Everyone who is anyone in tech wants a browser. Google has Chrome, Apple has Safari, Microsoft has Edge. AI provider Perplexity recently offered to buy Google’s Chrome browser and the Brave browser, among others. Now OpenAI has released its own browser, ChatGPT Atlas, which is built around ChatGPT and offers a context-aware, task-completing assistant integrated directly into your browsing experience.
It’s very similar to what alternative browser company Opera released earlier this summer: an agentic browser. Except that it’s from the largest and most innovative AI company on the planet.
An AI-driven browser is super-useful for people. It knows you, your context, and it lets you do work in the same browser window that you’re using to find data and resources. It knows what you’re seeing on a web page, and can summarize, interpret, or fact-check it. It can recall information from the web page you browsed last month that you can’t quite remember right now, and it can suggest new resources, information, or tools that will help you to accomplish the thing you’re working on right now. It can also act as your agent, shopping for the best deal, buying products, scheduling appointments, and, theoretically, accomplish any other task you can do online.
All super-useful.
It’s also super-helpful for big tech companies that want to expand their power and position. An AI-driven browser, or agentic browser, will help them know their users, capture huge quantities of data about populations, profit from people’s purchases, sell ads against their activity, and generally tighten the silken ropes that connect people to them as a customer.
There are at least seven reasons why OpenAI made the move to make the ChatGPT experience a browser experience:
- Better user experience: As mentioned above, offering the browser and the AI together in a single tool means less friction: no copy/paste between windows, more context and a smoother workflow. All of that makes ChatGPT’s core product better.
- Platform lock-in: If you use the Atlas browser, OpenAI becomes a core part of your daily life. You probably open a browser more frequently than the ChatGPT app: now you’ll do both at the same time. That means you’re a more secure OpenAI customer: you’re locked in.
- Traffic and attention: When OpenAI owns your browser, OpenAI captures a larger share of your attention and web navigation time. That, over time, can be monetized in multiple ways, including but not limited to advertising.
- New data and signals: All the big tech companies want to get more data. Having richer signals about what users are doing, what tasks they’re tackling, what websites they visit allows better personalization and more tailored models. It also future-proofs big tech companies against future technological and business changes, because they see what huge swaths of the population do at early inflection points, and can then acquire or hobble potential competitors.
- Expansion of agentic capabilities: Owning the browser environment means OpenAI can unleash advanced agentic features on anything that is possible to do on the web: work, commerce, chores, research … the list is endless. All of that is hard or impossible in an app that doesn’t have native access to the richness and diversity of the open web.
- Disrupting the search and advertising model: At some level of size, tech companies all start to compete with the giants, like Google, Apple, and Microsoft. Shifting users to ChatGPT Atlas ensures they’ll spend less time on Google’s and Apple’s browsers, less time clicking through search results, and less time viewing ads. That means OpenAI becomes the new interface for web interactions, and Google gets disrupted.
- New business models: If OpenAI is your new agent for product research, purchase, ticket-buying, and more … there’s room for additional monetization. Agent-mode commerce, subscriptions, and premium tasks are all pathways to taking a larger chunk out of your wallet beyond a $20/month subscription.
On a larger competitive level, this moves OpenAI up the platform value chain. They don’t have devices and they don’t have operating systems (yet), which are the ultimate positions of power in the super-lucrative tech ecosystem. But since we spend so much of our time in browsers–at least when on laptops and desktops–this gives OpenAI a meaningful shot at a powerful platform position of their own.
Also, OpenAI can see what Opera did. They can project what Google might do with Gemini and the Chrome browser. Defensively, you could argue that OpenAI pretty much had to follow this path.
There’s a school of thought in tech that AI is becoming the new interface. In other words, our primary mode of using technology is through AI tools.
If that’s true, Google can make Gemini the default interaction mode for getting information, accomplishing tasks, and opening environments on Android. Apple can try to build a better Siri and put it front and center on iPhones.
OpenAI has to prepare for a world in which the companies that own the platforms that mediate its access to its customers do everything they legally can to limit ChatGPT’s growth.
This is a smart first step. There’s probably more coming.
