Okay, great. Perfect. I think we're good to start. My name is Alec Brondolo. I cover small and mid-cap internet at Wells Fargo. I'm pleased to be joined by Frode Jacobsen, the CFO of Opera. Frode, thank you so much for joining us today.
Thanks for having us.
Maybe if we could start, for those that are less familiar, could you give us a brief background on the business, the consumer-facing Opera browser, and how you monetize?
Yeah, sure. Opera is a Norwegian company. That's where the headquarter is. We make web browsers. That's the most important focus of the company. Been around for 30 years. Have close to 300 million monthly users. How we make money is about driving traffic to partners. It can be to search partners, advertisers, and what else? I think we've been public since 2018, where our CAGR growth is over 20%. Over that time period, we're nicely profitable. We actually pay a dividend, which is quite unusual, I guess, especially with in excess of 5% recurring yield.
Yep. Perfect. So Perplexity has a browser called Comet. OpenAI has a browser called Atlas. Obviously, Opera has Neon. As we think about these AI-native companies, it's clear that the browser is a focus. Why do you think that is?
I think it's just the same as what we saw in the past with search, that Google creates Chrome as a way to essentially bring traffic to their platform. I think these companies are doing the same. I mean, we are excited about that move. Like Google, 25 years now, we've been happy on both sides to work together with Google as a partner. The browser has a lot of potential to drive traffic, but also sort of that broader logged-in experience. You have the context, et cetera. If you just exist on a website as ChatGPT or Claude or whatever, you have quite limited insight to the broader picture. I think the target of these new players in the platform space is to know as much as possible about the end user and to exist at a higher level than as a simple website.
Just like we've been happy to work with Google, for us, this just represents a broader spectrum of potential partners that we can work with, maybe even more competition for traffic. For us, the current situation is a much better situation to be in than the situation where there is kind of just one platform with one browser.
Yeah. I think the distribution vehicle that the browser represents, that's becoming more strategic. That's positive, right? Then on the other side, there's competition, right? There's new browsers. How do you think about differentiating Opera further as some of these AI-native companies release their own products in the space?
I think it's all very new, right? Our core focus is just we want to make the best browser. There are 30 years of history going into that, not just a product to showcase an AI platform, right? It is broader than that. I think actually our independence is a benefit because people probably do not want to use one browser for Perplexity and one for Claude and one for ChatGPT and one for Google Search. Being agnostic and independent and working with partners in the ecosystem on a non-exclusive basis, we think, is an advantage for us. That is what we are going to focus on.
Maybe with regard to foundational models, leveraging multiple models, which are you finding the most useful as you go and you build the browser technology? Do one or two of them stand out as best in kind of the browser environment, best in class?
Yeah. For now, in our own built-in assistant and agentic capabilities in our browsers, we use a combination of Gemini and ChatGPT with all the latest image, video, and all of those things. We also use the open-source models like Llama, for example. I think unlike the other ones, we actually operate across them. We're not completely locked to one. The user doesn't see it. We try to create it seamless for the user. I think on the technical level, with sort of these MCP protocols, et cetera, that we are able to let LLM essentially control the browser and an agentic experience. I think that can be quite interesting also as we look ahead because that means that if the user is, I'm using ChatGPT, for example, then with an Opera browser, we could, of course, let ChatGPT control the browser.
Like the user could almost bring their own AI to it, right? I think so, that's the technical side. Then beyond that, on the agentic side with just how to operate in the ecosystem. Like we saw Amazon suing Perplexity. That's a good example of something that kind of did not go so well. I think that's where the way we would do it would be different. It's much more important to—I mean, the agentic part of it should assist the user and help them along the way. At the same time, not shield them from the e-commerce players, for example, if they want to visit. It's good for both the browser, I think, and the broader landscape.
If the user, if somebody is invested in a platform to show alternatives, et cetera, that you have the AI part of it to simplify the process, but not have it happen in a black box that the user does not see.
Yep. How sophisticated is the technology today from a model routing perspective? I imagine that you want to use different models because there's different cost curves for different foundational models. Then each of the models has a strength, whether it's agentic capabilities or research. How sophisticated is the browser between delineating which model should be used for any given objective?
It works very seamlessly. I'm not a technical person, so I find it hard to say we're at a 7 out of 10. I think we are quite good at picking up the latest models from the partners that we work with and integrate that very quickly. In my own experience, and when I look at reviews of our agentic browser compared to other ones out there, I think we do stand up very well. I think we actually benefit from the ability to switch.
Yep. As long as I've been using web browsers, it's been a free product, right? You monetize the distribution of their search or advertising. You have a new product called Neon. I think the idea is that there's going to be incremental capabilities introduced. In exchange for that, there's going to be a monthly subscription, which is a new business model in this space. Could you maybe talk about how Neon is better, more advanced, more capable than the free browser? Kind of how are you differentiating that product and getting the customer to pay?
Yeah. Neon is the agentic sort of focus of our browser. Unlike Opera One, the flagship browser, the gaming browser, Opera GX, that are free and that have AI capabilities, the Neon browser is made for the biggest enthusiasts that appreciate the agentic capability. That has to do with if you're logged into services, your email, your Slack account, your subscriptions, that the agent can operate on your behalf, control the browser, open tabs, study them, things that are behind paywalls, summarize things, email it to your teams, put it in your calendar. You get that agentic support. Our goal is to move as much as possible of this into the free products as well. It becomes a balance between the revenue generation potential of that side versus the cost of it. Over time, probably these things will become more and more mainstream.
I think with regard to agentic capabilities broadly, as it exists today in many different modalities, sometimes it works, sometimes it doesn't, right? How far do you think we are away from having kind of a seamless agentic capability in the browser where it's going to be able to complete the transaction 10 out of 10 times, not 7 or 8 out of 10 times, which is probably where we are today?
I think much faster than we'd think. Like the pace of this. I mean, already the agentic experience is very good on broad tasks like study this topic, do research on something, or even things like look at these 20 companies and summarize their management teams. I see how the browser goes into everyone. There's a lot of copy-pasting and note-taking that's avoided when you use an agentic browser for something like that. You have the very basic tasks and the speed of that. I'm quite sure that these agentic browsers will soon order like egg and milk faster than I can do myself too. Not there yet. It's more kind of like start various processes and move on while it's working.
As we think about agentic capabilities improving, I think one component of it is the foundational models getting more sophisticated, right? Another component of it is MCP protocols advancing to a place where there is kind of a seamless layer for the agent to interact on the open web. Over the next two to three years, what do you think is going to drive more of the improvement? Is it going to be the foundational models getting smarter, or is it going to be more merchants kind of starting to adopt agentic protocols or a combination of both?
Probably a combination. I think it's very important that agentic browsers operate in a way that's constructive in the broader ecosystem. Developing more and more standards for how agents operate and how it can sort of take the users to the relevant sites and show them, I think will be important. We see on sort of the MCP layer, it is, as I mentioned before, already quite effective to let an LLM control a browser. Yeah, it doesn't seem like it's far away from being like to that 10 out of 10 level.
Yeah. Super helpful. Maybe if we shift the focus of the conversation to modernization, about one-third of revenue today comes from tech that is paid to you by Google. As user behavior shifts, the way that people search shift, how do you think about that bucket of revenue changing over time?
We think of it a bit broader, query revenue, which is historically mostly search. There is also the component of the user having a dialogue with our AI agent or is looking for something, and we serve that response directly instead of via search, which is still 5% or so of the category, but it has tripled over the past 12 months. It is growing very quickly. I think as the browser plays a more and more active role in helping the user navigate to what they are looking for, that revenue category actually has a lot of potential.
How do you think, as you observe the behavior in the browser, how much of the AI searching is incremental to the traditional search relative to cannibalistic? Is there kind of a sense for is the total pie growing, or is this more of a mixed shift? Any insights there might be helpful.
It's hard to say at that level of detail. What we have seen is that we've been able to—we are growing both our query revenue and our advertising revenue. Advertising is growing faster. Within the query, we see that there are perhaps fewer queries. I think that's also because the search engines are getting better at directly promoting. Like first, you don't click next page so often. At an eCPM growth, that's offsetting it effectively.
Yep. Maybe if we focus on the advertising business, the other roughly two-thirds of revenue, I think traditionally advertising for Opera has been display-based, CPM-based. Now that's starting to change where you're not transitioning, but the business is shifting towards e-commerce. Could you help us understand what percent of advertising revenue that e-commerce business represents today and how fast that's growing?
It was close to half in the third quarter. Probably a fairly safe bet that it will reach the milestone of representing half of advertising revenue in Q4 given the holiday shopping and all that. I think our advertising revenue has always been performance-based. We like that. We always demonstrate what we can bring in terms of traffic and value to the broader partner landscape. As you mentioned, e-commerce has grown very fast. That is something that we really started to focus on last year when we saw that we had a big enough user basis in both U.S. and European markets to make a real difference to partners. That has been growing from probably a fifth or a quarter of our advertising revenue to now a half. It has been going at over 100% year- over- year now for about a year's time.
Could you maybe help people understand shifting from the display kind of business to more of a pay-for-performance, pay-for-a-conversion? Why is that latter business model kind of maybe more attractive to you or growing faster than the traditional display advertising business?
We never really had the display advertising business. It's always been integrating partners, revenue shares, or rewards per other metrics. I'm sure it's nice to get paid no matter how effective your campaign is in a sense. The good thing about a performance-based advertising business is that, I mean, it's bound to be ROI positive for the partner. If it works well, it works well for both sides. I'm sure you can get lured into some very creative behavior if you're only paid on how much you show and not what results you drive for a partner. That would be quite short-sighted.
What do you think has been the enabling factor for the e-commerce business specifically? Has it been your scale in Western markets? Are there other changes you've made to the browser? Like what maybe you indicated several quarters ago, e-commerce was a quarter, a fifth, a quarter of the business. Now it's approaching half. What has been the enabling factor for the acceleration in kind of that part of the business?
I think it was as simple as we were of big enough size that it became interesting to partners to work with us. If you are a retailer, right, you appreciate to have multiple sources of traffic, not overly rely on one or two platforms for your promotional activities. Once we could drive traffic of meaningful scale also in key, let's say, e-commerce countries and regions, that started to open up doors. It becomes a flywheel that when we can demonstrate the effect we've had and how we're able to drive traffic to a partner well in excess of our market share in that region, it becomes a proof point that it's an effective channel.
Who are some of the key partners in the e-commerce business? Are there any brands or platforms you could call out that are meaningful?
Yeah. We focus on the big ones because we have a global business. I mean, it's Amazon, Booking, which other ones? Temu, Shein. We have some regional big ones like Lazada, Shopee that are both Southeast Asia, but also Latin to some extent. We focus on the biggest players. Then we go from starting in one market to adding another and a third and so on.
You know, I think that in, I'll say in retail media business models, whether it's Amazon's kind of e-commerce advertising business or Walmart or Target or Instacart, there's this question of as we shift to agentic commerce, how is that kind of business model going to change, right? Because if an agent is completing a transaction instead of a consumer, well, you lose the ability to place the impression, right? The person can't see the ad on the checkout page because it's an agent that's navigating there. How do you think about that transition in your e-commerce business? Is that something that the management team is thinking about?
Yeah. I think that is where it becomes very important to create an agent that saves the user time and helps them navigate to what they're looking for, but also not shut them out from seeing the alternatives, for example, right? Finding ways to, as we work with partners, that we find ways that these things can be balanced. Because it's also not good for us if our traffic is not generating value to the broader partner ecosystem. For the user also, if you ask for tennis records, as an example, like you might appreciate that one shows up on your door in two days, but maybe you'd also like to see in the screen like, do you want the basic model or this one or this one? That you have some option in there.
I think what you're saying is that by presenting the consumer with an array of options and then having the agent more focused on executing the transaction automatically, that leaves you with the ability to kind of create an impression or display an impression, display a recommendation that you can get paid for and make the brand feel good about having kind of some say in the process as opposed to being disintermediated by the agent.
Yeah. Because the agent can as a level one figure out, okay, so what's a good place for you to shop for that stuff, right? Which is a key value driver, right? You are bringing traffic to a partner. Here is a good place to look for that tennis racket. But then once you're there and you choose a racket, then maybe there's, do you have enough tennis balls? What about your tennis shoes? So you can still be exposed to that, which as a user, this tends to be a good experience, right? And then carry out the rest. And then I think the browser value contribution is bringing the user to the site and being able to make it seamless, but not taking away from the retailer the opportunity to interact with their customer.
Yep. That makes sense. How do you characterize, are there aspects of the story or your business model that you think that the market is misappreciating or misunderstanding as we go through, I guess, this transition towards agentic commerce to AI? There are aspects of the story that the market is getting wrong.
I think so. Or overlooking as a small player in the space. I always sometimes feel a bit misunderstood. I think the biggest one is we are not an LLM challenger. We are not like a VC-funded cash-burning me-too run in the LLM space. Sometimes when I read like, oh, it's tough on AI, it's like, but why does that involve us? The browser layer, we are excited about the emergence of many players in that space and new platforms that want traffic. As a number one, like our financials look totally different. We're actually.
Profitable.
Generating real revenue and profitable. Number two, like our strategic play is not to compete with ChatGPT or any other Gemini or any other model. It is to sort of exist on top of it like we have successfully done in the era of, let's say, search. Now that space is broadening, which is to our benefit and not to our detriment. I think maybe sometimes when people look at us at a glance, they could maybe misunderstand that part. I think beyond that, the quality of our partnerships. Now we are entering the 25th year with Google on the e-commerce side and the other big partners. They are all, many of them are decade-long relationships. I think you can see in our history that we must be quite good at working with other companies for mutual benefit.
I think that maybe gets a bit forgotten when we look at the broader space of players out there now that these are all, hopefully, in our view, potential partners because we think it can be mutually beneficial.
Historically, the management team has operated the business with a tremendous amount of financial discipline. I think you're going to be at a 24% EBITDA margin this year. Is there a thought to, as the industry starts to change, to maybe lean into investments, perhaps run the business with a more growth-oriented mindset? Any thoughts there about how you manage through this?
Yeah. It's the same. It's almost the same as the prior question in that. No, because we're not making those very, very billion-dollar kind of expensive kind of things. We are creating a software layer that exists on top of platforms like we've always done. Now these platforms happen to also include AI capabilities. Of course, yes, we are building the protocols and the ability for an LLM platform to control the browser and assist the user in that way. This is in terms of cost of running the business. It's the same as it's always been, let's say.
You have a new business called MiniPay. Could you maybe help the audience understand what the use case is, what the end market is?
Yeah. MiniPay is a non-custodial wallet that's made for stablecoins. It is originally, at least, focused on emerging markets, Africa, but also, let's say, family members and stuff outside the region. What it does is it makes it extremely easy for a person to take real money and put it in a USD stablecoin or a euro-based, etc., to pay and transact and to also use it for some services with completely immaterial transaction fees. That's what the product is, yeah?
How much synergy is there between the browser business and MiniPay? Is there shared customer, shared marketing? How much of an integrated kind of unit is this operating as, or is this more of a startup that sits within Opera?
It has a big benefit of coming from Opera just because we've been in these regions for decades. Of course, it has to do with money, so trust. We are known in Africa where we focus our efforts around MiniPay to be a stable, solid European company that's been their gateway to the internet for a very long time. We are trusted. It's grown to 10 million wallets, which I think is one of the fastest growing non-custodial wallets, at least, out there. Huge transaction flow and interesting. It's something we did because we saw the barriers to get into holding crypto assets or stablecoins as the most relevant one. It's very difficult, complicated plugins in your browser and all of that.
We just wanted to make it really easy because we saw that the user base there could have a big benefit of being able to hold a more stable currency.
Yeah. Could you maybe describe, I think that kind of the probably the leverage to crypto, the use of stablecoins is probably not that well understood. Could you maybe explain how stablecoins are kind of an enabling factor, an enabling asset for the MiniPay business?
I mean, the MiniPay business is a wallet. We are focusing on stablecoin because we are focusing on, let's say, the mass markets. We're just focusing on people who, yes, I'd like to keep my $10 worth of excess cash this month in dollars instead of an IRR or Kenyan shilling or another currency. The way for us, it currently represents a business is other players in the ecosystem who want to be promoted to the user, businesses who can take payment in those currencies, or sort of the broader adoption of stablecoin. We are essentially generating advertising revenue around the product, but we're not charging the users anything.
Yeah. Is this a business where you want to operate it and own it entirely into perpetuity? Could you potentially look for partners to finance the business at least partially? How are you thinking about kind of capital allocation and the ownership framework around this asset?
I mean, it's a capitalized business like our entire business. It's software that we develop. It's 100% proprietary. And it's already profitable. It's already generating profits for us. We wouldn't need investment as a financing mechanism. We are on our own also a well-capitalized company given our size. That would be more, there could of course be strategic opportunities with industry partners as a way to accelerate the adoption. It would be more in that direction, if anything, than like a pure kind of financing thing.
Maybe I could ask about OPay. You own 9.4% of OPay. It's valued on the balance sheet at $258 million. Do you want to keep the OPay stake or are you open to monetizing it over time? An update there would be helpful.
Yeah. On the topics of overlooked things, this probably also makes the list because we founded, Opera founded OPay back in 2017. It stood for Opera Pay at the time, right? Very different from MiniPay in that it actually is a very physical, present business. It is the biggest mobile wallet in Nigeria. It has hundreds of thousands of essentially human agents for people to move cash in and out. It is transacting in the hundreds of billions of dollars every year of money. It has been a super successful business. We played an active role in the beginning. Now the company is operating completely autonomously and independently from Opera. I think we have been commenting on sort of a natural progression for OPay to go public itself.
Sort of following that, in connection with that, over time, Opera will realize our hundreds of millions of dollars of gains from that and not be a passive shareholder over time.
The $258 million, the mark that the asset is sitting on the balance sheet at today, when was that marked? What valuation round? How is that? Is that recent? Was that several years in the past? How updated is the valuation that $258 million is based on?
We do an annual valuation every year. We try to be very cautious in this. Just we follow. What we saw in the past was that every time there was a funding round, we had to, in percent, quite massively raise our valuations. I think the last time that happened was in 2023 where we marked it up by $100 million or so. Of course, once the company goes public, we hope to repeat that exercise.
That'd be good.
But yeah.
Yeah. Yeah. Uses of cash going forward. You're going to get some proceeds whenever OPay comes public. Obviously, the business is cash-generative today. How should investors be thinking about M&A relative to share repurchases, dividends, and capital returns?
Yeah. I think since 2020, we've returned about $500 million to our shareholders, which is a meaningful amount relative to our market cap. I think all the cash we have is sort of, it's the shareholders' cash in a way. We are running a profitable business. We're not burning any of it. We've used both buybacks and dividends, special and recurring, which is the ongoing one now. M&A is, we bought some companies along the way, but it's been more for team acquisition or a particular license rather than.
Partners. Yeah.
I think the only company we ever bought that had some revenue, it was like $2 million a year. We bought the technology and we made it free. Our revenue growth has always been organic. It's not off the table, but we don't have a tradition to be very active in terms of P&L material M&A.
Perfect. I think I'm going to leave the conversation there, Frode. We enormously appreciate you taking the time.
Yeah. No. Thanks again.
Yeah. Perfect. Thanks.