From Perplexity's takeover offer for Chrome, let's have an in-depth discussion about the new round of the AI browser war.

In Silicon Valley, many people think Google is falling behind.

In mid - August, Perplexity, an AI - search unicorn, said it would acquire Google's Chrome browser for $34.5 billion, almost twice Perplexity's own valuation. Previously, Perplexity had also said it wanted to buy TikTok... Is this round of browser acquisition battle just pure hype, or is there really a possibility of a snake swallowing an elephant?

By the end of August, Anthropic newly released a browser product, Claude for Chrome, joining the battle. With the booming and enhanced capabilities of the AI agent track, the term "AI browser" has made "browser", a word that no one has discussed in the past few years, suddenly become a tech darling in Silicon Valley again.

Is the battle of AI browsers really coming? What new opportunities does it imply? How will the trillion - dollar search business be disrupted?

In this article, we will discuss the new round of browser battles with Guru Chahal, a partner at Lightspeed Venture Partners (the lead investor of Anthropic), and Howie Xu, the Chief AI and Innovation Officer of Gen Digital.

01

What is an AI Browser?

In the last browser war, Firefox, Google Chrome, Microsoft Edge, Safari, and Opera were the main competitors in the market. The market shares of these major players have become very stable in recent years, until the emergence of AI.

Chen Qian:

Please help us define what an "AI browser" is.

Guru Chahal:

I think it goes back to a very simple premise. In the past 20 or even 25 years, the way humans or users interact with computers has systematically shifted from the operating system to the browser. So, everything that used to happen at the operating - system level is now happening at the browser level. We used to rely mainly on the operating system for security, observability, data collection, and running applications, but now all these are done in the browser.

Therefore, my initial hypothesis is that the main mechanism for AI to interact with tools originally designed for humans will also be the browser. The browser thus becomes an interesting "fortress" where you can collect data, automate tasks, ensure the security of AI - system interactions, etc. This is an investment framework we believe in.

Actually, we have already invested in this field. Just as I talked to Howie before, a few years ago, we saw some very good teams working on enterprise browsers. And companies like Lightspeed are willing to use these browsers because our IT and security teams can control what users can and cannot do in the browser.

So we invested in Talon, a company that was later acquired by Palo Alto Networks, and we were the lead investor at that time. We witnessed this market grow very fast, and today it has become a very large enterprise - browser market.

And I think the core question we need to discuss today is: Can AI be another reason to drive the birth and growth of a new category of browsers? So I'm really looking forward to this discussion.

Howie Xu:

Yeah, you know Talon was acquired not because of AI, right?

Guru Chahal:

That's right. It was for cybersecurity reasons at that time.

Howie Xu:

There were other reasons too, but one interesting thing we can see is that we haven't seen browser - technology acquisitions for a long time. Suddenly, the browser has become an interesting technology again.

We've known each other for so many years and talked about countless topics, but we've never talked about browsers in the past 20 years. When I saw the article Guru published, I thought, "What? He's interested in browsers?"

Now you explain that you were originally interested in enterprise browsers, which is understandable. But now you're interested in AI browsers. Why is that?

Guru Chahal:

This actually stems from two key understandings and a basic belief.

At Lightspeed, our basic belief is that large models and AI as a whole represent a new type of ability that can bring in "cognitive labor" and make cognitive labor exploitable by technology. This is the first time in the history of technology that machines can process semi - structured language, understand language, and operate like humans.

So the question is: Where are the boundaries of this ability? On the one hand, how smart can the model become? We think we're just scratching the surface, and the model will continue to get stronger. We're also one of the largest investors in companies like Anthropic.

On the other hand, what limits the development of these models? If we look closely, there are actually two points: First, it's the model's ability to absorb data, learn, and perform better in an enterprise environment, that is, observation and learning; second, it's the model's ability to actually execute actions, to do things, that is, "action/execution".

And there is a technology that happens to be a universal window for observation and action, and that is the browser.

Because think about it, in the past few years, even just before you came to this interview, about 95% of the things you did on your computer were done through the browser. This means that if I were the browser, I would know exactly what you're doing and which software you're interacting with, and then I would be able to interact with these software in the same way.

Of course, because there are still desktop applications, this number won't be 100%, but if you're looking for an entry point, it must be the browser. I think this is why the browser has always been a very strategic "digital real - estate", and this is why Google did the entire Chrome project.

Think about it, the only reason Google did Chrome was that they wanted to control search, but the browser itself is such a strategic entry point. So, in the AI era, its importance is several orders of magnitude higher than in the past. The browser has regained attention because it solves two key problems at the same time — it allows you to collect data in a passive and continuous way, and it can act both as a collaborative co - pilot to help you take continuous actions, and as an autonomous agent to perform work for you.

02

Perplexity's Ambition

and Google's Innovation Dilemma

The news at the beginning that Perplexity wanted to acquire Google Chrome, although it sounds a bit more credible than Perplexity's previous statement of wanting to buy TikTok, the valuation and scale of Chrome are definitely beyond what Perplexity can handle. Besides being a public - relations strategy, it also sends a signal: Perplexity believes that AI browsers will be a track they will focus on in the future.

In July this year, Perplexity released an AI - browser product, Comet, based on Chromium. Currently, it supports Windows and MacOS, and there are plans to expand to Linux, Android, and iOS systems in the future.

Currently, this AI browser is only available to high - end Perplexity Max users who pay $200 per month. Other users can only join the waitlist to apply.

According to public information, Comet is a browser product with an embedded AI assistant that can provide continuous help in the sidebar without the need to switch tabs frequently.

It supports various tasks such as web search, content summarization, meeting scheduling, email writing, etc. At the same time, it can also let the AI agent automatically click, input, and submit forms on the web for users, such as completing the entire shopping process.

It also supports using functions to let the AI refer to the content of the currently open tabs, maintain the context - relevance of questions, highlight words and sentences on any web page to get instant explanations and extended thinking, and support multi - angle questioning.

Finally, it supports everything from summary content, research, translation, to schedule management and multi - platform linkage, such as email, Slack, social media, etc.

Although such an AI browser is still a relatively early attempt, limited by the capabilities and technology of the AI agent, it does show the potential of future traffic entry points.

Chen Qian:

A few days ago, I was chatting with Howie. There was news that Perplexity unexpectedly offered $34.5 billion to acquire Chrome. What was your reaction when you first saw this news?

Guru Chahal:

We didn't invest in Perplexity, so we don't have more information than you two. But what I'm going to say is that saying such things and doing such things is risk - free. Even if it doesn't succeed, it can get a lot of attention. After all, Perplexity is a consumer - facing company, so anything that can get more attention is worth doing.

Howie Xu:

Attention is all you need.

Guru Chahal:

I like this statement, Attention is all you need. So what's the harm? There's no harm. And in the extremely unlikely event of a benefit, there might be a situation of forced or voluntary divestiture. In that case, why not strive for it? So I think this is a very smart and shrewd move.

Howie Xu:

Although I'm the Chief AI Officer of a listed company now, even without this position, I would strive for everyone's attention. So I completely agree with Guru: There's no harm in this. Even if you get nothing, the worst - case scenario is getting a lot of attention.

I think Google is in a very interesting period. On the one hand, their search business is being impacted. Personally, I use search much less frequently than before, and I think this will become more and more of a fact in the next five to ten years. But at the same time, Google is also a provider of AI, and they have a large number of excellent engineers.

In the browser war, Chrome has a market share of over 70%, clearly the giant in this market. The browser war is real. In ten years, 70% of people will not use Chrome as their default browser, and it may even be far from this proportion.

This is because Google's annual revenue is about $400 billion, and a large part of its search revenue actually comes from Chrome. So any adjustment Google makes to Chrome will immediately have a huge impact on its revenue.

The general market consensus is that of the $400 billion, about $80 billion comes directly from Chrome. So if I change the user interface of Chrome, I'm actually changing the revenue of $80 billion. How many companies in the world have a revenue of $80 billion? So this is a big deal.

I think they are experiencing what is called the "Innovator's Dilemma". They're not unaware of what a native browser should look like, and they understand the logic we mentioned, that is, the browser should help users browse or enter the digital world more easily, but they can't make too many changes. From a first - principles perspective, this is actually very obvious.

When I communicate with people inside Google, they tell me straightforwardly: "My hands are tied. I can't do much. I know what to do, but I can't do it." So in a sense, we're going through a very interesting stage: The industry leader is about to be phased out. This is my assumption for the next ten years.

So who will be the next? Obviously, there are Comet, Dia, and Norton Neo, but I think we're all still in a very early stage.

I joined Gen Digital as the Chief Innovation Officer about a year ago. A year ago, I started researching the browser project. We all observed internally that the browser urgently needs the addition of AI, not just adding a button for summarizing pages. That's not a browser at all. That's just an assistant.

I think the world needs a native browser. We must rethink the browser