.

Thank you, Ralph, and good afternoon, everyone. As Ralph mentioned, we are pleased with our first quarter 2025 results. Our strong financial performance reflects the success of our ongoing strategic initiatives, operational efficiency measures and our commitment to delivering value to our shareholders. Revenues were $28.3 million, representing a 12% increase over the $25.4 million in the first quarter of 2024. As Ralph mentioned, our 2024 year-end financial results were impacted primarily due to the delayed renewal of two contracts with New York City Police Department. Both of the NYPD contracts were signed in the first quarter with approximately $3.5 million of the catch-up revenue from the contracts. We're also pleased to report that both of those contracts were 3-year renewals, representing over $64 million in contract awards. The bookings of all of our SafetySmart Platform solutions, some of which are multiyear contracts are also growing healthily. Gross profit was $16.6 million or 59% of revenue compared to $14.9 million or 59% of revenue for the prior year period. Our adjusted EBITDA was $4.5 million compared to $3 million in the first quarter of 2024. Our adjusted EBITDA increase was related to the delayed contracts. As a reminder, adjusted EBITDA, a non-GAAP financial measure, is calculated by taking our GAAP net income or loss and adjusting out interest income, income taxes, depreciation, amortization and impairment, restructuring costs and losses, including those related to fixed asset disposals, stock-based compensation expenses and acquisition-related expenses, including adjustments to our contingent consideration obligations. Our adjusted EBITDA was a bit lower than consensus estimates as we held our company all-hands meeting, which cost over $700,000 in the first quarter. This onetime event happens only once a year, generally in the first quarter. Our legal and accounting expenses are also historically higher in the first quarter as we complete our annual financial audit and our preparation for and filing of our 10-K. And as Ralph mentioned, we are investing more in our AI capabilities as well. Our operating expenses were $17.8 million or 63% of revenues compared to $17.5 million or 69% of revenues in the first quarter of 2024. Breaking down our expenses, sales and marketing expense in the first quarter was $7.3 million or 26% of total revenue compared to $7.1 million or 28% of total revenue, in-line with the prior year period. Our R&D expenses were $4.1 million or 14% of total revenue compared to $3.6 million or 14% of total revenue in the prior year period. G&A expenses for the quarter were $6.5 million or 23% of total revenue compared to $6.8 million or 27% of total revenue for the prior year period. As a reminder, we expect our G&A expenses to grow less than our revenue on a percentage basis as our company grows. It's important to recognize that the percentage of revenue of each of our OpEx categories is at or below the level of the first quarter of last year. Our GAAP net loss was approximately $1.5 million or a loss of $0.12 per basic and diluted share for the quarter based on 12.6 million basic and diluted weighted average shares outstanding. This compares to a net loss of $2.9 million or $0.23 per basic and diluted share based on 12.8 million basic and diluted weighted average shares outstanding for the prior year period. Deferred revenue as of March 31, 2025, was largely in line at $45.4 million compared to $44.2 million at the end of fourth quarter 2024. We ended the first quarter with $11.7 million in cash and cash equivalents compared to $13.2 million at the end of fourth quarter 2024 and much higher than the $5.7 million that we had at the end of 2023. We repurchased 33,493 of our shares at an average price of $15.04 for approximately $504,000 in the first quarter of 2025. Currently, we have approximately $21 million available on our line of credit as we have approximately $4 million in debt outstanding, all of on our line of credit. Now turning to our guidance for the full year of 2025. We are reaffirming our full year revenue guidance range of $111 million to $113 million. We are reducing our full year 2025 adjusted EBITDA margin guidance range to 20% to 22% related to potential costs associated with tariff changes and the investments that we are making in AI modeling and tools that we are incorporating in our products and internal operational use. We are reaffirming our expectation for our annual recurring revenue, or ARR, to increase from $95.6 million at the beginning of 2025 to approximately $110 million at the beginning of 2026. As a reminder, this guidance is in spite of the loss of approximately $9.7 million from the loss of the Chicago ShotSpotter contract in 2024. Overall, we are pleased with the progress we have made on each of our strategic initiatives and operational performance of the business. With that, we're now happy to open the call for questions. Operator, will you please open the line for Q&A?

Certainly. We'll now be conducting a question-and-answer session. [Operator Instructions] Our first question today is coming from Richard Baldry from ROTH Capital Partners. Your line is now live.

Thanks. Now that you've got a broader suite of tools, can you talk about how you manage sort of the pipeline there to make sure you're not overweighting or underweighting resources in each of those areas?

Yes. Thanks for that question, Rich. This is Ralph. So each product solution has its own pipeline metrics that we're tracking. And I would say that the pipeline is fairly solid across the platform and particularly strong in a couple of segments, including ResourceRouter and CrimeTracer. We're targeting kind of 3x to 4x coverage based on the annual contract value quotas that we've issued out to the sales team.

And then maybe looking at SafePointe specifically, I think you've revamped that product in the past year. So is there any early feedback on that? Maybe any like customer cases, maybe without naming names, maybe verticals you sold into at that?

Yes. So thanks for that question again. This is Ralph. So as you know, we're focused on the health care vertical, casino vertical and then high-end commercial properties. And we've seen really success across the board, particularly in the health care vertical, where I think our discrete architecture is critically important to a health care concern. I think we talked about on our last call that we have a couple of proof-of-concept deployments going in on one of the top five, top 10 health care chains here in the U.S. One of those has actually kicked off and it's going pretty well. We're hoping to light up the second one here in Q2. So more to come on SafePointe in the health care vertical.

Next question today is coming from Mike Latimore from Northland Capital. Your line is now live.

Hey. Hi, this is Aditya on behalf of Mike Latimore. Could you give some color on if do you expect meaningful bookings to come from PlateRanger this year?

Do you want to take that, Alan?

Yes, sure. So this is Alan. We are still working on the actual bookings. We have a significant amount of pipeline in there. But -- so I would say it is still going to be somewhere in the 1 million-plus range in terms of actual bookings, and we're expecting slightly a little less than that in revenue this year, but significant increase in revenue into next year. We are seeing a lot of interest there, and we are running a couple of different kinds of marketing aspects to get more customers into the actual contracts.

Got it. And do you see additional expansion in Puerto Rico this year?

Yes, this is Ralph...

I think we're -- go ahead, Ralph.

No, go ahead. Sorry.

No. I was saying that Puerto Rico, as we mentioned before, the contract that we have right now ends at the end of June. We are in ongoing discussions with them to not only extend that contract, but to make that into a multiyear extension renewal. We are also in discussion with them. I won't give you too much more details on that, but with some potential expansions beyond the 30 miles that we have there. So things are going positive there. But as we all know, it just takes some time, especially dealing with Puerto Rico.

Thank you. [Operator Instructions]. Our next question is coming from Trevor Walsh from Citizens. Your line is now live.

Thanks. Hey Alan and Ralph, thanks for taking the questions. Maybe rotating back to the SafePointe piece. Great to hear the legislative sort of win, I guess, in California. Can you maybe, Ralph, just give us a sense of based on kind of where you see the sales cycle around SafePointe kind of executing now currently and then based on that 2027 requirement, when you think deals kind of really start to materialize around that? I would imagine maybe more into 2026 just based on the timing, but how -- I guess, how far in advance do customers need to start kind of prepping for that new requirement?

Yes. Thanks, Trevor, for that question. This is Ralph. So I think you're absolutely right. With respect to the AB 2975, that's really much more of a 2026 opportunity for us. But make no mistake about it, we're making really, really good traction in other places in the country with SafePointe, not only in hospital vertical, not only in the hospital vertical, but also in the gaming casino vertical as well. So what we're -- our game plan there is to have a fairly comprehensive campaign in California to really position ourselves within that market because now as we've often said in previous calls, our greatest competition with respect to SafePointe is do nothing. And now this particular unfunded mandate coming from the state of California is taking away that competitive energy or that competitive thread there would do nothing because now at least 400 hospitals have to do something along the weapons detection space. And we think because of the uniqueness of our offering with respect to the passive sensors, it's discrete deployment and not having a checkpoint that we're well positioned to get our fair share of that particular market. But we're going to have to invest some time and energy, certainly working with CalOSHA and the various participants that were kind of behind getting this bill passed in the first place, along with hospital chains in California. So we're pretty excited about it, as you might imagine.

Awesome. Great. And Alan, maybe my follow-up for you, if I can. I recall from last quarter when we were discussing the New York kind of pushing of the deals, I guess, into this quarter that there was going to be some at least benefits of the deal sliding around the margins because of kind of the costs being already outlaid and then the revenues hitting kind of later. But it didn't seem like that necessarily would materialize by at least just kind of looking at kind of where gross margins lined up in the quarter. So just wondering if -- maybe it was other things that you had called out in the quarter around some of the other costs and the annual meeting, et cetera, that maybe kind of washed that out a little bit, but just curious kind of what -- how that materialized in the numbers?

Yes. This is Alan. Thanks for that question, and it's a valid question, too. I think what we are seeing already, at least in Q1, we hit the adjusted EBITDA of the $4.5 million, but we had about $1 million in additional expenses. Our all-hands meeting was north of $700,000. We also did have some increase as we do the AI algorithms and the modeling. That increased our actual bandwidth usage of like AWS and Azure. That was probably a couple of hundred thousand dollars as well. So when you start adding the $1 million in terms of expenses, had those not occurred, then the adjusted EBITDA would have been significantly higher, specifically related to the New York City contracts, a lot of that did flow to the bottom line, as you would expect. There was still some things that we had to cost a little bit related to some actual commissions that were owed to the prior company that was doing some of that as well as a little bit of cost of goods sold cost related to that.

Next question is coming from Eric Martinuzzi from Lake Street. Your line is now live.

Curious to know about the revenue progression given the outsized kind of the $3.5 million lumpiness that occurred in Q1. How should we be thinking about the progression to Q2? I know at least what I've got for consensus is 27.8%, which would be a step-up from the 28.3% minus 3.5%. Just wanted to make sure you're comfortable with that step-up.

Yes. So this is Alan. It's a great question. And having that 3.5% out would have put us back down to around 24.8%. So north of 27% is a little bit too high in terms of what we need in order to hit our revenue guidance for the year. I would -- we don't give quarterly revenue guidance. But for the annual, it would be more likely to see a reduction into Q2, Q2 going up to Q3 and Q3 going up more significantly into Q4 as all the stuff that we're investing in right now, especially the SafePointe starts recognizing a lot more of the revenue.

Got you. And then on the margin guide, the 100 bps, did you -- maybe I missed it, but did you clarify what the gross margin implied in? Is -- are we still kind of a 59%-ish for gross margin and the impact is in the operating expenses?

Yes, it's another great question. We're still expecting our gross margin to stay around there for the year. As we continue to invest, we are seeing some of those costs. When you talk about the change in the adjusted EBITDA percentage, it's only a little over $1 million. But you've got some of that's going to go through cost of goods sold, such as things like the increase in sensors that we deploy for the ShotSpotter. That might be somewhere around $250,000, maybe $300,000 between new sensors and replacement sensors. So that's going to affect that a little bit. And then the rest of the stuff that we're doing in terms of the AI algorithms and then investing there for all the appropriate things, especially for SafePointe into the algorithms and modeling. So that's going to be from the bandwidth costs, AWS, Azure as well as personnel. You can add that another probably $700-plus thousand throughout the year, not obviously quarter-by-quarter, but that's something that is going to keep the cost of goods sold a little lower. But I think it's still important to understand that overall, our operating expense is growing and will continue to grow less than the revenue, even with those additional costs. So it's something that we are managing. And you can even see that from Q1 last year, our total OpEx was $17.5 million. This year was down to $17.2 million even though revenue went up.

Next question is coming from Jeremy Hamblin from Craig-Hallum, your line is now live.

Thanks for taking the questions. And I want to come back to AB 2975, SafePointe. And just understand in terms of the total number of facilities within the state of California that you believe to be required under this new law to put in place detection. That's kind of part one. And then part two is, I think you mentioned Maryland, New Jersey, Virginia as other states looking at similar type of legislation. If you add those states in, then what's kind of the total number of addressable facilities that we're talking about?

Yes. Thanks, Jeremy. This is Ralph. With respect to the state of California, we're estimating -- this is directionally true, I'd say, we're estimating that there's about 400 hospitals and estimating that each of those hospitals kind of given the fact that there's maybe one to three buildings kind of per hospital and location would have on average something on the order of kind of 10 egress -- ingress -- egress points that they'd have to protect. There is a carve-out certainly for a number of hospitals that are in rural settings or more kind of clinics that don't have to comply with this particular legislation. But we think that there's probably about kind of 400 hospitals in California that would have to comply, and we're estimating about 10 lanes per. So about a 4,000-lane opportunity in California. Quite honestly, we're not looking -- well, we're not really looking beyond California with respect to this type of legislation. We just want to make note that kind of given the workplace violent issues that exist pre-eminently in hospital settings that, it wouldn't be a surprise if there are some other states that are kind of in the early discussions of doing something similar to California might flip over and have similar legislation. We know in the state of Illinois, for example, although it doesn't apply to hospitals, there's a similar regime with respect to requiring weapons detections for casinos in the state of Illinois. So this is something that's going to be, I think, part and parcel for this particular market around how certain entities need to become compliant and they can't afford to do nothing. So it's a bit of a kind of unfunded mandate that we're going to be tracking on, of course.

And then as a follow-up, how does the annualized cost of the SafePointe per entry way compare versus kind of traditional solutions, whether that's a Garrett or Cheya kind of traditional metal detector weapons detection or an Evolv system that's a bit more AI tools involved. Can you give us a sense in terms of bidding on these opportunities where you fall in the spectrum?

Yes. Thanks for that follow-on question. So I can tell you what our pricing model is and business model. We're charging about $20,000 per year per lane. And that's kind of similar in many ways to ShotSpotter when you think about our cost to deploy and how we recapture that cost investment is very similar to ShotSpotter. We're breaking even on a lane inside of 12 months and then kind of the ongoing annual subscription fee to the extent that people would renew would be super high margin for us. Interestingly enough, we're finding and dealing with commercial opportunities that a number of these customers are signing up for multiyear deals, two years, three years, et cetera. From a total cost comparison, and this is pretty early and directionally, I would say that we offer a fairly compelling solution kind of given the unique nature of what we do and not having to have people walk through a checkpoint. Total cost of ownership, we think, is extremely attractive for our solution versus other solutions that require much more manually intensive resources there at the checkpoints. We don't require that. We've learned a lot from our experience in ShotSpotter in terms of just not providing the base capability of detecting weapons, but also the reporting and analytics that go along with that. So we have a kind of an Insight solution. Well, it's not kind of like Insight, it is Insight that you'd be familiar with ShotSpotter. We reimagined that to the SafePointe solution now where agencies can -- excuse me, corporate entities can basically check their ingress flow, the number of detections per passes, the number of times they can apply ground truth where they've actually pulled someone over and found a weapon, et cetera. This is an incredibly powerful reporting tool that goes beyond just the weapons detection piece that we think is, again, is going to be very attractive to folks in California, OSHA that are really about protecting these facilities and having robust data to ensure that they're appropriately reducing risk of potential workplace violence in hospitals.

Great. Great color. I wanted to also just ask a little bit more about the international opportunities that you mentioned. right? So deploying in Brazil here in Q2. And then as you think about some of the other opportunities, I know you've probably looked at some expansion in South Africa. You've noted previously some other Latin American countries that you might be looking at as well as the Caribbean. And just to get a sense for timing, I know typically, the international bidding is a little bit longer sales cycle. So I wanted to get a sense for how we should be thinking about that.

Yes. So you're absolutely correct. The international sales cycles tend to even be longer than the long sales cycles domestically, but they're super impactful. Just to remind you and others that our pricing leverage there is we typically charge 3x to what we charge domestically. So when the deals happen, they happen in fairly big ways. Hey, look, we're really excited about Latin America, in particular, kind of going back to Brazil. We were in Brazil in a couple of cities back when I joined the company in 2010 on our legacy business model. So to be able to return to Brazil is really exciting because of all the Latin American countries that have issues with gun violence. I would say Brazil is most similar to the U.S. than almost any other country in Latin America or frankly, even South Africa in terms of the intensity and scale of the problem. Brazil is a very big country. There's a lot of cities there that have a lot of square miles of coverage need there. So to actually get our first mover in Niteroi, Brazil, which is importantly that -- excuse me, Portuguese-speaking, is going to be important for us to hopefully reopen that market opportunity up. So -- and we're already in discussions without saying too much more. We're already in discussions with a couple of other cities in Brazil, kind of doing a quick follow, hopefully, to Niterói. We saw the quick expansion that happened in Montevideo, Uruguay. We're excited about that, of course, and they represent a very compelling first mover early adopter customer in Spanish-speaking Latin America. We've had a number of municipalities from other neighboring Latin America, Spanish-speaking Latin American countries -- South American countries, excuse me, that have visited Montevideo, and they've actually gone out to various Latin America security conferences talking about their success. I do also want to point out that it's just not ShotSpotter, but we're really excited about a couple of pretty interesting opportunities with ResourceRouter in both an existing Latin American country where that's already a ShotSpotter customer and then one non-ShotSpotter customer that's pretty substantial and then also true with a PlateRanger as well. And if you follow our press in anything or follow the press in South Africa, you'll see that there is a real, real need for an expanded footprint in and around Cape Town that we're hopefully going to get the Western province to support. So we've got a lot of interesting activities going on. Timing-wise, I think the best way to think about this is late, late 2025, early 2026 for some of those cards to flip over. Did that answer your question?

Yes, very helpful. Last one for me is back to the operating expenses for a second. So you saw an uptick in R&D costs both on a run rate from last year, but also from the end of '24. Was that where you're reflecting some of the AI tool investments? Is that falling into that bucket? Or is there any other color you might be able to share on whether or not that kind of $4 million run rate in R&D is kind of where we should be thinking?

Yes. Great question. This is Alan. It was primarily related to two things. One would be some of -- it is primarily for the personnel that are helping us with the AI algorithms. We're doing a lot of modeling right now related to SafePointe primarily. Most of the cost is related to that. So that's basically where it is. Also some of the -- like the AWS and some of the other bandwidth type costs ends up in the R&D bucket itself. That's primarily where it is. I do think you're going to see the $4 million -- a little more than $4 million, maybe $4.1 million go down a little bit as we stabilize those, but it's not going to go down significantly throughout the rest of the year. But it will definitely still grow less than the revenue.

We've reached the end of our question-and-answer session. I'd like to turn the floor back over to Ralph for any further or closing comments.

Great. Yes. Thank you very much, and thank you all for joining us again today. As you've heard, we're executing well across the board. We're expanding our platform, deepening customer relationships and pursuing new opportunities in AI-driven public safety. We certainly appreciate your continued support and look forward to keeping you updated on our progress. Thanks again for joining the call.