The new face of the browser: Who’s building AI-first browsers, what they do and how they could upend advertising

Big tech and nimble startups alike have moved quickly in 2025 to weave large-language models and agentic assistants into the thing most of us use to reach the web: the browser. Here's a list of the main AI-powered browsers and browser modes currently in market or announced, with ownership, their strengths and known weaknesses (security, privacy and product limits) and how each could change advertising and the ad ecosystem.

OpenAI — ChatGPT Atlas

Ownership and what it is: Atlas is a full web browser released by OpenAI that embeds ChatGPT at its core and ships initially for macOS with planned rollouts to Windows, iOS and Android. It adds a persistent ChatGPT sidebar and an “agent mode” that can perform multi-step web tasks on users’ behalf.

Strengths: Tight integration with ChatGPT and its memory features gives Atlas powerful contextual assistance (summaries, comparisons, task automation). OpenAI positions the product as a “super-assistant” that unifies search, tools and user context inside the browser.

Weaknesses and concerns: Reporters and analysts note privacy and trust questions (how memories and browsing data are handled) and the broader strategic challenge of displacing Google’s entrenched Chrome/SERP dominance; regulators and rivals will scrutinize data-use and ad implications. As with other AI browsers, there are also technical and security risks when agents are given permission to act across sites.

Advertising potential and disruption: By routing queries and task flows through ChatGPT, OpenAI stands to capture a larger share of attention and the referral paths that drive ad impressions and search revenue. News outlets and analysts highlighted that Atlas could threaten Google’s search advertising franchise by substituting direct, synthesized responses for traditional search results — a structural risk to the ad market.

Microsoft — Edge with Copilot Mode

Ownership and what it is: Microsoft’s Edge browser has been retrofitted into an “AI browser” through a built-in Copilot/“Copilot Mode” feature that integrates Microsoft’s Copilot assistant into the browser UI. It is available on Windows and Mac (opt-in) and can be allowed to access tab content, history and other context if the user consents.

Strengths: Microsoft can leverage enterprise integrations (Office, Teams, Windows) and deep cloud resources to deliver productivity gains (summaries, drafts, task automation) and a consistent Copilot across workflows. The feature benefits from Microsoft’s scale and existing user base.

Weaknesses and concerns: Copilot’s capabilities are permission-gated and Microsoft has flagged some features as time-limited or experimental; privacy and product-placement concerns (prompts to try Copilot, monetization hints) have drawn user complaints. There are also questions about whether retrofitting AI into an incumbent browser can match the experience of AI-native browsers.

Advertising potential and disruption: Microsoft can route more user attention through its own services and partner ecosystem (Bing, Microsoft Ads). Copilot could both create new ad placements inside an assistant UI and reduce traditional search ad clicks if users accept AI answers in place of SERPs — though Microsoft’s enterprise focus means some monetization could come from subscriptions and services rather than pure ad dollars.

Perplexity — Comet

Ownership and what it is: Perplexity, a startup that built an AI search/assistant stack and is backed by Nvidia among others, launched Comet, an AI-first browser that tightly integrates its assistant to “travel the web” with users rather than act as an add-on. Comet was released publicly in 2025 after an initial limited launch.

Strengths: Comet is designed around conversational, search-augmented browsing and early users praised its speed at summarizing pages and doing multi-step tasks. Perplexity has optimized for delivering concise, grounded answers and agentic helpers.

Weaknesses and concerns: Independent security reviews and audits (reported by rivals and security outfits) flagged vulnerabilities and prompt-injection risks in Comet’s handling of web content and screenshots; critics warned that agentic features can open the door to phishing or code-execution style attacks if not carefully sandboxed. Perplexity has been responsive but the issues underscore a broader security trade-off for deep assistant control.

Advertising potential and disruption: Comet positions itself as a direct competitor for user attention and search flows. If users accept assistant answers instead of clicking to publisher pages, the traditional referral and ad revenue model is at risk — but Perplexity has also explored subscription tiers and partnerships (some distribution deals announced with payment platforms) as alternative revenue streams.

The Browser Company — Dia

Ownership and what it is: The Browser Company (the team behind Arc) launched Dia, an AI-native browser built around chat with open tabs and an assistant that understands page context. Reported as a product pivot from Arc, Dia emphasizes an assistant-first interface and was released in beta in 2025.

Strengths: Dia’s design foregrounds a conversational workflow where users “chat with their tabs,” a novel UX that can reduce tab overload and speed research tasks. It won early praise for product design and the way it weaves AI into core navigation.

Weaknesses and concerns: As with other boutique browsers, scale and monetization remain open questions; Dia adopted subscription pricing for advanced features and the company has signalled limits on free usage. Startups in this space also face integration and security challenges when agents interact across arbitrary web pages.

Advertising potential and disruption: Dia’s user experience could reduce page visits to ad-funded sites if users depend on assistant summaries; conversely, Dia could enable new in-browser commerce and partner experiences (subscriptions, embedded services) that monetize outside traditional display/search ads. Atlassian’s reported acquisition interest suggests enterprise distribution plays are also possible.

Opera — Aria / Neon

Ownership and what it is: Opera has been integrating its own AI assistant, Aria, across Opera desktop, Opera GX and Opera Mini for some time; in 2025 Opera also launched Neon, an AI-centric browser focused on agentic tasks and power-user workflows. Opera’s approach mixes in-browser AI features with a freemium model.

Strengths: Opera’s advantage is familiarity and distribution (Opera still has a large installed base in some markets) and the company offers Aria for free across platforms. Aria pulls from multiple models and can generate images, synthesize responses and manage tabs.

Weaknesses and concerns: Opera’s position as a smaller browser means it must differentiate by specialized features; critics ask whether its AI offerings can match the scale and model quality of Google, Microsoft or OpenAI.

Advertising potential and disruption: Opera has historically experimented with alternate revenue models (bundled services, partnerships and integrated content). Its free Aria model and Neon subscription offerings show hybrid approaches: ad replacement via premium features, new commerce hooks, or tighter partnerships with content and commerce providers rather than relying purely on display ads.

Brave — Leo / Ask Brave

Ownership and what it is: Brave, the privacy-focused browser maker, embeds an AI assistant called Leo and has been evolving “Ask Brave” as a search + AI experience that prioritizes privacy and independent indexing. Brave emphasizes privacy-preserving answers and its own Brave Search backbone.

Strengths: Brave’s privacy posture and self-hosted search index let it offer AI assistance without routing everything through the big cloud providers; that appeals to users who want AI help but are wary of Big Tech data collection. Brave also moves quickly on product updates and privacy controls.

Weaknesses and concerns: Brave remains small relative to Chrome and the depth of assistant features depends on model and indexing partnerships. Privacy-first positioning can limit certain monetization paths (targeted ads), pushing Brave toward subscriptions or contextual, non-tracking ad models.

Advertising potential and disruption: Brave’s approach could accelerate ad models that don’t rely on cross-site tracking — for example, contextual ads or paid search/subscriptions. If Brave’s AI keeps users within synthesized answers, publishers’ referral traffic could fall, but Brave has repeatedly signalled it will experiment with privacy-respecting monetization rather than the surveillance ad model.

What matters for advertising

Attention capture vs. referral flows: Multiple reports and analysts flag the same structural issue: assistant answers and agentic task completion can short-circuit the click flows that feed publisher pages and search ad impressions. That represents both an opportunity (new assistant-native ad formats, commerce integrations, subscriptions) and an existential threat to the current search ad market.

Security and sandboxing are central: The push to give assistants the power to click, fill forms and act on pages creates new attack surfaces (prompt-injection, screenshot exploits, malicious pages). Several security audits and vendor disclosures recommend careful feature gating and sandboxing.

Privacy choices will steer monetization: Browsers with strong privacy positioning (Brave, Opera’s choices) are likelier to pursue subscriptions, partnerships and contextual ad formats; firms that can leverage enterprise bundles (Microsoft, Atlassian interest in The Browser Company) will mix subscription revenue with service monetization. OpenAI and Perplexity — by controlling the assistant layer — could also build direct commerce or subscription revenue that sidelines ad networks.

The Bottom-line

AI is no longer only an add-on: in 2025 multiple companies have turned the browser itself into an AI product. Each approach carries trade-offs — assistant convenience versus security and privacy risks, and new monetization experiments versus the entrenched ad-driven web economy. The most immediate advertising impact will be a reallocation of attention: if users accept synthesized answers and agentic task completion, advertisers and publishers will need to find new ways to reach users — whether through assistant-native placements, commerce integrations, subscriptions, or partnerships with the browsers themselves.