Okay, we are ready to begin with our next presentation, so please give a warm welcome to the stage, CEO of ROC, Scott Swann.
Well, welcome everyone. Thanks for thanks for indulging me here for the next 25 minutes. you know, I wanted to give you a little bit of background about ROC, but to really kind of set the tone. You know, every critical system, whether that be banking system, border system, law enforcement, defense, whatever it may be, typically, you need to answer three specific questions. Who am I dealing with? What do I know about them? Can I trust them? these are the kind of questions that ROC is really trying to assist in finding those answers to, and as rapidly as possible, within seconds, where a lot of times this can take hours or even days to really have those types of answers.
You know, identity, intelligence, they're fastly converging to be kind of this new backbone infrastructure layer, and we believe a really important one that will take us into the next decade. You know, it's important to note, and a lot of people are surprised when I say this, is that when we look at the way that we screen people across this country, the AI that's used to support that, it's almost 100% foreign owned. That's across FBI, DHS, DOD, and all of these various federal agencies. That is an area where the United States has just essentially fallen pretty drastically behind in that space. We're really trying to assist the United States in being the dominant provider in these identity technologies within the United States. You know, we grew ROC organically. It's about a 10-year-old company.
We never took any outside capital, no investment, no private equity. We became public this past February, all in a effort to really scale the company and address this problem. A forward-looking statement. You know, we're building this Vision AI platform to really unify, video, biometric technology, so think fingerprints, face recognition, iris, the objects from videos, which ranges across a large spectrum of objects, along with digital evidence to really unify into a single platform. Today, typically, if you look at agencies that have to deploy any of these types of technologies, they're usually stovepipe. That data doesn't come together. That data is much more powerful when it does come together. That's really the platform that we're trying to build is that we have built actually, is the ability to remove all this data into one single unified platform.
Our strategy that we've really put into place here at ROC is, you know, focus on not only the components, but to build out the full application software stack. You know, we got our start as a face SDK company building algorithms. What we learned was that other people were coming along and we were doing the hard part, building the algorithms, and getting contracts that were pennies on the dollar compared to the kind of contracts they would get with much better ARR and durability. We wanted to build out the components into a full platform. We also knew that our algorithms need to be among the best in the world. You don't have to take my word for that.
In the government space, there's third-party entities such as the National Institute of Standards and Technology that publish leaderboards on who are the best in this particular space. ROC is consistently being evaluated as one of the number one technology providers in this area. And then pool all this together into a single platform. Let me really give you a real-world practical example of how all this comes together into a unified platform. Many of you have probably heard of latent fingerprint technology. This is when you pull a fingerprint from a crime scene, and then you search against a really large database.
In our first evaluation by this group, NIST, that does these third-party evaluations, searching against a database of about 30 million records, the global leader in the world in this particular space, took about an hour and a half to search this database. Here at ROC, we searched that database in 15 seconds and maintain parity and the accuracy for that. That's a really major game changer. That's not just speed. First of all, that can change the way that policing happens across this country. It is true cost savings for our customers. Instead of competing based off of licensing costs, there's an overall total savings cost ownership just by the efficiency of our algorithms in this space. You can see there's a whole list of capabilities that we specialize in at ROC.
We want to get beyond just face recognition. We build fingerprint algorithms. You know, we've been benchmarked for our weapon detection, you know, when it's through visible cameras, just being able to detect whether or not someone's holding a gun. We do license plate recognition, a range of other objects. The concept is, if it's something that you can see, you know, with your eyes, we should be able to write algorithms, build this into our platform, and be able to provide services based upon that. You know, completely outside the technology, it's really, I think, one of the things that make ROC special is our people. You know, we built this platform largely from people who have lived this problem.
You know, I have people on our team at ROC that's been involved in every major event to include September 11th and helping make sure that the data is fully exploited to help support investigations. You know, we actually care about getting this right. You know, we actually wish that we had some of these kind of tools to help support these challenges when we were in service. We've combined, you know, the legacy of subject matter expertise with young artificial intelligence computer scientists to help build out our overall capability. We operate across four primary mission areas.
National security obviously is a very large one for us, but other public safety space, digital identity, physical security, and each one of these may have their own specialized mission space, but we service all this through one single unified Vision AI platform. As we go beyond just providing components, which would happen in our software development kit or SDK, which by the way, we have no intentions of stopping selling our SDK. We have a lot of pipeline associated with the SDK. This is how our platform would show up to our customers though with the other products. You know, in 2025, we essentially went to market with one of these products, which was ROC Watch. This would be our video analytic platform.
This allows us to do real-time recognition of a variety of objects or face recognition or other kind of capabilities, as well as post-process. Think about Boston Marathon bombing. It was one of my last assignments when I was at the FBI. This was the first time that the FBI opened up its tip line to the public. We were inundated with the amount of still video, still pictures and videos that were received, but we didn't have tools to actually help support that investigation because of that. ROC builds a product to support that. ROC Evidence, this would be our tool to help support the overall judicial system. You know, you have a large investigation and large events. All that data to be exploited goes into the system.
It manages all the rules, permissions, and all that capability to help support the overall judicial process. ROC Enroll is our identity verification platform. I'm sure a lot of you use this. This is essentially where you take a picture of your face and you take a picture of some sort of a credential like a driver's license or a passport, and you verify that identity. We've white labeled across a big portion of the fintech community for that to this point. Now we offer full stack solutions in that space. The particular product I get the most excited about is ROC ABIS. ROC ABIS, you can think about national identity systems, or policing systems. In the U.S. alone, we have about 19,000 criminal justice agency systems that require these kinds of capabilities.
This is where you search those fingerprints or faces against large repositories to get some sort of deposition on who a person actually is. It's important to note we're not a startup. You know, we're already a revenue-generating company. You know, we're moving toward a business model where we're trying to achieve real recurring revenue, which is a characteristic of going from components to products. You know, the that shift is really what's driving a lot of our long-term scalability. When we think about our financial, you know, 2026 is a really interesting start of a year. You know, we're coming out of a 2025 where there was no fiscal year approved by the federal government.
Those that understand federal financing understand also that the federal government has to spend and obligate their money before September 30th of this calendar year. We have, at our size, a lot of lumpiness 'cause we're very, you know, small as a overall company with some of our current product portfolio. As we move toward products and the pipeline that we're actually supporting right now, we see that there is a lot of scale and a lot of more predictability in the, in the out years. The, the story that we really are selling that we believe in is the product-based revenue. In year-over-year quarters, we saw about 77% increase in ROC Watch. It's important to note ROC Watch was the only product that we were to market with in 2025.
As we come into 2026, we're bringing four additional products to market. We've already monetized two of those additional products. Our ROC ABIS platform, we have our first couple of wins in ROC ABIS already this year, as well as we just signed our first digital evidence customer with the DEA in April. 2026 across all of our product portfolio is really intended to help establish those beachhead customers. You also see here, we're a sweetheart for the federal government for performing complex research, we have a pretty substantial research portfolio in the past. I think the government shutdowns have kind of delayed some of that award in those particular places, we're very confident that we'll have a healthy year within research.
We've also been able to take a lot of that research and transition that into real-world production. Really our gross margins validate, you know, this model. You know, we're operating with software-level dynamics here, about 80% gross margins. Same time, we've been intentionally reinvesting back into the company. You know, organically growing, we were pretty much operating at the wire trying to manage most of our money back in. As we take in the IPO funds, for the next 18 months, we're really scaling this out. We really think, we believe that we'll be on a path to return to profitability within 18 months from the IPO. This is really where this goes. You know, large-scale investigations today, they involve massive amounts of data.
You're talking petabytes worth of data, but you still have to move really quickly through that kind of data to be able to get to the right result, and you can't afford not to get that correct. We're unifying all that into a single platform. We're operating with a land and expand model also. You know, we start with pilots, we start with small proof of concepts, we prove out the technology, and we are able to convert these small component sales into larger platform-based product sales with nice ARR and durable revenue. But we really are focused on turning these pilot programs into programs of record. This is how we believe we'll build that durable revenue.
What really matters here when we think about products and platforms is that there's some fundamental changes that happen when you go from components to products. You know, first of all, the contract dynamics look much, much different. This differs a little bit based off the products that we are actually going to market with. In our video analytic products, for example, there's a lot of people in that field. Our go-to-market strategy is really to own the tactical operations side of that. If you think about all those three letter agencies, for example, we specialize because we have a lot of operators who worked in that space of being able to provide video analytics solutions in that space. They also have the most complex requirements that you could build to.
That means the rest of the market's very adjacent to us working in commercial security. It's not uncommon that if you go out to the biggest commercial firms, that they have somebody that was former FBI or former chief of police that's actually running those technology groups. That scales three year type of contracts typically in that space. In ROC Evidence, ROC Watch, those other two products that we get really excited about, you know, these are five to 10 year types of contracts. You see, you know, typically not often, or very common is, you know, a base contract with a 25% maintenance tail. Sometimes agencies like us to smooth that out across four years and see a little bit more consistent cost to them per year.
In some cases, in our ROC Evidence, there's a variable component of that too, based off of consumption. You know, as I said, some of these cases are extremely large, petabytes worth of data. We have a strategic partnership with Microsoft in this space. As they consume additional storage space or consume analytics against that data, that helps create this variable revenue dynamic along with that as well. From a capital allocation perspective, you know, we kind of have been ingrained in our, in our culture to be very frugal with our funding. We've been ramping up our staffing. We have lots of new engineers, lots of delivery specialists, lots of forward solution engineers, as well as some additional business development.
We've never really invested that much in marketing and business development, but we're ramping those up a bit as well. In the research side of the house, we need to get our team a little bit more compute resource. We operate a little bit different than other AI, where it takes massive amounts of data center capability to actually create your models. We're a little bit more narrow focused. You know, we're spending $2 million to get our team, you know, some really substantially new compute resources. We've already achieved, even through our organic approach, some parity with the leading billion-dollar foreign companies in the world in this space. This will go a long way in helping us create faster and more accurate models. It's a very large market.
We're talking about $100 billion with multiple segments. The one benefit that we have is that we're able to bridge these various segments, and we have capability that can bring that data together to make it a lot more useful and powerful for our customers. You know, earlier I mentioned those three questions. You know, who is this? What do I know about them? Can I trust them? Historically, those questions could only be answered through separate systems. We're providing those same answers within a single platform, and that's really what's creating fundamentally a new category of capability here. This land and expand model, as they take one of our products and provide it into their atmosphere, we have this ability to pull all that together and help grow those contracts with our customers.
We already talked a little bit about the team. I'd say a lot of us have worked together for many years. You know, we believe this is one of the special parts. We have extensive experience across federal government as well as the commercial sector. As you also take a look at our board, you know, we're talking about, you know, people who have been the, you know, director of the science technology of the CIA, commissioner of Boston Police Department, highly esteemed FBI executive, number three in the FBI that also became the global security director of MGM. Lots of industry experience. You know, they understand the kind of mission that we're supporting and have pretty extensive networks to be able to assist us as well. You know, as a public company, we're really positioned well to execute on this strategy.
You know, we currently have a market cap approximately around $130 million. You know, we have strong insider ownership, around 75%. This really creates a tight alignment between, you know, our leadership and our shareholders to really scale this business. Really just to wrap this up and open it up for some questions, you know, we're building a mission-critical AI environment, really where failure is not an option. You know, we have a clear path to achieving this durable annual recurring revenue. We operate with software-level dynamics with really high margins, around 80%. We've structured this strategy around a platform that compounds and can grow with our customer needs. Really building that infrastructure layer for identity intelligence, and we're positioning the United States to being the dominant player in identity intelligence for several decades to come.
With that, I'll open up for any questions. Sir.
What do you think the implications of world models are, both as a opportunity and a threat, I guess?
Could you repeat, please?
The world models, the AI models, the visual-based models. Is that a opportunity, a threat, or both for you guys?
Visual-based models, I'm not sure I really completely understand the question when you say world-based models.
More general AI intelligence, so moving from large language models to visual, like think of a LLM, but using visualization as opposed to just text.
Yeah, you know, the LLM market, you know, they're it's a massive undertaking the way that they are building out those models. I think, you know, where we differ a little bit is, like, they're taking massive amounts of data, and they're bootstrapping and training off of that data. We do similar things, but for us, precision is extremely important, and so, you know, going to that extra level of precision and training those particular models are very labor-intensive also. You have to do a little bit of consolidation of both.
I think the idea that a lot of the users that we operate with have to be completely disparate, separated, they can't touch the internet, that also creates a little bit of a near-term benefit for us also, whereas, you know, most of those models will require some amount of connectivity in the future.