Opera (OPRA) 2026 年第二季財報電話會議：營收成長 25% 與調升全年財測指引

重點總覽

• 2026 財年第二季營收年增 25% 至 1.781 億美元，高於 Opera 財測指引區間上限。廣告營收成長 27% 至 1.15 億美元，搜尋查詢營收則成長 21% 至 6,200 萬美元。
• 調整後 EBITDA 創單季新高達 4,240 萬美元，年增 32%，利潤率為 24%。調整後淨利成長 27% 至 3,000 萬美元，調整後稀釋每股盈餘 (EPS) 成長 25% 至 0.33 美元。
• Opera 上調全年展望，預期營收為 7.34 億至 7.42 億美元，調整後 EBITDA 為 1.72 億至 1.75 億美元。以中位數計算，相當於營收成長約 20%，調整後 EBITDA 利潤率為 24%。
• 每月活躍用戶數 (MAU) 總計達 2.88 億。西方市場 MAU 成長 4% 至 6,100 萬，其中行動端成長 8%；年化每用戶平均收入 (ARPU) 成長 25% 至 2.46 美元。
• 管理層表示，西方市場的 AI 用戶在 Opera 上停留的時間更長且進行更多搜尋。西方市場搜尋查詢營收成長 29%，而全球成長率為 21%。
• 在第一方庫存、OEM 白牌解決方案、內容合作夥伴關係以及 Opera 應用程式更廣泛的 SDK 觸及率支援下，Opera 的可服務受眾突破 7 億，高於六個月前的 5 億多。

重要財務數據

Opera 在第二季以每股 17.44 美元的價格回購 636,000 股，總金額 1,110 萬美元。截至 6 月 30 日，流通在外股數降至 8,890 萬股。包含 7 月發放的股利在內，該公司表示自 2020 年以來已透過股利與回購向股東發還 5.77 億美元。

業務與營運表現

廣告成長由電子商務領軍。Opera 的自有電商平台涵蓋超過 100 家商家和 1 億多種商品。管理層正在開發 AI 支援的比價與其他高購買意圖廣告格式，旨在於購物者更接近購買決策時將其與商家相連。

Opera 也在擴展旅遊領域，並已在其自有用戶群之外展開初步的推廣活動。管理層表示，第三方庫存拓展了 Opera Ads 的觸及範圍，且除了庫存成本外，所需的額外營運費用非常有限。

公司的 AI 策略核心是保持瀏覽器作為開放上下文與執行層。Browser Connector 允許 Anthropic 的 Claude 和 OpenAI 的 ChatGPT 等服務理解即時瀏覽上下文，並代表用戶與 Opera 互動。Opera 還推出了 Opera Browser AI，這是一個開源命令列介面，用於將即時瀏覽器活動與 AI 程式碼代理和自動化工具連接。

管理層報告指出，前沿 AI 服務、開源模型和本地託管模型的偏好使用量持續成長。管理層指出，Opera 的獨立定位使其能夠支援多個提供者，而非依賴單一 AI 生態系統。

Opera GX 本季新增近 200 萬用戶，MAU 達到 3,700 萬。桌上型電腦與行動端均實現成長，並獲得 AI 整合以及提供遊戲內獎勵與增益的遊戲合作夥伴關係支撐。

行動端動能以西方市場最為強勁。過去一年中，英國的 Android 和 iOS 合計 MAU 成長 66%，美國成長 40%。iOS 版 Opera One 在全歐洲成長 42%，儘管 Opera 約 90% 的智慧型手機用戶群仍使用 Android。

自上次更新以來，MiniPay 新增 300 萬個錢包啟用數和 8,800 萬筆交易，達到 1,800 萬個錢包和 5.18 億筆交易。這款自管穩定幣錢包在 66 個以上的國家營運，並包含超過 57 個上線的迷你應用程式。Opera 於 6 月在整個歐盟推出 Visa 卡，並正在非洲、拉丁美洲和亞洲的受支援市場逐步推行。

管理層財測指引

管理層預計現金薪酬費用與行銷費用合計占比將從 2025 年占營收的 36% 降至 2026 年的約 32%。在營運槓桿以及 Opera Ads 營收帶來的額外營運費用有限之支援下，Opera 預計其調整後 EBITDA 利潤率將比 2025 年提升 25 至 40 個基點。

風險與關注焦點

• 管理層預計現金轉化率將出現單季波動，但表示上半年的轉化比率可能會在下半年趨於穩定並有所上升。
• 隨著第三方廣告庫存擴張及季節性廣告活動增加，全年營收成本預計將升至營收的 39% 左右。
• Opera GX 通常在 7 月和 8 月使用量較低，因為用戶在暑假期間花在電腦前的時間較少。
• MiniPay 仍處於早期階段。管理層表示，需要進一步擴大規模、基礎設施與合作夥伴連結，基於交易的變現效益才會更加顯著。
• Opera 預計 OPay 最終將推動 IPO，但管理層無法對時間表發表評論。

分析師問答環節亮點

管理層將 Opera 可服務受眾擴展突破 7 億歸因於新的合作夥伴關係，包括 AI 生成影片和 AI 社群服務。Opera 的目標是在其更廣泛的可服務網路中達到 10 億用戶，但管理層強調該努力仍處於早期階段。

關於搜尋查詢經濟效益，財務長 Frode Jacobsen 表示，成長主要由單次搜尋價值提升所帶動。部分領域（特別是智慧型手機）的每用戶搜尋頻率也在增加。非搜尋查詢營收年增超過 200%，但規模仍維持在數百萬美元的單位數水準。

關於 Browser Connector，管理層表示透過第三方 AI 服務發起的活動仍在 Opera 的瀏覽器環境中進行。當用戶透過自己的訂閱存取服務時，Opera 無需承擔 AI 模型的算力成本，而基於瀏覽器的商業協議仍可適用於產生的活動。

關於廣告利潤率，管理層表示第三方 Opera Ads 庫存的毛利率正在改善。儘管如此，整體毛利率反映了變動中的營收結構，而有限的額外營運費用使 Opera 能夠設定更高的調整後 EBITDA 利潤率目標。

管理層形容 MiniPay 自推出以來即實現獲利，但表示該平台目前仍以規模與網路發展為優先。一旦生態系統達到足夠規模，未來潛在的變現來源可能來自交易量與抽成率 (take rate)。

電話會議完整逐字稿

完整財報電話會議逐字稿

管理層陳述

Operator

Welcome to the Opera Limited Second Quarter 2026 Earnings Call. [Operator Instructions] Please be advised that today's call is being recorded. [Operator Instructions] I would now like to turn the call over to your speaker today, Matt Wolfson, Head of Investor Relations. Please go ahead.

Matthew Wolfson

Thank you, Erica, and thank you, everyone, for joining us this morning. I'm joined by our CEO, Song Lin; and our CFO, Frode Jacobsen. Before I hand over the call to Song Lin, I would like to remind you that some of the statements that we make today regarding our business, operations and financial performance may be considered forward-looking. Such statements are based on current expectations and assumptions that are subject to a number of risks and uncertainties.

Actual results could differ materially as a result of various factors, including those set forth in today's earnings press release and in our most recent annual report on Form 20-F filed with the SEC. We undertake no obligation to update any forward-looking statement. During this call, we will present both IFRS and non-IFRS financial measures. A reconciliation of IFRS to non-IFRS measures is included in today's earnings press release. The earnings press release and an accompanying investor presentation are available on our Investor Relations website at investor.opera.com. Our comments will be on year-over-year comparisons unless we state otherwise.

With that, let me turn the call over to our CEO, Song Lin, who will cover our second quarter operational highlights and strategy, and then to Frode Jacobsen, who will discuss the details of our financials and expectations for the third quarter and full year. Song?

Lin Song

Thank you, and good morning, everyone. We have been looking forward to sharing today's report with you. Our second quarter results reaffirm that being an independent, well-established and innovative browser provider with a user base of nearly 300 million people and a significant advertising reach is a very attractive position to hold in a rapidly evolving and expanding ecosystem. Instead of placing bets on which AI services or infrastructure plays will be leading in the future, we cultivate our position as a tech enabler and platform that facilitates choice for the end user and access to our vast user base for our partners.

In this landscape, the browser is becoming more valuable as AI changes how people search work and act online and Opera is already translating that shift into greater engagement and monetization. By continuing to give the most demanding users new reasons to switch from the operating system default browser and by expanding the functionality of our advertising platform, we broaden our partner ecosystem and grow monetization opportunities every month, adding to the foundation of our long-term trajectory as well. Both revenue and adjusted EBITDA exceeded the top of our guidance range, with growth fueled by acceleration of revenue growth from 23% year-over-year in Q1 to 25% year-over-year in Q2.

The strength was broad-based. Advertising revenue grew 27% to $115 million, while Query revenue grew 21% to $62 million. With that, second quarter revenue exceeded $178 million and surpassed the seasonal Q4 peak a quarter earlier than what we've seen in most prior years. Adjusted EBITDA was also a quarterly record at $42.4 million, representing a margin of 24% and growing 32% year-over-year. And importantly, our confidence in this elevated trajectory allows us to raise full year guidance beyond the Q2 overperformance, which Frode will get back to. Advertising growth was again led by e-commerce in particular. As we look ahead, our road map includes additional high-intent formats, including AI-supported price comparisons designed to help shoppers evaluate the products while helping merchants reach users closer to a purchase decision.

Our in-house commerce platform already helps match our users with the best deals across 100 merchants with over 100 million products. Within travel, another high potential vertical for us, we work with the top online travel agencies and have started initial campaigns beyond Opera's own user base. Our partners continue to expand their work with us because our performance-based campaigns deliver measurable outcomes. As a combined platform for first and third-party inventory, we are able to inform and allocate campaigns with a solid understanding of the relevant audiences. In fact, our total addressable audience when taking into account the millions of users that access our content platform through OEM white-label solutions and the broader SDK ratio of Opera apps has now reached beyond 700 million, up from the 500 million we announced just 6 months ago. This scale and growth reinforces our position among the largest online platforms. Our query revenue representing the monetization of our users' proactive intent continue to grow ahead of underlying search market benchmarks as we benefit from natively integrating key partners as a part of the browser interface. This revenue category directly captures the traffic monetization potential of increased engagement in our browsers with native AI functionalities, benefiting both time spent and the browser's ability to connect the right partners with our users at the right time.

This is also true as it relates to the evolution of our search partnerships where the secular tailwind from longer and more complex user journeys continues to build. Search is evolving from short keywords to questions and increasingly to conversations. We combine this with rapid product innovation, driving both user appreciation and increased usage of our browsers, all of which results in more engagement within the address bar and omnibox with more opportunities to connect high-intent users with relevant results. Look at the ARPU-driven Western markets, as mentioned before, we see that the users who engage with AI within our browsers spend significantly more time in the browser and even conduct many more searches versus comparable users who are not yet engaged in AI, all of which directly contributes to ARPU growth. As an overall result, we see that query revenue is growing at 1.4x the pace in Western markets, this is the global average, up 29% year-over-year as opposed to 21% globally. It's a point worth making that Opera is already monetizing AI-driven query activity today, not simply describing a future opportunity.

In the second quarter, Google announced a new commercialization of its AI mode, widening the basis for our query revenue stream. We also expanded our AI strategy by announcing Browser Connector for leading AI services, including Anthropic’s Claude and OpenAI’s ChatGPT. Browser Connector enables users to securely connect these AI services to the Opera browser, elevating those services to become agentic by understanding the live browsing context and interacting with the browser on the user's behalf. This represents a shift from closed single vendor AI experiences toward an open ecosystem where users can choose AI services that best meet their needs while retaining the browser as a central interface.

We are fully committed to such interoperability as the best basis for growth of new AI platforms and services, allowing the users to have a deeply integrated experience without juggling multiple browsers and enabling new platforms to access the users in a native way without having to drive both adoption of the platform and a dedicated and perhaps narrow browser experience on top. This open approach aligns with Opera's position as an independent browser vendor and appeals to our most technologically sophisticated user base, many of whom value flexibility and avoiding dependency on a single AI provider. As AI assistant become increasingly capable, the browser is well positioned to serve as a trusted context and execution layer connecting users with multiple AI services, and we expect adoption of such integrations to eventually be commonplace for all users.

Opera also introduced Opera Browser AI and open source command-line interface that enables developers and AI enthusiasts to integrate the browser directly into AI-driven workflows. By allowing AI coding agents and automation tools to interact with a live browser, Opera Browser AI extends the browser's role beyond traditional browsing and reinforces Opera's strategy of making the browser programmable infrastructure for the next generation of AI applications.

Turning to our user base. Opera had 188 million (sic) [ 288 million ] monthly active users in the quarter. Our Western user base grew 4% year-over-year to 61 million with both desktop and mobile platforms contributing. Mobile was particularly strong, growing 8% year-over-year across Western markets, while the [ low ARPU Asia ] home base continues to phase out. This continued mix shift towards higher-value users helped increase annualized ARPU by 25% to $2.46. Opera GX reached 37 million monthly active users, adding almost 2 million users during the quarter. Both desktop and mobile grew with a larger absolute contribution coming from desktop. Partnerships with games like Forsaken, rewards players with in-game items like free skins and game boost, which resonates with our target audience.

Our momentum is especially visible in some of the world's most competitive mobile markets. Over the past year, combined Android and iOS MAUs grew 66% in the United Kingdom and 40% in the United States. Across Europe, Opera One for iOS grew 42%, demonstrating our potential to broaden our smartphone base, which is still about 90% Android. Users continue to choose Opera for differentiated features, including our free no-log VPN, intuitive tab management and building browser AI. Our iOS user base growth shows how even a highly restrictive ecosystem has materialized and the opportunity for us to grow both users and overall ARPU.

Our browser reach and brand trust also enables us to scale new services. MiniPay, our self-custodial stablecoin wallet solved the problem of access to international currency for users in emerging markets, removing complexities for the end user and making P2P transfers and Web3 access very easy. Given our ability to rapidly scale, we are also able to work closely with key partners such as Celo and Tether to drive adoption of these services. MiniPay's growth continued in the second quarter with 3 million new wallet activations and 88 million transactions since our last update, bringing the total to 18 million wallets and 518 million transactions. MiniPay now reaches more than 66 countries and includes more than 57 live mini apps and is rapidly expanding its capabilities.

In June, we launched a card in collaboration with Visa that bridges the gap between stablecoin holding and daily spending. The card is now available across the EU and is being gradually introduced in supporting markets in Africa, Latin America and Asia. Stablecoin access has different use cases in different economies, but as a global yet locally integrated network, we remove friction from international money transfers while both parties are unbanked, and our users can travel globally like true locals in markets where mobile payment options are expanding. All of this strengthen our conviction that MiniPay can make stablecoins useful for everyday savings transfers and spending. It is still early, but the product scale, utility and ecosystem participation continue to progress rapidly.

With that, I would like to turn the call over to Frode Jacobsen, our CFO, to discuss our financial results, guidance and capital allocation in greater detail. Frode?

Frode Jacobsen

Thanks, Song. We remain very pleased with how our strategy and business performance converts to healthy financials, having exceeded our guidance ranges in both quarters of 2026 to date. As Song Lin pointed out, being an independent partner-oriented and tech-first browser in the evolving AI landscape is a great position that we will continue to cultivate. And yet again, we are able to lift our full year guidance, reflecting both the Q2 overperformance and our trajectory as we enter the second half of the year.

Opera's growth is all organic and comes with healthy profitability. We invest in our growth through rapid and continuous product development, creative and engaging marketing and by acquiring the third-party inventories as we scale our ads business. Those who have followed us over time know that we balance this carefully to ensure a strong growth trajectory while also driving profits and cash generation that we return to our shareholders through our recurring dividend and share buybacks. Last quarter, I talked about how our 10-year average annual revenue growth stands at 21%. And in fact, if you zoom into the post-COVID period and look at the CAGR across the last 4 full years, the average annual revenue growth has been 23%, and profit metrics have grown even faster.

At EPS level, our share buybacks have amplified that trend with average annual adjusted EPS growth of 33%, which excludes valuation gains from our stake in OPay that we eliminate from our adjusted metrics. Zooming back into our Q2 results, revenue grew 25% to $178.1 million. Overall expenses came in according to expectations, resulting in adjusted EBITDA growth of 32% to $42.4 million or a margin of 24%. In terms of cost categories, cost of revenue items combined came in at 38% of revenue, exactly as previously indicated. Marketing spend came in at $36.2 million, representing a sequential decline of 6% relative to Q1 with continued discipline. Cash-based compensation was $23.1 million, which included accelerated annual bonus accruals following the strong underlying performance in the quarter. The sum of all the smaller OpEx items, pre-adjusted EBITDA came in at $8.9 million and was overall flat versus Q1. Below the EBITDA line, we achieved adjusted net income of $30 million or 27% growth year-over-year and adjusted diluted EPS was $0.33, representing 25% growth. Operating cash flow was $22 million in the quarter with free cash flow from operations of $17 million. Year-to-date, we have converted 76% of adjusted EBITDA to operating cash flow and 62% of adjusted EBITDA to free cash flow from operations, both ratios nearly the same as in the first half of 2025.

While we continue to expect fluctuations in cash conversion between quarters, the year-to-date ratios will stabilize and likely tick up in the second half of the year as they also did in 2025. In terms of capital allocation, our low CapEx business model allows us to return significant value to our shareholders through our recurring dividend and share buyback program. In fact, since 2020 and including our recent July dividend, we have returned $577 million to our shareholders, of which $320 million through dividends and $256 million spent to buy back a total of 37.2 million shares of Opera with an average cost per share of $6.88 and representing 31% of shares outstanding at the start of 2020. Our July semiannual dividend of $0.40 per share or $35.6 million total represented an annualized yield of 3.9% on the record date.

During Q2, we repurchased 636,000 shares for a total spend of $11.1 million pro rata distributed between public buybacks and repurchases from our majority shareholder at the same price per share, which was $17.44. This corresponded to 0.7% of shares outstanding at the start of the quarter and reduced the total number of shares outstanding as of 30th of June to 88.9 million. You'll see $14.2 million of buyback spend in our Q2 cash flow, which includes $4.1 million of Q1 repurchases that settled in Q2 and excludes $1 million of Q2 purchases that will settle in Q3. Now turning to guidance. As we revise our full year ranges, we combine the Q2 beat on both revenue and adjusted EBITDA with additional upside in the second half of the year, in line with how we also raised guidance at this time last year.

So while we build in a more normalized Q4 spike than the extraordinary year-end growth spikes we saw in 2024 and ultimately also in 2025, we also reflect our most recent momentum. For the full year, we guide revenue of $734 million to $742 million or 20% growth at the midpoint, adding $2 million to $5 million in addition to the Q2 overperformance. We guide adjusted EBITDA of $172 million to $175 million, representing a 24% margin on the elevated revenue midpoint. For the third quarter, we guide revenue of $181 million to $183 million or 19% to 20% growth. We guide adjusted EBITDA of $41 million to $43 million, representing a 23% margin at the midpoint. In terms of costs, we then implicitly guide to a full year OpEx base pre-adjusted EBITDA of $565 million at the midpoint, of which $140 million in Q3. At the new midpoint, we expect cost of revenue items combined to represent about 39% of revenue for the year and the quarterly percentages ticking up with seasonality in advertising.

Marketing cost is expected to remain relatively stable around the Q2 level, resulting in mid-single-digit annual growth and representing about 20% of full year revenue. Cash-based compensation expense is expected to modestly reduce relative to Q2 with annual growth in the low double digits and representing about 12% of full year revenue. The sum of all other OpEx items, pre-adjusted EBITDA is expected to remain stable at about 5% of revenue. In sum, cash-based compensation and marketing will then decline from representing 36% of revenue last year to representing about 32% of revenue this year, supported by economies of scale and the inflow of revenue from Opera Ads that carries cost of revenue but limited incremental OpEx. This enables us to guide to an increase in adjusted EBITDA margin relative to 2025 of 25 to 40 basis points. Taken together, we are very pleased with the second quarter and our momentum and opportunity as we enter the second half of the year. We have continued returning capital to our shareholders while investing in the product and commercial opportunities that can fuel Opera's growth well into the future.

With that, I'll turn the call back over to the operator for your questions.

Operator

[Operator Instructions] We'll take our first question from Naved Khan with B. Riley Securities.

分析師問答

Naved Khan

Great. I have a couple of questions. So maybe just on this audience number you gave, you said you have more than -- a reach of more than 700 million. which is up from 500 million plus that you had 6 months ago. Did you sign any new partners to drive this kind of reach? Can you maybe talk about that a little bit?

And then on a related question -- related sort of note, the 100 or so advertisers that you have with 100 million or so listings, item listings, how does that compare with the last quarter and the year ago period?

And then maybe finally on Opera -- sorry, OPay, any -- can you give us any sense of timing on when that might happen in terms of going public? Is it this year or next year? Just give us some thoughts there.

Lin Song

Well, yes, so it's Song Lin. I think I'll try to answer the first 2 questions and Frode can also address a bit OPay for whatever he can comment about. So yes. So for the reach, yes, yes. We have actually expanded quite a lot of new partners in the field on the back of actually our strength of Opera Ads and also with the fact that with the help of AI and algorithm, we were able to bring a lot of demand and also make it much easier for our partners to work with us because we can also help them monetize. So I would almost say like the broader new partnerships that we see coming, actually, many of them are very encouraging because they are -- like many of them are very new AI services that they see a benefit of combined with our strengths.

With those -- it could be in the field of AI-generated videos. It could be in the field of AI social and many others. And it's actually enhancement on both sides that they are very happy, but it also allow us also to reach advertisers, which are very keen on those audiences. So I think that we're actually very pleased and it's almost ahead of what we project. So reasonably happy about it. But again, it's still in early stage, right? Because I think our goal is just to reach billion, hopefully, ASAP, and then we should be the top-tier players in the field. And that's our goal. And then -- and also maybe also to briefly comment a bit about your question about 100 merchant and 100 million products. So just to be specific, that's actually particularly designed to power our AI services, right? So almost better if you imagine that as a way to show that with the help of AI for whatever previously may be only available if you do it from a search, now those are also available that we can directly pop up under the context.

And it's also very relevant because it's directly combining like relevant context with a particular product with the right price and with the right information, so which is actually only possible with the help of AI. And it's rather new. So I would say it's almost no -- it's not really a comparison in the past because in the past, we are not really doing this because of many limitations. But now with AI, we can. That with AI is actually possible for us to -- under the context of whatever user is browsing or solving to try to give him as accurate information as possible. So view this as the future approach where we try to give you the relevant information and hopefully also be able to commercialize it in the right approach and in connection with many of our partners. So it's actually -- it's still early days, but it's a very important initiative from our side.

So yes, as a summary, I think both of the 2 questions are actually relevant with our fast growth on AI. So the first one is actually we are able to work actually added many interesting partners on AI field, which has a very good positive loop on both sides. And the second one actually allow us to provide AI relevant commercial content, e-commerce content in the right context and pave a potential monetization base in the future. So quite excited. And with that, I think Frode can also help address the last question.

Frode Jacobsen

Yes. In terms of OPay and question around an IPO, we continue to expect that OPay will ultimately go public. We are very impressed with what OPay has achieved. And at Opera, we're also proud to have been part of its founding. As a shareholder, we will welcome an IPO. It will lead to an immediate transparency as to the value of our founding stake in the company, but I can't really comment on timing. That will be more up to the OPay team to judge.

Operator

And we'll take our next question from the line of Eric Sheridan with Goldman Sachs.

Unknown Analyst

This is Alex on for Eric. I wanted to dig into the strength you saw in the quarter of mobile MAUs in U.S. and Europe. Can you talk about some of the key catalysts that have driven this growth recently? Is it just broader adoption of Chrome and Safari alternatives post the regulatory environment? Are there any active investments you're making in the regions to drive this growth? And any differences in AI adoption and consumer behavior within AI that you're seeing across the 2 regions would be helpful.

Lin Song

Yes. So yes, I think I'll comment a bit on it. So yes, so in general, I think it's a bit of both, right? So in Europe, it's definitely because of the actions opening up. We do -- we always see from even last year on, we see a continued trend that users become aware, especially on iOS that there are alternative browsers and where we keep a nice growing trajectory. And this actually has been further helped by the advance of AI as, again, right, AI becoming much more visible to everybody that there is alternatives even on the open like iOS that you can choose as a browser player, right? So that we definitely see a very, very nice growth trajectory as reported, both in Europe, but also see the same trend happening in U.S. So very, very exciting about it.

And we -- yes, so like we have a big hope on how that will continue to grow further. And then -- and then on top, we think that it's -- I would also say that it's also the trend that we see that it's a self-reinforcing loop that we also feel very encouraging that users come to the platform come to Opera browser instead of system default browser. Usually those systems are mobile, typically OS because of AI. But then what we also see is that the moment they use AI, they actually spend much longer time and even in traditional search and also they have much more engagement compared with those which do not come from AI, but from some other regular cases. So this actually overall create almost a positive feedback loop that they come to Opera for the AI and the more they use it, the more they actually engage with it, which actually makes it very encouraging.

So I think that's also why we will probably likely continue to double down this by providing a better products for the end user and hopefully, also will nicely see a growth trajectory on those platforms.

Operator

And we'll take our next question from Ron Josey with Citi.

Ronald Josey

Song, you mentioned earlier just about greater engagement from users who engage with AI versus those who don't. I want to hear a little bit more from you on just the adoption, what tools those users are using within the browser, the insights of users who have adopted those tools, meaning who are they versus those that have not? And specifically, are they Western users? And just more insights on the plans to drive greater adoption of the AI tools given the impact of the shift, I think, toward an open AI ecosystem, which you talked about in the letter.

And then just as a follow-up on Agentic commerce and the 100-plus merchants and 100 million, I think, products. Just talk to us a little bit more how Opera's positioning here as Agentic commerce evolves into a bigger part of everyone's shopping experience.

Lin Song

Yes, sure. I think I'll try to get some answers there. So I think, first of all, we like -- at least internally, we have also done some stats and we have some -- as a browser, I think we are lucky that we are in a position to be able to have firsthand information of what user behave in those environments, right? So I think one thing we can definitely confirm is that the use of AI is definitely moving forward at a fast pace, right? So like both, I would say, in terms of using the popular service like the big players that we usually see, both from Google Gemini, but also from Anthropic and also from ChatGPT, -- most of them actually use it by visiting the web as we comment, and we can definitely see that both of the users growth largely in the last quarter and both year-over-year, but also very visible quarter-over-quarter.

So that's definitely something which we can confirm, right? But then maybe I'll also comment a bit in a broader level, right? So there are also some very interesting other trends that we see during the Q2. So number one is that we also clearly see a pattern that on top of us using those big online service players, we also see an interesting trend that we saw an increased use of many open source services, right, that is out there. So that is very visible. And we believe that -- and for the combination of user behaviors, we believe that there seems to be a sense that us are now becoming to very -- like, let's say, if previously they only use one chat or whatever as a whole of their use of AI. Now it seems that the behavior becoming -- they use chat for something, they use even different chat for different things. Imagine one may be for company work, the other may be for private work.

But then we also see that they now also started to use many other, let's say, alternative open source models as a potential tools in whatever context, right, that -- so that is actually, I think, quite interesting that we see in Q2. We believe that's partly just because of -- it's an illustration of diversification that's probably a function of more and more of many of the agentic functionality, whatever are better served by some open source ones either because of because they are becoming much cheaper and token price becoming much more cautious among the users. But potentially, I think also because of some of the agentic framework or whatever are better supported by open source, we believe maybe also by intention that because many of those guys do not want all their work whatever access to be available of one big single big AI player, maybe in the fair that those guys will take their share or whatever, right? So that's one thing which we see quite interesting.

I would also say that the other thing which we see on the browser end is also that there seems to be also now more and more opening of both online model, but also local models, which also become very interesting. So like user will use big frontier models for the really top of note stuff. But we also see that many of them actually prefer to use local models in many other ways, like voice input and a few others. And again, primarily probably both for the concern of privacy, but also for potentially token price considerations because local model, of course, doesn't cost anything. It just cost electricity on your laptop. And this is even more obvious when it comes to the latest Mac machines and a few others, which are all well supported out those. So by and large, we think these are actually quite relevant and interesting, and that's actually partly why we define our strategy because we think both of them are very positive to Opera number one, being an independent player, we are a very natural place to serve all of those big frontier models, allow people to access it, which we see already happening.

But we are also very not biased towards any other open source models among others. And we are also we access all the browser, which do support the local models, which are hosted on local machines in combination with other technologies. So we think all of those are very interesting directions and prompt us to move further in this area of being the browser infrastructure to support all those functionalities. So both the open source one and the big frontier models, but also both the cloud one, but also local ones. So we think we're very uniquely positioned on those space and very excited also about the trend of this moving in the future.

Operator

And we'll take our next question from Jim Callahan from Piper Sandler.

James Callahan

I guess starting with GX users with a strong uptick. I think you added as many users Q-over-Q as you did in all 2025. Any further commentary on kind of what drove the strength this quarter and maybe like the sustainability of that going forward?

Lin Song

Yes. Song here. I think I'll still try to answer. So yes, no, I think we are very excited to see the fast growth of GX in Q2, quite pleased about it. Yes. So I think fundamentally, it's a combination of things that while I think we definitely see that GX users are very AI conscious and the continued integration of the latest AI services that we have been providing has been able to resonate with the end users, which we are very excited. But I think there's also a fact that we are also now starting to work with more and more games and game developers by have provided more better integrated gaming ecosystem.

For instance, a typical case would be that now if you are a GX user, you would be able to participate in many interesting in Roblox games, for instance, where you can have game boosts and also daily rewards among others, right? So we basically see that GX almost becoming more and better integrated into the gaming ecosystem, and that definitely helps both for the access of those games, which provide hopefully future monetization opportunity, but also bring more users to GX, which we are very pleased about. So I would say it's a combination of both. So both more integration of AI, which is actually very helpful and very mindful to the end users, but also by we are maybe better embedded into the gaming world and gaming partners, which helps expand the user growth.

And then super quickly, you also mentioned about the sustainability. So we think the model is definitely very sustainable. The only thing we are mindful is just that, of course, during the summertime, yes, like it's summer and holiday is always a low season for GX, just to say. So we are right now in July and August, we will always be low season. So just a reminder, that's a physical limitation because we have people are not in front -- not home [ summer home ] not in front of the computers. There's limited stuff we can do about.

James Callahan

Great. That makes sense. And then with a couple of quick ones on the search business. Any comment on pricing versus impressions in terms of what's making up the revenue growth? And then I might have missed this, but any math we can do to back into like the other query part of the business would be helpful.

Frode Jacobsen

Yes. This is Frode here. I can comment on the search side. So I think overall, we see search revenue being driven predominantly by the value per search. As I mentioned, we also through engagement have tendencies of increases in search per user, in particular, on the smartphone side. But the general trend has been better matching with early search results, fewer queries needed per search, but then more than offset by better monetization on a per search query. I think the non-search part of query revenue has continued to grow well over 200% year-over-year. It's still in the single million dollars, but an increasingly important part of our revenue potential.

Operator

[Operator Instructions] We'll take our next question from Lance Vitanza with TD Cowen.

Lance Vitanza

I have 2 questions, please. The first is on the durability of growth and this valuation disconnect. At 6.5x next year's EBITDA, the stock still appears to imply skepticism around the durability of your growth trajectory. This despite the fact that Frode, you pointed out, right, the 21% growth CAGR over the past 10 years. But what gives management confidence that the current level of growth can persist beyond the next few quarters?

And what metrics should investors focus on to assess whether the growth is becoming more structural rather than cyclical?

Frode Jacobsen

That's a difficult question to answer. As Song talked a lot about, and I touched on too, I think the environment that we operate within has not been this exciting for a company like Opera for many years. So much is happening, so quick evolution around us and the browser playing a bigger and bigger role in people's daily life. So I think that we are very excited about. We see that our ability to turn that engagement into monetization and revenue has been very strong over the many years and continues to be.

And as we look ahead, we -- we also, in a way, take comfort in the fact that while we are very pleased with our growth, we've talked about e-commerce and how quickly that scales. We've mentioned travel as an opportunity that we think we also under-index in. Even if we are very pleased with the momentum in terms of the global market, we are still a very small player. And so what we see is that we still have the ability to navigate that opportunity space and sort of address opportunities, not just one by one, but as our capacity allows.

Lance Vitanza

Great. And then on MiniPay, it's now reached 18 million wallets. It's in 60 countries. You've got several dozen mini apps, and you recently launched a Visa card. At what point do you think the platform will have achieved sufficient scale so that you can begin prioritizing monetization alongside user growth? Are we still in the early innings of user acquisition? Or are we approaching an inflection point where the economic contribution could become more visible?

Lin Song

Yes, it's Song Lin. I'll try to comment a bit, right? So interesting. So okay. So first of all, I would say it's definitely still early stage, to be clear, right? So I think basically, -- it's -- we also learned it from our earlier experience in OPay and a few others that, of course, you almost have to be a bit more patient with fintech, especially with the fintech that we are doing, which is basically almost the infrastructure play and also the play to use technology to connect in the world almost, right? So I think those things do play a lot on very, very patient and careful growing of of user base, but also connecting all the partners across all the different content and regions, whatever and interconnecting them with technology because it's all about technology, which is what Mini Pay is about, right? And so -- and also it's also about building up all the partnerships, which we are very pleased that Visa, I guess, is a good example that we work with now to launch the Visa card across different countries among others.

So it's still early, but I think those are the nature of those kind of fintech services that it needs a huge infrastructure to be able to scale, but I think we also take comfort on a few things. So number one, I think from day 1, MiniPay is profitable and reasonably profitable. So we like -- so I think itself has been proven that it has a sound business model, and we have always been very disciplined. So that's number one, which we are very -- take comfort. And number two is, of course, that the trajectory of what happened in some other fintech investments we used to have also give us confidence that the moment this has scale and reach its network effect, it can happen relatively fast, right? Because it's all about like it has already have transaction volumes, it has already have right GMVs.

So it's all about at a certain time of the day we turn around the right trigger and starting to monetize by transaction volumes and by potential take rate and a few others, which is rather standard in the fintech space, right? So I think that can happen very fast. Once we think that it actually has that volume and connections in the world. So overall, very positive. It's still very early stage, but we think there's a lot more potential that we can see in the future.

Operator

And we'll take our next question from Jacob Stephan with Lake Street Capital Markets.

Jacob Stephan

Maybe just to start out on kind of the Browser Connector economics. I guess to start, when a user resolves a query inside of Claude, OpenAI, ChatGPT, whatever, through the actual connector versus your own environment, do you monetize that session today? And is the monetization rate any different between LLMs, I guess?

Lin Song

Yes. Okay. So I can try to answer that a bit, right? So I think that has a few benefits, both for the revenue and the others, right? So I think number one, as we also commented a bit that number one important is, of course, to solve the end user, right? Because many users say they like Opera, but they also like to use the AI of their choice, right, may be ChatGPT or may be others that from there, they can control it, right, that they can access the browser context and to visit the page and do a few things, right? So that's quite relevant. And we are happy, very happy to support that. And we think that's actually important functionality of Opera being stand-alone independent browser providers, right? We are very happy to be that infrastructure. So that's number one, that it's very important to the end user. It does have a benefit economically for the sense that, number one, it, of course, in those case, it doesn't cost Opera money because all the calculation and whatever, of course, based on the user subscription and from those cloud service.

So there is no additional cost to it, except providing that browser infrastructure, but no total cost among others, which is very effective. But also that be aware that all those activities are still within Opera browser, right, everything. Like you can -- a typical scenario is that in ChatGPT, you can in that interface to ask browser to go to a certain web page and to search and to what level, right? And of course, all those are still happening inside the browser environment and subject to whatever commercial deals browser would have with particular partners. So that's why we are very happy to also see that as far as the whole infrastructure and environment remain inside the browser, we think that can still have future benefits -- well, it has both current and future benefit, and it's just part of the whole browser play as if it's in regular web pages.

The only difference is just that in this case, it's not controlled by the end user, but controlled by the agent of choice or the AI of choice from the end user, all else equal. So I think both from us to be fairly positive about it, both for -- most important probably for be able to give you the choice, but also for the fact that as everything happened within the browser environment, we think there's plenty of opportunity for us to monetize.

Jacob Stephan

Got it. And maybe just touching on kind of the advertising growth versus kind of the margin quality of that. Obviously, advertising revenue is up 27%. Cost of inventory has kind of continued to climb here. I guess as Opera Ads expands beyond your own owned inventory, should we expect kind of gross margin to continue to kind of structurally decline? Or how should we think about kind of the incremental EBITDA margins, I guess, from that revenue growth?

Frode Jacobsen

Yes, Jacob, I can chime in on that. Even within Opera Ads, what we see also on third-party inventory is that our trend is an improving gross margin. So it's just about the mix between the different revenue types in our totality. In Q2, we had 38% cost of revenue, which is exactly what we expected. And we have guided it to tick up by about another percentage point or so for the year as a whole. But I would say we are able to do this while still increasing our adjusted EBITDA margin expectations because of economies of scale in the business as a whole and the fact that the Opera Ads platform has quite limited other OpEx from growing.

So I think that's something that we always managed carefully. We focus mainly on adjusted EBITDA on our cash flow, our net earnings as opposed to the gross margin percentage. But even within the gross margin percentage, I think you will see when you look at our history that from being quite insignificant in our P&L, it started to scale when we launched Opera Ads and that went through its initial growth phase. And now you see a much more stable and softer trend.

Operator

At this time, we have no further questions. So I'd like to turn it back over to Song Lin for any additional or closing remarks.

Lin Song

Sure. So like again, I think I would just like to take the chance to thank everybody for joining us. For us, it's quite straightforward. Our focus for the second half is about execution. We need to continue to improve our products, deeper engagement, deliver for our commercial partners and also convert the opportunities in front of us into sustainable, profitable growth. We are energized by our progress and by the work ahead, and we look forward to keeping you updated. Have a good day, everyone.

Operator

We'd like to thank everybody for their participation on today's conference call. Please feel free to disconnect your line at any time.