The AI Browser Wars Explained

In a recent AI Moment Podcast, we dived into AI Browser (I dubbed the AI browser wars).

So right now it is Dia vs Comet vs Atlas vs Chrome (with some flirting from Microsoft Edge, Firefox with some tweaks and Opera (remember them?) will a number of updates.

I dived into the what, who, why, and how of the AI Browsers.

The Top Line:

• Development in browsers has previously been strategically slow - no need to and limited competition (plus Google and Apple ecosystem didn’t need it to be - apart from Apple’s search deal with Google)

• Value for the company for data and insights, but not many revenue opportunities apart from ads, paid agreements for search deals, paid extensions and paid for services.

• For most companies, it’s a data play and they then own a deeper experience with the brand and tech.

• Browsers have struggled as a business line and struggled to gain traction with millions of users. Unless you are the default like Edge, Chrome or Safari it is an expensive business as acquisition costs will be high. Even Brave which has been going for years and touts huge numbers have struggled to really grow

• Browsers can also be marketing moments. Often Google will promote their services on on their start page. Apple alignment with privacy didn’t align it as a revenue maker away from the Google search deal

  • Ask the Chrome team how it won the default browser race…growth tactics like download for better experiences, switch etc were used on their homepage

Who & What

Why?

Things to be aware of

• Risks: There have already been examples and high-profile cases of security risks of agentic browsers (AI browser) - Perplexity has been called out by Brave browser for security issues.

• Tech Legal Battle: Amazon are now suing Perplexity for Comet’s in browser assistant being able to find discounts and taking actions (maybe with scraping while on websites)

• Desktop only at the moment - Comet is apparently coming to iOS soon

• AI Tech Stack Race To Top & Bottom

  • Browsers are sticky on desktop and can control what you do and how you do it

  • Building a browser is a way to not be limited by reach or removed by Google searches & app stores etc - you can act around or above them (albeit built on Chrome tech)

  • All of the AI browsers are all built on top of Chrome (the open source Chromium project)

  • Claude have created a different level of attack with Chrome extension & most recently tests with Excel. Their desktop app is a layer under the operating system, allowing them more control with their tool and what access and collaboration it could have

• Tested & Now Work In: I work in many, but I tend to use Comet now Dia has had a number of features blocked, specifically limiting summarising content & YouTube - Google does not take kindly to ad blockers

• Product Recyclement: Once one creates a feature, the others copy quickly - assistant sidebar, working between tabs

• Power Of Brand Vs Privacy: Apple are still slow to move, suspect it is down to privacy stance and allowing others to develop good experiences within Athe pple ecosystem - Apple still owns the user with App Stores and OS’s

Strategic Points To Consider

• G Power 💪: Google has the browser market to lose - but will there be a breakout browser built off Chrome? Unlikely in the near future

• Agentic Future 🤖? Shift to agentic makes most sense in the browser you trust and you could see it happen if big companies don’t blindly block - expect this to be a trend. The agentic web is being built under the surface of the web, these companies need to invest massively in building for agentic behaviours.

• Winner 🏆! It will be winner will take most - best features and sticky users will mean that the browser will win many new customers

• Frustrations 😡: AI browser actions are frequently slow and often take too long and don’t complete tasks. This will get better, but AI is still considered slow by many

• Litmus Test For OpenAI 🚦: This is a good litmus test for ChatGPT and how sticky they are

  • ChatGPT pushed building inside of their new ecosystem (create decks with Canva by tagging on ChatGPT) - Atlas might have to be their action engine - their way of controlling how you work with and on ChatGPT

• We are seeing another bifurcation ↪️: “app-lication of the web” while the open “information web” is declining under AI companies’ huge investments and development

• Enable Or Block ❌: Will be the question many companies have to ask themselves from 2025 → 2026

• Hybrid 🆕: AI browsers are future leaning so for agentic usage as well as human - can this scale properly with companies trying to protect IP (Cloudflare and Perplexity already had a public feud over Comet and the way they gather information and show up as bots)

• Apple Power 📲: Apple will still influence the space with OS ownership, app-store ownership and controlling privacy. Apple could shift people back towards apps to control more or could still with SLMs on your devices removing the need for many browsers etc

• Switching costs ↔ are low but enough to put many changing from Chrome to Atlas or Dia etc - humans hate change!

• Transfer From Model To Browser 🔎: The question might be does LLM use mean you will be driven to download and use: Gemini → Google Chrome / Atlas → ChatGPT / Comet → Perplexity

• Legal 👩‍⚖️: Law suits might play an important part, perplexity and OpenAI being sued by publishers.

  • Reddit is suing many providers currently, will this impact progress?

• Blocking ⛔️: Web sites and apps have to decide to allow or block agents, this will be a big moment for the future web as browsers will be taking actions and becoming agentic - so will act like human and assistant/agent/bot.

• Premium Browsing 💳: There could be a case where you are given a premium version of the AI Browser with subscriptions

• New UX 🔓? We haven’t seen new user experiences yet, the browser is still point and click and scroll and swipe it hasn’t been amended yet

  • Once the UX changes, companies and attribution leaders will have to change how it tracks - a way to think is if Atlas drives a chatGPT purchases will it be seen as an agentic attribution or will it recorded differently

My prediction Chrome will make periodic changes that will act as a strategic defence for them, they will mimic competitors’ features and lock in the Google ecosystem and likely block other browsers enabling working inside of Gmail, Google Calendar and Docs suite.

I don’t bet against (Fidji and her team) Atlas but this latest version shows how ChatGPT have released an alpha and its not ready.

When the battle comes for the home browser not just the work browser this is where Apple will have to wake up, Google and Apple’s to lose but my own use case with Gemini + AI Browser assistant reviewing tasks and sitting across my writings etc has added another layer of rigour which wasn’t possible, once users see these use cases away from Chrome is where another browser might just start taking meaningful marketshare.

If you would like to go deeper into AI and why this is just part of the battle: Here is my recent deck from my AI & Marketing presentation beyond the AI hype