There’s a lot happening in the browser world. Just weeks ago, the conversation centred on Chrome — and whether Google might be forced to offload it because of its overwhelming market share. Now, the landscape looks different. OpenAI has unveiled ChatGPT Atlas, Opera is pushing Neon, and Perplexity has launched Comet. The race to redefine how we access the web is on.
Browsers want to be part of your day from the moment you wake up to the minute you call it a night. It’s a scramble for attention and, eventually, revenue. If 2023 and 2024 belonged to AI agents, 2025 is shaping up to be the year of the AI browser. Every new player is trying to chip away at Google’s dominance.
A browser is, by nature, agentic — designed to act for you. The idea? A digital assistant that can handle your bookings, pull up research, or retrieve forgotten pages without a click.
ChatGPT Atlas: One tab to rule them all
ChatGPT Atlas, OpenAI’s new browser, revolves around its chatbot. It remembers previous conversations and searches and, for now, it’s available only on Apple Macs, with iPhone, Windows, and Android versions coming soon.
What makes Atlas different? OpenAI is trying to fuse ChatGPT into everyday browsing — to learn from the way you explore the web.
AI sidebars and agent-like tools are the new members in the browser arms race. Chrome has them. Perplexity’s Comet flaunts them. Atlas joins the fray. It’s still missing a few essentials — extensions don’t yet work — but since it’s built on Chromium, that’s likely a temporary gap.
Most people still use ChatGPT in a single browser tab, jumping between it and others. Atlas turns that one tab into an entire environment. In Atlas, pull up the sidebar and you’ll find your full ChatGPT history. In essence, it’s that familiar ChatGPT window, expanded into a complete browser.
In practice, you could have Safari open on one side for work and Atlas on the other for ideas, context, and conversation.
OpenAI’s reasoning is clear. As AI models become more capable, they need richer user context. Google already has that through Chrome — your accounts, your searches, your history, your habits.
OpenAI’s thinking is simple: If ChatGPT is to be a true digital personal assistant, it needs the same depth of understanding. Atlas aims to deliver that, building “memories” of where you’ve been and how you interact online. But users must turn on the memory feature for it to work.
There’s one hurdle: We’ve spent years fine-tuning our brains for SEO-style searching, but Atlas doesn’t think in keywords. It’s slower, more conversational — a different rhythm entirely.
Where it shines is recall. Ask it to find a page you visited while researching something, and it can sift through your history and context to retrieve it.
A generation on the verge of switching
ChatGPT now has more than 800 million users. Chrome still towers at around three billion, but the direction of growth matters. The younger generation is gravitating toward ChatGPT — not necessarily Atlas yet, but it’s only a matter of time.
They already chat with AI daily; folding that habit into a browser is the next logical step. For OpenAI, it’s a clever play to challenge Google where it lives.
And Google isn’t sitting idle. It’s weaving Gemini AI across its ecosystem, including Chrome. The competition is fierce, the stakes enormous — and for the first time in years, the browser wars feel alive again.
Reports of OpenAI working on a browser have been trickling in for 11-odd months, but it still feels like they rushed this out, because it resembles the Dia Browser with a few extra tweaks. It tries to focus on the agentic mode, which remains not very useful.
Where OpenAI has an advantage is if they announce within ChatGPT that users should download Atlas — then the browser’s popularity may pick up.
Google used to be good, but as the need to earn more ad dollars grew, everything got bloated with worse results, sponsored posts, and so on. At the moment, in the LLM era, we’re in this perfect little box where there are no ads and it gives you clean information. We’ll eventually reach the point where it becomes like Google — but in natural language. It will ultimately have sponsored results.
Many of these new browsers use Chromium — they look the same and fundamentally work the same way, with all the AI-related features built in. The question is: who can make AI features genuinely useful first? There’s nothing here yet that makes one go, I must use an AI browser now.
There is the web where you just go and find information. It’s like going through a bunch of Instagram Reels and wasting time. That part of the web is in decline. And then there is that part where you just talk to chatbots and get information that it picks up from the web. The part of the web that’s growing is the one that’s distributing applications. You don’t have a phone on you, no problem: Go on the web and order your groceries from the Amazon website. How will this look when a robot does it? There is pressure from agentic browsers, but let’s see if it survives 2026.