Zedge, Inc. LD Micro 20th Annual Main Event Conference

Welcome. Thank you everyone. Thanks for joining.

I am sure the room will fill in more, not because of me, but because of the weather conditions outside. I hope that I make good use of your time. I am here to give you an update about Zedge, and our expansion of the business into the exploding world of AI as a data and services provider to core AI developers. Of course, everything that I discuss today is subject to our Safe Harbor statement. To start with, DataSeeds, our new initiative, is our core growth initiative at this point in time. What have we done? We come from the world of consumer mobile app publishing. We have got two super popular apps that are focused on creator communities and content. What we have done is we have taken that knowledge, and we have reapplied it to make ourselves relevant to AI model builders.

As we all know, any AI model depends on the data. Good data in, and then if you have got a good model, you will even get better data out. Bad data in, you are going to get even worse data out. That is where we are really making a difference. Our controlling shareholder, the Jonas family, Howard Jonas from IDT Corporation, just completed a private placement, capitalizing another 7.7 million or infusing in the company with another $7.7 million. We have got, give or take on a pro forma basis, close to $30 million on our balance sheet. Our core consumer offerings are profitable, and we continue to be dedicated to seeing to it that we grow those offerings accordingly. We are going to use some of the capital in order to bulk up on our talent, sales, marketing, and engineering. Some quick KPIs. Give or take, our creator community has around 20 million monthly active users.

We have got a subscription layer in the Zedge Marketplace app that has grown by around 41% Q3 over Q3. As a side point, we are going to have our fiscal 2026 annual earnings call taking place this coming Monday. Please tune in. We are excited to share the numbers with you. We have got, give or take, a $16 million enterprise value on a pro forma basis. Like I said, give or take, around $30 million of cash on a pro forma basis, without any debt on our balance sheet. We have got, give or take, a 20% cash flow yield. What are we doing? We are really focusing on scaling DataSeeds and seeing to it that we become a core provider to all of these AI developers with respect to data and data services that they need in order to scale accordingly.

Let's talk a little bit about the market. Grand View Research estimates that by 2033, the market for data will be, give or take, around $16 billion of revenue on an annual basis. We have built our solution so that it is purpose-built, and it mirrors what the AI developers or the robotics developers need with respect to data needs. We are able to do this by tapping into our community. Then finally, we announced this past Monday that we are doing model evaluations. Where we are differentiating ourselves is by taking advantage of our experts in the world of photography in order to provide expert evaluations.

As models continue to develop, the need to have humans in the loop that have deep expertise in understanding what they're looking at, as opposed to just regular rank and file people, in order to critique the model becomes critical. We have that. It's an unfair competitive advantage. As I'd mentioned, we're all purpose-built. We offer multimodal data. That's audio, that's video, that's imagery, even synthetic data, as well as data that is needed by robotics manufacturers in order to train their robots, known as egocentric or exocentric video.

It's not simply the video, but it's the video plus all of the sensor data that is needed to give the robot the smarts that it needs to make sure that when it is putting away the dishes, it doesn't throw the dish into the closet at too high of a velocity or hold onto the dish and crush it because it is holding too tightly. As part of this, we are also tapping into our community of creators in order to undertake evaluations. Coming at this from the vantage point of really understanding what they're looking at and offering a wide array of different variables by which they can critique models in a fashion which provides model builders with the competitive smarts that they need in order to further refine and perfect their models. We have several different ingredients in our ecosystem that feed into DataSeeds.

As you all know, we've got the Zedge Marketplace, which is the leading marketplace for mobile personalization, content, ringtones, wallpapers, app icons, battery charging icons, dynamic wallpapers, you name it, as well as a full gen AI Suite, that app has been downloaded close to 800 million times, primarily through organic channels. That marketplace has creators that are very interested in participating and helping create content for any technology company that's in the need of data. We also have GuruShots, the leading photo competition game available in the marketplace. We acquired GuruShots around five years ago. At the point that we acquired GuruShots, it came with a catalog of give or take 150 million photographs. We've been able to secure rights for approximately 50 million of that 150 million for AI training purposes.

In addition to that, we have a very vibrant community of expert photo enthusiasts that have outstanding trade craft and perception, and that is really fueling us on the eval side. We have an innovation team that we have been investing in, and their focus is going to shift more heavily into how can we provide new innovative solutions to fuel DataSeeds. Content and creators. Obviously, we have access to the 20 million in our community. We know very well how to work with creators, how to see to it that they are incentivized and can deliver mission-focused content solutions based upon the needs that we have, and benefit from that, either in terms of receiving virtual currency or hard payouts. Then we have the experience on mobile and the trust. Think about it. When we go out and bring creators onto the platform, we can do the user acquisition, we can do the onboarding, we can do the consent.

Consent is really important. As many of you know, many of the major AI companies have either lawsuits that are pending or that they have settled on because they scraped the internet. They acquired content that they were not supposed to acquire through questionable means. We, on the other hand, are creating content on a bespoke basis, and as part of that process, the content creators are providing consent, and that is auditable, such that if someone down the road were to say, "Where did you get this data from?" We can then go back chapter and verse and say, "Here is the creator, and here is what they consented when they uploaded that content." All of those are being used and repurposed to fuel DataSeeds.

What happens on the business side? We are speaking to the frontier model labs, the leading technology companies, major robotics manufacturers, as well as a whole slew of startups in the technology space that are focused on AI. We reach out to them, we explain our offering. I think I will position it as the following. I was meeting with the CEO of a multi-billion dollar AI player around 2 months ago. Part of their business is in creating datasets, and the CEO had said to me, "Jonathan, what you guys are doing is elegant.

We have to do it with brute force, but you have cracked the code in terms of how to do this and see to it that it is a very organic and natural process for the creators, without having to necessarily hire them on a full-time basis." We reach out to a prospective customer, they send us a very detailed brief. We review that brief, and there are all sorts of details in there in terms of what is the content that they need, how they need it created, what sort of metadata they need, and so on and so forth. After that, we can then begin to source. How are we going to produce that? We have 3 different ways of fulfilling that need. We have an off-the-shelf catalog, whether it be of imagery, audio, or video. That is one layer. We can then simply search through that catalog.

If we have the data, we'll sell it, or we will license it to the prospective customer. The second is what we call managed crowd, and this is really what our consumer user base is able to participate in. That is that we will go out and we will ask them, "Hey, you want to create some data? We can either pay you in virtual currency that you can use in our apps, or we will pay you hard dollars in order to create this data according to spec." On the back end, we have a whole set of technology to ensure that the quality is there and that we are going to ultimately deliver to the end customer what they need. They don't want to come back to us and say, "Hey, 50% of the data wasn't good." We don't want that.