This past week, Perplexity made an unsolicited bid to buy Chrome from Google for nearly $35 billion. This comes despite the company’s own attempt at building a new web browser, Comet. Most people dismissed the bid as a marketing stunt, an attempt to highlight Google’s antitrust problems, and perhaps even force a sale. Others took it as nothing more than a joke.
The bid is unlikely to go anywhere, but there’s no denying the sudden resurgence of interest in web browsers, both on the consumer side (e.g. Comet, Dia), and on the commercial side (e.g. headless, automated browsers for AI agents). Startups and larger companies are working on new, AI-native browsers, and for the first time in a long time, I do think that the way in which we “consume” the Internet is going to drastically change over the next few years.
The “good ol’ days” (late 90s - early 2000s)
I’m old enough to remember when the Internet was mostly a collection of text, just articles and information. You started your browsing experience on Netscape or Internet Explorer, and relied on Yahoo or MSN as your gateways to the world wide web.
Then came Google (and always-on, high-speed broadband internet), and it became easier than ever to ask questions on the Internet and be pointed to exactly the information you needed, exactly when you needed it. Gmail and Chrome followed, and most people haven’t changed their browsing habits since then.
Also, over the past 2 decades, we as consumers became accustomed to “doing things” on the Internet, not just using it for information. Things like booking airline tickets and hotels, paying bills, buying stuff, making trades, ordering groceries, and consuming gobs of content.
At work, we shifted from desktop-native software distributed on CDs to the cloud and web-based SaaS applications. Enterprises today pay millions of dollars per year for 3rd party software accessed purely through a browser.
In general, the browser became the operating system of the Internet, taking it from just a collection of text, to a place for consumption, work, and commerce.
All that and more, in your hand (2010s - early 2020s)
I got my first iPhone (the 3GS) in 2009, about 2 years after the first model launched. At the time, I was one of the few in my circle of friends to have one. Fast forward 15 years later, and nearly everyone has a smartphone. The Internet is always available. You can check emails, answer Slack messages, browse social media, book tickets, and buy things, from anywhere, at any time.
But the iPhone didn’t just give us a portable Internet browser, the App Store led to an explosion of custom Internet experiences, tailored to our needs. Want to access your bank balance? There’s a native app for that. Want to book tickets to a Lakers game? There’s a Ticketmaster native app for that. Each with its own branding, UI/UX, consumer experience, data, and company behind it.
My first iPhone had less than 10 apps. My phone today has 100s, one for each airline, restaurant, etc. I’ve interacted with, many just used once or twice. In this “mobile era”, the phone (iOS and Android) itself became our browser, offering custom experiences for nearly every interaction on the Internet.
Return to text (late 2022 - now)
I’ll be honest, I didn’t get the hype around ChatGPT when it first launched. It wasn’t until last year when I started using it more often. Nowadays, if I’m seeking information, I find myself first chatting with ChatGPT, rather than Googling. And even when I do use Google, the AI/LLM-based summaries appear at the top of results and continue to get better. So it’s no surprise that search traffic to content sites has cratered since ChatGPT became more mainstream.
Is LLM-search perfect? Not at all, it still makes mistakes and LLMs often hallucinate. Can ChatGPT tell me my bank balance? No, at least not yet. Using products like ChatGPT feels like a return to the text browsing Internet experience of the late 90s and early 2000s. If you’re looking for information or looking to learn, you can chat and prompt your way to learning. And with the newer deep research capabilities, you can let the system autonomously prompt itself into producing pretty good summaries.
What happens next?
Is chat going to be how we exclusively use the Internet now? I don’t think so. I also don’t think we’re going back to the mobile era of bespoke, native applications built and managed by teams of developers.
Rather, I think the technology now exists to support a new type of browser. A browser built on the same foundation as today’s browsers, with the ability to fetch and aggregate data from any server or “legacy web application,” but with the added ability to deliver completely unique browsing experiences to the end user.
For example, if I want to access my bank balance or investments, I shouldn’t have to log in to each account manually, or use a dated aggregator (RIP Mint). My browser, under my direction, with access to my credentials and accounts, should be able to build and render a custom budgeting/investing experience for me on the fly, one that is better suited to my workflows and preferences.
A true challenger to incumbent browsers will change what it means to browse and use the Internet, allowing for nearly infinite customization and personalization of the browsing experience. It will probably generate its own HTML and JavaScript. It may look more like a Claude Code/Lovable/Cursor hybrid than today’s Chrome. As code generation becomes faster and cheaper, it’ll be interesting to see how companies react. Do they become more API-centric and put restrictions around their data and platforms, or do they continue to invest in end user-facing web applications to control the user experience?
It’s unclear where things are headed, but it’s a fun time to explore and build. I’m not sure if Perplexity’s bid to buy Chrome is legit or not, but it’s certainly no joke. With OpenAI also reportedly prepping a new browser, a new browser war is brewing.