Open in interactive viewer → charts, metric popovers & call review

Opera's Growth Engine Changes Cylinders Mid-Flight

A 20% quarter and 10% of the market cap returned — but the e-commerce vertical that built the story is slowing, gaming is inheriting the torch, and OPay's IPO optionality just went quiet.
OPRA · Earnings Call · 2026-11-17

Opera Beat, Then Quietly Rotated Its Engine

Opera's third quarter looked like more of the same: revenue of $182.8 million, up 20% year over year, with ARPU up 22% — faster than revenue, which is the number that actually matters for a company that stopped chasing raw users years ago. Adjusted EBITDA of $43.2 million (a 24% margin) came in above the top end of guidance, and adjusted diluted EPS of $0.34 grew 16% in the quarter and 23% year to date. Cash sits at $109 million with no debt. On the surface, another beat-and-nudge.

Underneath, the composition changed. The company's advertising business grew 22% to $116.8 million, but the mix inside it is being reshuffled in real time — and management said so more plainly than usual.

The e-commerce rocket is sputtering, and gaming is catching the baton

E-commerce has been the entire Opera Ads story for three years. In February, CFO Frode Jacobsen described it as growing at “rates and a 100% year-over-year rate, including in the key fourth quarter” — Frode Jacobsen, Chief Financial Officer · 2026-02-26. That is over. On Tuesday he framed the change in regulatory terms: “some of the cross-border challenges, EU new rules, the market needs to adapt to that... and that led to a slowdown. We still have growth across all key verticals, including e-commerce, but slower than what we saw earlier in the year” — Frode Jacobsen, Chief Financial Officer · 2026-11-17. It echoes what happened in the U.S. at the start of 2025 — a policy shock that visibly reset a vertical's growth curve.

What's filling the hole is gaming. Song Lin: “we're definitely talking about more than 100% growth year-over-year... double digit or more percent of our total ads revenue and growing very fast” — Lin Song, Chief Executive Officer · 2026-11-17. Jacobsen added the sharpest line on the call, borrowed from history: gaming is “maybe similar in size now to where e-comm was in terms of its share of total advertising 3 years ago” — Frode Jacobsen, Chief Financial Officer · 2026-11-17. The enabling mechanism is new — cross platform attribution, stitching a desktop intent signal to a mobile install to a connected TV impression, which is also how CTV became a live inventory source this quarter.

The cost of that rotation is visible in the guide. Q4 revenue is guided to $198–205 million, roughly 14% growth at the midpoint versus 20% for the full year, and total ads revenue mix pushes cost of revenue to about 46% of the quarter's revenue from roughly 40%. Management is pre-paying the bill for new gaming clients and third-party inventory now to build 2027 vectors. Honest, but not free.

Local AI is the new Browser Connector — and the market is on the same wave

The keyword rotation here is unusually clean. Last quarter, Browser Connector was the company's number-one theme — the architectural pitch that Opera, unlike Chrome or Edge, lets you plug Claude or ChatGPT directly into a live browsing session. This quarter it barely surfaces. In its place: Local AI, web AI, and open weight model support all jumped to the top of the transcript.

The product behind it: “with the launch of Opera's WebAI, users are now able to access open-weight models on their machine with a simple toggle” — Lin Song, Chief Executive Officer · 2026-11-17. The pitch is that local inference is private, offline, and costs nothing but electricity — removing the Docker-install and terminal-command friction that kept it a hobbyist tool. Note this is not a fresh idea for Opera; local model talk dates back to the Llama 3 era in 2024. What is new is the packaging as a browser-native standard rather than a feature.

Importantly, this is a theme Opera is riding, not inventing. open-weight models entered the market-wide top-75 keyword set in the prior quarter — the broader conversation about running models locally instead of paying per token. Opera's edge is structural: no model lock-in, no capex race, no token bill. That positioning is why the browser arguably matters more in an AI world than it did before it.

The other half of the AI story is less philosophical and more arithmetic. AI Mode — Google's answer-generating, ad-bearing search experience — is already inside Opera's numbers. Song: “our query growth is excellent... that's obviously already included all the monetization that we're seeing in, say, AI Mode” — Lin Song, Chief Executive Officer · 2026-11-17. Query revenue rose 17% to $65 million, and the non-search slice of it is “growing very fast... in the 100-plus range... there is no cost of revenue. It's 100% gross margin revenue streams” — Frode Jacobsen, Chief Financial Officer · 2026-11-17. That is the highest-quality dollar the company books, and it is embedding itself into a category everyone assumed AI would destroy.

Ten percent of the market cap, returned

Opera's most under-discussed fact is capital return. Year to date it has handed back $137 million against a market capitalization of roughly $1.37 billion — about a tenth of the company — through a semiannual dividend (the July installment alone was $35.6 million) and $66 million of buybacks since March at an average of $17.33 per share. The share count fell 2.2% in a single quarter to 86.95 million. Cash conversion is not the constraint: operating cash flow of $37.4 million was 87% of adjusted EBITDA.

And there is a mechanical catalyst the tape may still be ignoring: the conversion from ADSs to ordinary shares listed on NASDAQ in early December, which “removes a barrier for inclusion in certain equity indices and investment funds whose mandates restrict investment in depository shares” — Frode Jacobsen, Chief Financial Officer · 2026-11-17. For a company this size with this cash generation, index-eligibility is not cosmetic.

The valuation debate has not moved. Last quarter an analyst put it bluntly: “At 6.5x next year's EBITDA, the stock still appears to imply skepticism around the durability of your growth trajectory” — Lance Vitanza, Analyst · 2026-08-19. At roughly seven times the $172–175 million adjusted EBITDA guide with net cash behind it, the market is still pricing a business that fades — while management buys its own shares at that price.

OPay stops being a story

The quietest change on the call was the most telling. For three consecutive quarters, OPay's eventual IPO was a reliable talking point — in August, Jacobsen said Opera “continue[s] to expect that OPay will ultimately go public... It will lead to an immediate transparency as to the value of our founding stake” — Frode Jacobsen, Chief Financial Officer (CFO) · 2026-08-19, and in February noted the hire of public-company executives as a sign of the natural next step.

This quarter an analyst noted OPay had decided not to pursue the offering. The answer was a shrug:

nothing to report on our plans beyond that. So the IPO as they structured it was primary only. We're not in a rush, neither the other pre-IPO investors.

Frode Jacobsen, Chief Financial Officer · 2026-11-17

Worth noting the arithmetic that went unmentioned: the stake is carried at $301 million, or 8.5% — versus $300 million at 9.5% in August. A point of equity evaporated for no cash. Meanwhile the replacement optionality narrative is MiniPay, now at 20 million activated wallets and nearly 600 million cumulative transactions, with Stripe-enabled top-ups via card, Apple Pay and Google Pay. It rides the same stablecoin adoption wave visible across the market — but management was explicit that monetization remains secondary to adoption and utility.

So the real question entering 2027 is not whether Opera can grow 20% again. It is whether gaming, CTV and 100%-margin non-search query revenue can outrun an e-commerce vertical that just hit a regulatory wall — while the company hands a tenth of itself back to shareholders each year. The 14% Q4 guide is management's own answer, and it is a cautious one.