OpenAI didn’t plan to make a browser. Not at first, anyway. But ever since the company launched ChatGPT, it started to see lots of users do the same thing. “You’d have this tab open on chatgpt.com, and you’d be working on something else somewhere in your browser,” says Adam Fry, a product lead on the ChatGPT team. “And people are doing this behavior where they’re copying and pasting back and forth with ChatGPT. Hey, here’s what I’m working on, can you help me with this?”
Eventually OpenAI came to a realization, the same one that has swept through the tech industry over the last couple of years: Browsers are the secret to everything. Your web browser is practically guaranteed to be the most-used app on your computer. It’s definitely the only one that can access your email, your bank accounts, your confidential work spreadsheets, and your doomscrolling platforms of choice. It’s the richest source of information about you, your activity, and your life. It is, as Fry puts it, “the operating system for your life.”
For any general-purpose AI assistant to work, it needs to be where the users are. And the users are in browsers. So OpenAI made a browser, called ChatGPT Atlas. It’s a simple app, with a row of tabs at the top and a big ChatGPT text box every time you open a new one. You can use Atlas to ask questions about the tab you’re currently using or collate data across tabs. ChatGPT can fill out forms for you across the web, and it will even attempt to use Instacart to buy you groceries if you ask. You can use ChatGPT to query your entire browsing history, or use what it’s learned about you in your other work and conversations.
This is the pitch for the AI browser, a new category of web browser that has rapidly emerged over the past few months. You can do similar things, inside a similar-looking app, with Comet, Perplexity’s browser. Or with Dia, The Browser Company’s AI-first browser. Microsoft’s Edge is turning into an AI browser, and even Chrome has gotten a big injection of Gemini over the last year, too. All these browsers have roughly the same ideas, built on roughly the same tech — and they’re all hoping their bots are so helpful you might think about switching.
You probably spent the last decade or so not thinking much about your web browser. You use the one that comes with your device, or more likely, use that built-in one to download Google Chrome. Then you never switch again. Web browsers show webpages, and mostly do it fine. They’ve seemed like a solved problem, or at least one not worth solving any further.
At the same time, though, companies around Silicon Valley have spent years trying to figure out the next big platform. After a generation ruled by the duopoly of Android and iOS, companies everywhere are desperate to stop paying app store royalties and start getting more control of their fate. First they tried to make the metaverse happen and made a lot of noise about Web3. Then, one by one, as AI captured the imaginations of tech companies everywhere, they trained their sights on one of the oldest, least exciting, and most important apps we all use all the time. They’re all coming for your tabs.
Image: Cath Virginia / The Verge
In the annals of the browser wars, the conflict has spiked at two important moments in the development of the internet. The first was, of course, at the beginning of the internet when the whole idea of a “web browser” became a thing. That was in the late 1990s, when an academic project known as Mosaic turned into a browser called Netscape Navigator. That became a huge hit and opened millions of people’s eyes to the possibilities of this thing called the World Wide Web. Then Microsoft built its own browser, Internet Explorer, and used every ounce of its considerable might to bully Netscape out of the market. That turned into a landmark antitrust case, which left both Netscape and Microsoft somewhat debilitated in their efforts to win the web.
In the mid-2000s, with the browser makers stuck in litigation, the browsers themselves languished even as the web’s population grew fast. Two new combatants picked up the fight. There was Mozilla Firefox, a project born out of Netscape’s tech (and with the help of a lot of its team). And there was a team inside of Google, led by a budding executive named Sundar Pichai, that built a browser called Chrome.
If the first browser war had been about giving people access to websites, the second was about giving you more to do with them. Instead of static pages designed to be downloaded, developers were using new standards and tools to create interactive applications — Gmail, Facebook, and other platforms were showing off just how much you could do with the web. Pichai and the Chrome team focused relentlessly on making these applications load fast and work smoothly, and eventually helped turn desktop computing from a series of apps to a series of browser tabs.
Google won the browser wars in about 2012
For all intents and purposes, Google won the browser wars in about 2012. Statistically speaking, you’re probably reading this in Chrome, which has approximately 4 billion users and accounts for between two-thirds and three-quarters of the entire browser market. And for more than a decade, it has had no real competition. Microsoft rebooted its browser plans with Edge, Firefox continued to develop, Apple still ships Safari, and a few enterprising startups have tried to compete, but nothing has made much of a dent.
In the last couple of years, though, a few companies have become convinced that Chrome is vulnerable. For one, the regulatory environment has shifted in a big way. Choice screens are becoming common on smartphones, giving browser makers a chance to hawk their wares to new users. Google is now embroiled in the exact kind of never-ending antitrust litigation that tends to make companies slow. Add in the fact that it’s actually pretty easy to build a browser these days, since practically everything is based on the Chromium architecture, and it’s the perfect time for the competition to pounce.
We’re also potentially about to once again change the way we interact with the web. If Browser War I was about webpages, and BWII was about web apps, BWIII could be about web agents. The AI companies all believe we’re about to spend less time typing URLs and more asking chatbots to do things on our behalf. If any of that is ever going to work, you’re going to need an AI assistant that has access to all of your information and every app and service required to help you out. The web is the only place all that exists, and the browser’s the simplest way to get access to all of it.
Image: Cath Virginia / The Verge
There are three reasons a browser might be valuable to a company that believes in an AI-centric future. The first is that your browser contains a vast trove of data about you. It always has — another reason Google made a browser is to collect data on everything you do outside of Google products. Your browser knows everywhere you go, and everything you do, at an unbeatably specific and frankly somewhat terrifying level. This data is useful to businesses for lots of reasons, but the easy money’s in the ads. “We want to get data even outside the [Perplexity] app to better understand you,” Perplexity CEO Aravind Srinivas said on TBPN earlier this year. He said the company planned “to use all the context to build a better user profile, and through our Discover feed we could show some ads there.” If you’re looking for huge sums of personal information, you basically can’t beat browsing data.
The second is that it’s a powerful app platform: as soon as you log in to your bank, your email, or your Instagram, your browser stores a cookie that gives it nearly unfettered access to that app. If you’re trying to stand up an AI agent that can buy you groceries, book you flights, or help you with homework, you need a tool that can both access and interact with websites alongside the user. “If you put browsers in chatbots,” says Brave CEO Brendan Eich, a veteran of all three browser wars, “you get these crappy little web views that don’t act like a full first-class browser. You want a chatbot in a browser.”
The third, and maybe the most valuable, is that your browser contains the most important input system on the internet. In many ways, Chrome exists because Google wanted to make it outrageously easy to do Google searches. If AI interactions are going to usurp Google searches, they have to be that easy. “The most valuable thing in this new world,” The Browser Company CTO Hursh Agrawal told me earlier this year, “is the fact that the browser owns CMD-T and the omnibox, because that’s the single entry point into your computer where you express intent — it’s the most-used text box on your computer.”
“It’s the most-used text box on your computer.”
Add it all up, and of course any company betting on all-encompassing AI should build a browser! The theory is a good one, but none of it matters unless there is actually some huge change coming to the way we use technology. Betting you can beat Chrome has been a fool’s errand for 15 years, and it’s not at all certain that the behavioral or technological changes these companies are betting on are actually going to appear. Agentic AI largely doesn’t work, and it may not anytime soon. There are plenty of players who would prefer it never does.
It’s also not clear whether people will actually give their lives over to AI en masse. LLMs and agents make the web more opaque, more automated, and potentially more prone to lots of new problems. “If you recommend me a pair of shoes,” says Anthony Enzor-DeMeo, the general manager for Firefox at Mozilla, “do I know that that’s because that’s the best shoe for me, or because there’s, like, a signed deal for Nike?” Many security experts are worried about the possibility of prompt injection, by which bad actors subvert and change an AI model’s process for their own benefits. And, as Eich puts it, “it’s just a matter of, do you really want OpenAI getting all your queries and prompts?”
While all this supposed revolution unfolds, Chrome just set a market share record. If there is even a race here, Google is starting with an unfathomably huge lead.
Let’s just say this does all play out according to the wildest dreams of the upstarts and AI companies. In the next few years, everybody reevaluates their browser choice, and the era of Chrome’s supremacy ends dramatically. The most immediate change will be the big one: The Google search bar is replaced nearly everywhere by the AI model of your choice. This would be a crushing blow to Google’s business, and to the tight search / browser combination that has ruled the web for years. Google was the portal to the web; now it’s not.
You’ll also quickly start to see some AI-powered features in the browser itself: More than one person I spoke to is working on using LLMs to organize and sort your browser tabs, and everyone seems to love the idea of being able to ask questions or do searches through your browser history. Give it a few years, and even the basic shape of a browser window — a row of tabs, an address bar, bookmarks — might start to change.
After that, our relationship with the web starts to change as well. When we’ve had browser wars before, it was because the web was becoming more powerful and more useful. But if you’re OpenAI or Perplexity or any other bot-first browser maker, you may not have much interest in making Google Meets run more smoothly in your browser. The ultra-popular text box might matter much more to these companies than the vast expanse of the web. They might want to make it hard for you to leave the text box.
Everyone I spoke to for this story said they believe in the web and care about keeping it open and accessible. Nearly all also said they worry about other companies that don’t feel that way. “We’re used to the web just being accessible,” Enzor-DeMeo says, “but now we’re starting to see browsers with subscriptions.” He says Mozilla is eager to integrate AI into Firefox, but will do so in a way that both gives users control — over the models they use and how their data is collected — and preserves the open web.
But what if you’re Parisa Tabriz, who runs the Chrome team at Google? You already have billions of users, who mostly use your browser to go to websites. Your browser is a crucial entry point to Google Search, the best business in the history of the internet — and if you help kill the web, you’re helping kill Search. But you’re also being pressured to adopt AI, to prove you’re not behind and that you won’t end up like Internet Explorer.
When I put this question to Tabriz, she acknowledges the tension. “I think we are probably moving more deliberately and thoughtfully,” she says, “and I think AI opens a ton of opportunities but presents a lot of risks, too.” Like others in the space, she’s concerned about the security risks in agentic browsers, which can be derailed by bad prompts or sketchy behind-the-scenes dealings. But she’s as bullish as anyone on the team’s recent integration of Gemini into Chrome. “The web is awesome, browsing is awesome, and I see it as just this next computing shift: with browsers and AI we’re going to be able to do things we could only dream about, in terms of helping people get things done.”
Right now, Chrome is still the dominant force in browsers, antitrust bruises and all. But the browser wars are back in a big way. These are the wars for the very soul of the browser. There are more players than ever, with more divergent ideas about not only how browsers look and feel but what they exist to do. To the winners will go the most important text box on your computer — and the right to decide what it means to browse in the future.
