Perplexity’s Comet Browser: big bet or big mistake?
Perplexity’s founder, Aravind Srinivas, is all-in on AI browsers. OpenAI and Google Chrome are right behind him. But is building a browser the right move for Perplexity?
Remember the good ol’ days when you had to choose between Internet Explorer, Chrome or Firefox to surf the web?
Well now we have a whole new roster of AI browsers vying for your loyalty, and it’s mostly a new set of names: OpenAI, Perplexity, or old reliable Google Chrome.
Perplexity, the AI-driven search engine, has just announced their latest product splash: the Comet internet browser to take on Google Chrome. Perplexity’s vision for Comet is to create an internet browser that can think and act on your behalf, enabled by webpage and multi-tab context plus access to your personalized apps like email and calendar.
A few use cases to illustrate:
Book a meeting if you say “Schedule a meeting with John next week” by interacting with your calendar and email (with permissions, of course).
Create a walking route in Google Maps for you of the top tourist destinations in London
Build a grocery cart for you on Instacart…because manually selecting groceries on an app to be delivered to your doorstep was exhausting 😅
So…why is everyone building a browser?
Google and OpenAI are also racing to build AI-first browsers, with Google announcing features to integrate more AI-powered features into Chrome, and OpenAI announcing their AI browser is weeks away from launching.
The real reason Perplexity, OpenAI, and Google are all racing towards building browsers? It comes down to owning distribution and data.
Google needs Chrome to collect more data on its users, so that it can target ads better.
Browser users = more data = more effective ads.
So it makes a lot of sense why OpenAI and Perplexity would want to enter this arena, and why Google needs to defend this space.
Here’s a breakdown of each competitors strategy:
Perplexity’s Comet: high-risk, high reward?
In his interview with Y Combinator, Aravind Srinivas (Perplexity’s co-founder & CEO) that the Comet browser is their #1 focus. For Perplexity, they view the browser as the operating system of the future.
In his interview with YC, Aravind Srivinas outlined a few core product vision principles and reasons for investing in the browser:
The browser becomes an omni-box, capable of handling many tasks in parallel
The browser is the operating system for how we interact with AI and the internet
He speculated that the future of their business model could be:
subscription-based (what they do now)
usage-based (where users pay agents to execute a specific task)
commission-based (think affiliate links, but AI-mode where Perplexity could take a cut out of transactions executed)
ad-supported using cost-per-action instead of the traditional cost-per-click models Google’s ad business has relied upon.
Browsers are relatively difficult to copy compared to an AI-chatbot product, but he’s also fully aware of the competition and knows their only moat is speed & agility
When asked about competition from Google and OpenAI he said:
“If something is worth doing, it's only natural that everyone is doing it.”
Touché Aravind, touché.
“A browser is the dumbest thing you could build in 2025”
I was nodding my head at Aravind’s outline of Perplexity’s strategy.
Actually, I wasn’t even going to write about Perplexity’s browser announcement, because I thought the move was completely obvious and well, un-interesting…
🔥 until I came across a hot take from Chamath Palihapitiya (a Silicon Valley investor) and Travis Kalanick (the founder of Uber) on this week’s All-In Podcast.
Chamath did not hold back:
“Building a browser is an absolutely stupid capital allocation decision … A browser is the dumbest thing to build in 2025.”
Chamath pointed out that browsers will quickly become outdated, as they essentially handle rendering, security, and networking. In his view, these are all back-end “under-the-water” elements that users will soon no longer need to worry about in a world where agents can execute tasks for us.
Travis Kalanick, the founder of Uber, shared a similar sentiment. Kalanick believes that AI agents will soon leapfrog browsers altogether.
Consider the case of booking a flight: rather than navigating through a series of flight options, for example, an AI agent could simply book the flight for you, making the browser unnecessary.
So, is Perplexity throwing all its chips into the wrong game?
What Should Perplexity Do?
Travis’ and Chamath’s take gave me some pause and got me brainstorming some alternative product strategies for Perplexity.
Option 1) Leapfrog browsers entirely 🐸
Instead of focusing on a user-driven browser, Perplexity could conceptualize a browser-free internet.
Building a browser is no small task. And importantly, it’s rooted in the assumption that users want to browse the web to complete a task.
But do users actually want to browse the internet to complete digital chores and errands?
Perplexity could leapfrog the browser by conceptualizing next-gen AI agents that complete “the boring stuff” for you.
Instead of summarizing the best keyboards for my Macbook, just find me the best keyboard, tell me why it’s the best in 1-2 sentences, and help me hit buy.
Instead of summarizing hotel results, find the best hotel for my preferences and book it.
Instead of summarizing fun things to do in Paris, plan my itinerary for me and book all my activities/excursions.
The product itself could be presented in the form of a voice UI, chat UI, and integrated in a web app, native desktop or mobile app, or even a browser extension or “always-on” overlay like Cluely.
The form factor doesn’t matter as much as the bet that users don’t actually want to browse the internet to complete mundane tasks like booking appointments or travel plans. Users may want to browse the internet for fun, social, shopping for exciting purchases, or news-gathering.
Most people want to complete “digital chores” like booking a flight, communicating with a colleague, or buying a gadget in the fastest, easiest way possible.
And that’s where a browser-free lily pad awaits. 🪷
Here’s what an even bigger leap could look like:
To be fair, Perplexity probably has this on their radar, and are likely planning to make the second leap when the consumers and the technology is ready (remember, products and services are still playing catch-up to enabling Model Context Protocol or even optimizing their website to be accessible by agents).
Perplexity absolutely could make the second leap after building a browser, but they risk sinking unnecessary resources into a dying form factor.
And you may argue say “well, they’re a well-funded start-up, backed by NVIDIA and Jeff Bezos! They can do everything!”
… but considering the latest war for AI talent (looking at you, Zuck), product decisions need to be more efficient than ever.
Let’s say they choose the 2-leap approach: build an AI browser first, build the no browser future second. By the time they’re ready to move on, they could have already leapfrogged their competitors by skipping the browser project altogether.
Or worse: their competitors move onto building a browser-free internet before they do.
Option 2) Build a kick-ass browser, primed for sale to Apple.
Perplexity could focus on building a standout browser, and then license or sell their technology to Apple as some early reports have speculated.
And the more I think about this match, the more I think this could be a match made in heaven. 💗
Think about it: Apple has been slower to ship cutting-edge AI features, Siri is still as clueless as it was in 2010, and the Safari browser has struggled to keep pace with competitors. Perplexity’s strength lies in its product design and their focus on user privacy, aligning well with Apple’s own stance on data protection. With its robust ad blocker, leadership in conversational UX, and privacy-centric features like local data storage, Perplexity could create a browser that complements Apple’s ecosystem (and bridges the gap in their AI missteps).
Option 3) Focus on industry-specific verticals
Alternatively, Perplexity could drop the browser project altogether and focus on industry-specific enterprise use cases where they already have a competitive advantage.
The perfect verticals are ones that require specific expertise and are heavy on info-visualization, like travel, medicine, financial data, and academic research.
One particularly ripe area for disruption? Financial data. As Chamath Palihapitiya suggested on the All-In Podcast, Perplexity could replace Bloomberg with a more intuitive and AI-powered platform. Which, they’ve already done, but the point is more around doubling-down in this area:
“For Perplexity, I think their path to building a legacy business is to replace Bloomberg. Everything that they’ve done in Financial Data has been excellent. As someone who has paid $25,000 to Bloomberg for many years, the terminal is atrocious.”
And to Perplexity’s credit, they’ve already invested into enterprise verticals like Finance and data integrations for academics like Wiley. Their advantage around information visualization, data privacy, and conversational UX makes them well-suited to continue drilling down into enterprise use cases that require niche data integrations.
WWYD: What Would I Do if I Were Perplexity?
Of course, it’s a lot easier speculating from the side-lines, but it’s a fun thought experiment for any Product Builder.
My take: I’m all about the power of niche and industry specialization. If I were in Perplexity’s shoes, I’d go all-in on Option 3 by focusing on industry-specific use cases and building enterprise-grade products with gnarly data integrations. This approach would allow Perplexity to win because giants like Google and OpenAI will overlook those niche use cases while seeking world domination.
And as the old saying goes, the riches are in the niches.
What Product Builders Need to Know
Perplexity’s push to build an AI-internet browser is bold, but they may be investing resources into a user behaviour that is on the decline. Additionally, Perplexity is up against some formidable competitors, and they may be taking on too broad of a market for their size & scale.
This week, I challenge you as a Product Builder to reflect on two key questions:
Am I future-proofing for my user’s behaviour to change over time, especially with more AI-fication?
Should I be narrowing my team’s efforts to a more specific niche?
💬 And lastly, what would you do if you’re in Perplexity’s shoes? Let me know in the comments!