The setup: In the span of 48 hours this week, OpenAI unveiled Atlas and Microsoft relaunched Edge's Copilot Mode. Both browsers can now see your tabs, synthesize information across pages, and take actions on your behalf. Google's been quietly building similar features into Chrome (though you probably haven't noticed).
The tension: Every major tech company is racing to put AI directly into your browser. But here's what nobody's talking about: This doesn't just change how we browse the web. It might fundamentally break the entire economic model that's kept publishers alive for the past 20 years.
Did we say break? We actually meant "reinvent". We'll explain down below. But first, here's what's new in the AI browser wars...
Microsoft Edge: From Copilot Mode Launch to Full AI Browser (July-October 2025)
Microsoft actually took two swings at the browser wars in the span of three months.
July 28, 2025 - The Initial Launch:
Microsoft introduced Copilot Mode as an experimental feature in Edge, available for free (for a limited time) on Windows and Mac. It was fully opt-in.
What launched in July:
- Streamlined new tab page: Single input box combining chat, search, and web navigation.
- Multi-tab context: With permission, Copilot could see all open tabs and compare information across them.
- Natural voice navigation: Speak to Copilot instead of clicking and typing.
- Dynamic pane: Open Copilot in any webpage without losing your place.
- Actions (promised soon): Would access history and credentials to book reservations and manage errands.
- Journeys (promised soon): Would organize browsing into topic-based journeys.
As for privacy? Data protected under Microsoft Privacy Statement with visual cues when active. Full on/off control in settings.
October 23, 2025 - The Relaunch:
Three months later, Microsoft came back with the features they promised—and positioned Edge as "your AI browser" instead of just a browser with AI features. Same name, completely different ambition.
What's actually new in October:
Copilot Actions in Edge (now live in U.S. limited preview):
- Actions with Voice: Perform simple tasks like opening webpages or jumping to specific information without scrolling.
- Complex actions via chat: Tell Copilot to unsubscribe from shopping newsletters by going through your inbox, or make restaurant reservations for you.
- Coming soon: Browser-based voice for completing complex tasks hands-free.
Journeys (now live in U.S. limited preview):
- Past browsing projects automatically grouped into helpful topics.
- Jump back in exactly where you left off.
- Suggests smart next steps.
- With explicit permission.
Browsing History Integration (new):
- Copilot can now use your browsing history (with permission) to provide higher quality responses.
- "Chat with Copilot about that blue hoodie you were looking at last week".
- Opt-in via Page Context in Settings, turn off anytime.
Scareware Blocker (new):
- Uses local AI to protect from full-screen scam takeovers.
- Enabled by default on qualifying devices.
Enhanced Password Management:
- Create strong passwords, store safely, monitor for breaches 24/7.
- Instant alerts and easy updates.
The positioning shift is significant. In July, Microsoft called Copilot Mode "a new way to browse." In October, they're calling Edge "your AI browser" and emphasizing that it "anticipates, assists and accelerates your experience online."
OpenAI Launches ChatGPT Atlas (October 21, 2025)
https://www.youtube.com/live/8UWKxJbjriY?si=VJMypXpG44ewMOVH
OpenAI didn't just add features to an existing browser—they built a new one from scratch with ChatGPT at its core. Atlas launched worldwide on macOS for Free, Plus, Pro, and Go users, with beta access for Business (and Enterprise/Edu if enabled by admins). Windows, iOS, and Android versions are coming soon.
What it does:
- Integrated ChatGPT memory: Conversations draw on past chats and details to help with new tasks.
- Browser memories (optional): ChatGPT remembers context from sites you visit and brings it back when needed. Completely under your control—view, archive, or delete anytime.
- Agent mode (preview): Available for Plus, Pro, and Business users. Can research, analyze, automate tasks, plan events, and book appointments while you browse.
- Multi-tab intelligence: Works across tabs to consolidate information (like turning flight, hotel, and activity tabs into a single itinerary).
- Deep app integration: Works with Google Calendar, YouTube, Maps without leaving your current page.
The interface: New tab page lets you ask questions or enter URLs for faster, more useful results in one place. Select from tabs for search links, images, videos, and news.
Privacy and control:
- Toggle what ChatGPT can see per-site in the address bar.
- Incognito mode to temporarily log out of ChatGPT.
- By default, content isn't used to train models (opt-in available).
- Parental controls work in Atlas
Agent mode safety:
- Can't run code, download files, or install extensions.
- Can't access other apps or file system.
- Pauses on sensitive sites like financial institutions.
- Can use logged-out mode to limit access.
Google Launched "Chrome's Biggest Upgrade Ever" (September 18, 2025)
Google isn't taking all this Chrome competition without a fight, as you'd expect. They also quietly integrated Gemini across Chrome with 10 major AI updates, calling it "the biggest upgrade to Chrome in its history."
What rolled out:
- Gemini in Chrome: Now on Mac, Windows (U.S., English), mobile (Android/iOS), and coming to businesses via Google Workspace.
- Agentic capabilities (coming soon): Will handle tedious tasks like booking haircuts or ordering groceries.
- Multi-tab work: Compares and summarizes info across multiple websites.
- Tab recall: Finds webpages you visited previously with prompts like "what was that blog I read on back to school shopping?"
- Google apps integration: Works with Calendar, YouTube, Maps without changing tabs.
- AI Mode in omnibox: Access Google Search's AI Mode directly from the address bar (rolled out late September).
- Page questions: Ask questions about your current page from the omnibox, get AI Overview in side panel.
- Gemini Nano scam protection: Enhanced Protection mode expanding to stop fake viruses and giveaway scams.
- Spam notification detection: Reduced unwanted notifications by ~3 billion per day on Android using AI.
- One-click password changes: Change compromised passwords with single click on supported sites (Coursera, Spotify, Duolingo, H&M, more).
The Underdogs: Dia and Perplexity Comet
Before the tech giants joined, two browsers pioneered full agentic AI:
Dia Browser (from the makers of Arc, now in General Availability on macOS):
- Writing partner in every text box with in-line copy editor.
- Tutor in every tab for learning.
- Personal assistant that's already up to speed on your context.
- Retail concierge for shopping decisions.
- Privacy-focused: "exists to make the internet better, not to profit off your personal data."
- AI that understands, organizes, builds, emails, creates, and shops.
- Can organize tabs by category, build websites, draft email replies, create study plans, shop for products.
- Described as "the browser that works for you."
More on both of those in our previous coverage of the AI browser wars here.
Now Here's Why This All Matters
The real question is this: How will agentic AI actually change how we consume content? Will browsers become pure chat interfaces that replace traditional web reading? Or will “anchor” pages become more valuable because you can process more info in one place, with users staying longer to chat with AI about the content on the page itself, creating monetizable dwell time for “information hub” style sites?
Most likely? Browser providers become the new gatekeepers of banner ads, with higher subscription tiers for no ads (hello, streaming model) while free tiers get ad-blasted. Publishers and advertisers will then work through the browser as a new middleman vs Meta & Google’s duopoly… with all the costs that come with it.
Allow us to explain...
How We Got Here: The SEO Playbook That Built the Modern Web
For the last two decades, the web publishing playbook has been brutally simple:
- Identify high-volume search queries.
- Create individual pages targeting each query.
- Optimize for Google's algorithm.
- Monetize through ads or affiliate links.
- Repeat 50 times.
A travel site doesn't create one comprehensive guide to Paris. They create:
- "Best hotels in Paris"
- "Top 10 Paris restaurants"
- "Paris in 3 days itinerary"
- "Cheap flights to Paris"
- "Paris tourist attractions"
Each page is an entry point. Each entry point is a chance to serve ads. More pages = more traffic = more revenue.
This model works because Google sends users to the specific page that answers their specific query. Publishers win by casting the widest net possible.
The AI Browser Model: Why Individual Pages Stop Mattering
Now imagine you're planning that Paris trip in Atlas or Edge with Copilot Mode enabled.
You open a tab with a Paris travel guide. Instead of clicking through 15 different pages, you just... ask the AI questions:
You: "What's the best 3-day itinerary if I'm staying near the Louvre?"
AI: Reads the page you're on, cross-references your open tabs, synthesizes information. "Based on your location near the Louvre, here's a day-by-day breakdown..."
You: "What about vegetarian restaurants in that area?"
AI: Doesn't open a new page. "I found 6 highly-rated vegetarian spots within walking distance..."
You: "Book me a table at the top-rated one for tomorrow at 7pm."
AI: Takes action directly. "Done. Reservation confirmed at Gentle Gourmet."
You never left that first page. You got everything you needed through conversation with AI that could see and reason over the content.
The business case problem: If users don't need to visit 15 separate pages anymore, why would publishers create 15 separate pages? That's wasted time and effort for middling return.
The Shift: From Search Farming to Information Hubs
Here's where it gets interesting. In an AI browser world, your content strategy flips:
Old model:
- 50 thin pages targeting individual keywords.
- Goal: Rank for as many searches as possible.
- Monetization: Pageview volume × ad impressions.
- User behavior: Land, consume, leave.
New model:
- 5 comprehensive information hubs.
- Goal: Become an authoritative source AI references.
- Monetization: Engagement depth × time on page.
- User behavior: Land, stay, chat.
Instead of creating "Best Hotels in Paris," "Best Restaurants in Paris," and "Paris in 3 Days" as separate pages, you create one massive, encyclopedic Paris Travel Hub that becomes the authoritative source AI models pull from—and where users actually stay to have conversations with the AI about the content.
What This Looks Like In Practice
Let's say you run a tech review site.
Today's SEO approach:
- "iPhone 16 review"
- "iPhone 16 vs iPhone 15"
- "iPhone 16 camera quality"
- "Is iPhone 16 worth it?"
- "iPhone 16 battery life"
- "Best iPhone 16 cases"
That's 6+ pages, each optimized for different search queries.
AI browser approach:
One comprehensive iPhone 16 Deep Dive that includes specs, comparisons, camera analysis, battery tests, buying recommendations, and accessory guides. Users land once and chat with AI about any aspect:
- "Compare the camera to Samsung S24"
- "Is the battery better than the 15?"
- "Show me cases under $30"
The AI references your comprehensive hub, synthesizes answers, and the user never leaves your page. You're not optimizing for keyword rankings—you're optimizing for being the source the AI trusts and users engage with.
The Monetization Flip: Volume to Depth
This completely changes how publishers make money.
Pageview economy:
- User visits 10 pages.
- Sees 30 ads total (3 per page).
- Average time: 45 seconds per page.
- Revenue: $0.50 CPM × 10 pageviews = $0.005 per user.
Engagement economy:
- User visits 1 hub page.
- Spends 15 minutes chatting with AI about content.
- Sees premium contextual ads or sponsored AI responses.
- Revenue: $5 CPM × sustained engagement = $0.075 per user.
You make 15× more per user by keeping them engaged on one authoritative page than bouncing them through a dozen thin ones... if they even visit in the first place.
The Gatekeeper Problem: Who Controls The Money?
Here's where things get messy.
If browsers become AI-powered chat interfaces, the browser provider (Microsoft, Google, OpenAI) becomes the new middleman between publishers and audiences.
Think about what Microsoft's Copilot or OpenAI's Atlas can do:
- See your content.
- Synthesize it for users.
- Insert ads or sponsored responses.
- Take actions (book hotels, make purchases).
Who gets paid when AI books that hotel using information from your travel guide?
Most likely scenario: Browser providers adopt the "streaming service" model of freemium ad-supported streaming, with higher paid ad-free tiers as subscription.
- Free tier: Ad-supported browsing with AI features.
- Premium tier: Ad-free AI assistance for $20/month.
Publishers who want their content surfaced? They'll work through the browser's ad network. Just like how Netflix and Hulu became gatekeepers for TV content, Edge and Atlas could become gatekeepers for web content.
Translation: More costs, less control, and a new platform taking a cut.
Who Wins, Who Loses?
If this plays out like we imagine it, there will be some clearly defined winners and losers.
Winners:
- Browser providers (Microsoft, Google, OpenAI) who control the AI layer.
- Premium publishers who can create truly comprehensive, authoritative hubs.
- Niche experts who already build deep, engaging content.
Losers:
- SEO farms churning out thin, keyword-stuffed pages.
- Affiliate sites that rely on driving clicks to product pages.
- Ad arbitrage players who profit from volume, not depth.
What Publishers Should Do Right Now
If you're a publisher or content creator, here's the play (just our POV; do what you will with this information!):
- Audit your content portfolio. Which pages could be consolidated into comprehensive hubs?
- Invest in depth over breadth. One 10,000-word authoritative guide beats ten 1,000-word thin posts.
- Optimize for AI comprehension. Structure content so AI can easily parse, reference, and synthesize it. Think clear headers, data tables, and comprehensive FAQs.
- Prepare for engagement metrics to matter more than pageviews. Time on page, scroll depth, and interaction will become the new KPIs.
- Experiment with interactive elements. If users will chat with AI about your content, make your content worth chatting about. Calculators, comparison tools, interactive data—these keep users engaged.
This Isn't Hypothetical: Wikipedia Just Lost 8% of Its Traffic
Pretty much every online publisher knows this is a problem. It's already devastating content creators.
For example, last week, the Wikimedia Foundation revealed that Wikipedia, the internet's free encyclopedia that literally everyone uses, just lost roughly 8% of human pageviews compared to 2024.
At first, they thought traffic was spiking, especially from Brazil. But after updating their bot detection systems, they realized the truth: those weren't humans. They were sophisticated bots designed to scrape content while pretending to be real users.
Once they filtered out the fake traffic, the real picture emerged: fewer actual people are visiting Wikipedia. And the culprit is exactly what we're talking about—AI-powered search summaries that give you the answer without ever sending you to the source.
The Death by a Thousand AI Summaries
Here's what the data shows, and it's brutal:
- Recent Pew Research found that when Google shows an AI Overview summary, only 8% of users click through to actual websites. Compare that to 15% when there's no AI summary. That's a 50% drop in clicks just from adding AI.
- For questions starting with "who," "what," "when," or "why," Google now triggers AI summaries 60% of the time. These are exactly the kinds of questions Wikipedia excels at answering.
- Users rarely click the sources cited in AI summaries. It happens in just 1% of visits to pages with AI Overviews.
Think about what this means: Wikipedia is still being used; it's just invisible now. Almost every large language model trains on Wikipedia. Search engines pull its information to answer questions in AI summaries. But if people never actually visit the site, who's going to keep creating and updating all that knowledge?
Wikipedia runs on volunteers. Volunteers who contribute because they want to share knowledge and be part of something bigger. But that intrinsic motivation only goes so far when your work becomes invisible infrastructure for billion-dollar AI companies.
The Web Fights Back: Enter Content Signals
The internet isn't taking this lying down. Publishers, content creators, and platforms are calling this moment a "web infrastructure revolt."
Cloudflare just launched Content Signals Policy—a new addition to robots.txt files that lets website owners express preferences for how their content gets used after it's accessed.
It's beautifully simple. Three signals you can set to "yes" or "no":
- search: Can you use this to build a search index?
- ai-input: Can you input this into AI models for real-time answers?
- ai-train: Can you use this to train or fine-tune AI models?
For example, if you want search engines to index your content but don't want it used for AI training, you'd add this to your robots.txt:
Content-Signal: search=yes, ai-train=no
Cloudflare is already deploying this for 3.8 million domains that use their managed robots.txt feature, automatically signaling that they don't want their content used for AI training.
The Deal Is Dead
For 25 years, the web operated on an implicit deal: You could scrape content, but you'd send referral traffic and give attribution.
AI companies broke that deal. They scrape everything, synthesize it into answers, keep users in their own gardens, and send nothing back.
Content signals aren't a technical block either. Bad actors can still ignore them. But they create a clear, standardized way for website owners to say "these are my rules." Combined with enforcement tools like Cloudflare's WAF and Bot Management, they give creators at least some control back.
But here's the bigger question nobody's asking...
Was Google Ever the Right Answer?
The open web is worth saving, but let's ask a harder question: Was using Google to find whatever random website happens to rank first ever the right way to discover trusted information?
Think about it. Wikipedia is the most-cited source in AI summaries and one of the most trusted sites on the internet. So why do we go to Google first instead of going straight to Wikipedia? Why don't credible sources have their own front doors that don't depend on the whims of Google's algorithm?
In our opinion, it's weird that the internet never got its "TV Guide"—a central, curated directory where you could browse trusted sources by category, recommended by experts instead of SEO tricks.
We've spent 25 years letting middlemen like Google decide what information surfaces, which worked fine when they sent traffic back to sources. But now that the middleman is keeping users in their own garden, the whole model is breaking.
Here's the wild part: AI could actually fix this—if done right.
Imagine finding information online similar to how you find it in real life: like a library. From this POV, you could look at AI like a librarian, working directly with the entire network of trusted sources to find you the exact "book" of information you need right off the shelf (in this context, properly attributed and linked). No middleman tax. No arbitrary ranking games. Just direct connections between you and the most credible sources on the topic you're trying to research.
Whether content signals will make a difference depends on whether AI companies respect them—or ignore them and face the consequences. It honestly feels like an uphill battle with the rise of AI browsers, especially ones who can process information locally on your machine. And these browsers do create a better experience for the user who just needs information, not an entire article whose sole purpose for existence was dedicated to answering a single question. That just feels inefficient, y'all, and it's why people hate Google now and like AI search tools.
So maybe the bigger opportunity here is rethinking how we access information entirely.
The search era gave us convenience... at first. The AI era could give us something better: direct relationships with the sources that actually create knowledge.
But—and this is critical—that only works if we solve the creator business model problem. Creators need to get paid when their work gets used, whether through attribution that drives traffic, direct licensing deals, or new compensation structures we haven't imagined yet.
If AI companies keep the ad revenue (or future subscription revenue) while creators get nothing, the entire ecosystem collapses. No Wikipedia volunteers. No independent publishers. No one creating the knowledge that makes AI useful in the first place.
Instead, this new AI agentic search era needs to priortize two vital things to protect the stability of the network:
- Tools and UIX that prioritize access to efficient, reliable information (instead of distract from it).
- Monetization that doesn't just extract new levies on existing publishers, users, and advertisers, but creates a virtuous system that enables all three to profit.
In a perfect world, AI browsers would increase the platform and status of publishers who provide meaningfully valuable information for readers who need it (over those simply farming content to get clicks and views), connect advertisers directly with buyers who want exactly what they're selling (instead of serving ads to thousands or millions who couldn't give a care otherwise), and reduces the overall cost and friction of doing business online through tools that make it as easy as possible to find what you need (reducing the tax that Google and Meta charge advertisers, and providing true ad / token revenue sharing content with publishers so they can keep doing what they need to do to make a living making valuable, trusted content). What, like it's hard??
The Bottom Line
We're watching the web's economic model shift in real time. From one perspective, the SEO gold rush that defined digital publishing for 20 years is ending. From another perspective, the SEO best practices that were always the most useful for users (and Google's bottom line) are about to become even more important. We actually just published a great interview with SEO titan Mark Williams-Cook on this topic if you want to learn more about how the agentic age is changing the web:
In our opinion, the era of the information hub is just beginning.
Atlas, Edge Copilot, and whatever Google ships next aren't just new browsers. They're the infrastructure for a completely different internet—one where AI sits between users and content, where depth beats breadth, and where engagement beats raw pageviews. With more bots online than ever before, it's pageviews just aren't a valuable metric anymore, and the entire web monetization ecosystem needs to change to reflect this.
Publishers who figure this out early will thrive. Those who keep churning out thin SEO content? They'll make a profit in the short term, but when the bot bill comes due, they'll fold under their thin brands and horrible user experience, and become irrelevant over time as their search traffic continues to decline.
That said, publishers and advertisers who embrace AI browsers still need to be wary of the new taxman.
Google has Chrome, and Chrome will still be a major player here. AdSense will simply move from the first result in Google to the banner ad on the side of your chat with Chrome. Meta may very well launch their own AI browser to expand their ad empire beyond FB and IG as traffic there plummets as well (I wouldn't rule Amazon out here either, since their ad business is booming, too). And you better believe OpenAI, Perplexity, Microsoft, and whoever else launches an AI browser will charge their own inevitable ad taxes as well.
More competition will be good for advertisers at first. It'll lower the overall rate that the gatekeepers can charge, and smarter AI-driven targeting will lower the cost to reach qualified buyers even further. But over time, as the industry consolidates again, there's a major risk of becoming too reliant on these new browsers as well. Act accordingly.
So as you can see, AI browser wars aren't just about which company builds the best chatbot. They're about who controls the next era of the web... and how much they'll charge for access.
