MC.94: The Browser Wars Return. And This Time, AI Is the Weapon

Hello everyone,

The browser wars are back. After 15 years of Chrome's unchallenged reign, 2025 has unleashed four serious contenders—each backed by billions, each betting that agentic AI represents the inflection point that finally breaks the incumbent's stranglehold. But here's the uncomfortable truth: most of them will fail, not because their technology is inferior, but because acquiring users is brutally, systematically harder than building better software.

The Problem With Browser Acquisition

Let's start with the brutal math.

Chrome commands 68% of the global browser market—3.45 billion users—and that share is still growing. Firefox, once a genuine competitor, has withered from 50% to 3% despite superior privacy features and technical excellence. Why? Because switching browsers is friction-heavy: users must actively seek alternatives, download them, install them, change default settings, and rebuild their digital muscle memory. It's a multi-step process that most people never complete.

The distribution advantages protecting Chrome are almost unfair. It pre-installs on 71% of global smartphones (Android). Google pays Apple $20 billion annually and Mozilla $100 million+ to remain default search. Developers optimize for Chrome because that's where users are, creating a self-reinforcing cycle. Network effects compound: 250,000+ Chrome extensions create switching costs, and deep integration with Gmail, Drive, Maps, and YouTube makes leaving feel like abandoning an ecosystem rather than just changing browsers.

Most critically: the browser market doesn't reward better technology. Firefox proved this. It had privacy, speed, and developer love. It lost anyway. The lesson is harsh—distribution beats differentiation in markets where switching costs are high and user awareness is low.

This is the mountain AI browsers must climb.

The Four Challengers (And Their Odds)

Dia (The Browser Company, acquired by Atlassian for $610M) is betting on simplicity. Built on Chromium, it embeds ChatGPT into the URL bar, adds AI-powered tab comparison, and includes a Skills system for reusable workflows. The Atlassian acquisition provides enterprise distribution through 300,000+ customers. But here's the catch: it's macOS-only with Apple Silicon M1+ chips. No Windows ETA. No mobile. The addressable market is roughly 5% of Chrome's. They're playing in a sandbox.

Comet (Perplexity, $1.22B raised) is the most aggressive. It launched free globally in October 2025 after initially charging $200/month. The product is genuinely impressive—true agentic browsing where the AI navigates autonomously, integrates with Gmail/Calendar, and completes multi-step tasks. Perplexity processes 780 million monthly queries and has 15+ million MAU. They're discussing Samsung Galaxy pre-installation, which could be game-changing. But Comet tracks "everything users do online" to sell hyper-personalized ads, and the company faces copyright lawsuits from Forbes, Wired, the New York Times, and others. The business model is controversial; the legal exposure is real.

Genspark (Eric Jing, ex-Baidu/Microsoft, $160M raised) is the efficiency play. 2+ million MAU, $36M ARR in 45 days post-launch, 50 employees. The differentiation: 169 on-device AI models running locally with zero internet required. Privacy-first, no subscriptions, no cloud dependencies. The problem: Trustpilot ratings average 2.5/5 with complaints about billing and security. LayerX Security found a 90%+ compromised webpage execution rate. It's fast and efficient, but reliability concerns are serious.

OpenAI Atlas (launched this week) is the heavyweight. ChatGPT-native design where the new tab page starts with ChatGPT, not a search box. Agent Mode automates tasks. Browser Memories provide persistent context. The distribution advantage is staggering: 800 million weekly ChatGPT users as a ready-made audience. This is the one Chrome should fear—not because it's technically superior, but because it's backed by OpenAI's resources and user base.

The Uncomfortable Truth About "AI Differentiation"

Here's what nobody wants to admit: most AI browsers are solving the wrong problem.

They're adding AI features to browsers. But the real issue isn't that Chrome lacks AI—it's that Chrome lacks agentic capabilities. Users don't want AI bolted onto a browser; they want AI to replace the browser for certain workflows. They want to say "book me a flight" and have it done, not "use AI to help me book a flight."

The differentiation angle in the knowledge base is correct: browser acquisition is hard because it requires massive effort to get people to switch. Most browsers are Chromium-based, so differentiation is marginal. And AI browser adoption has been minimal so far—despite billions in funding and impressive technology, none of these products have achieved meaningful market penetration outside early adopters.

Dia is enterprise-focused but platform-limited. Comet has traction but faces legal and privacy backlash. Genspark is efficient but unreliable. Atlas has distribution but is still unproven at scale.

The Real Winner (Spoiler: It's Probably Google)

Here's the scenario nobody discusses: Google wins by doing nothing special.

Chrome already has 68% market share. Google can integrate Gemini, add agentic capabilities, and maintain dominance without disrupting its $150 billion+ annual search and ad revenue. The company has the resources, the distribution, and the user base to evolve faster than any startup.

The more likely outcome isn't "AI browser disrupts Chrome"—it's "Chrome becomes an AI browser and nobody notices because it was already installed."

But there's a second scenario worth considering: market bifurcation. Chrome remains dominant for general users (70%+ share). AI browsers carve out 10-15% among professionals, developers, and enterprise segments where agentic capabilities justify switching friction. The market doesn't consolidate around one winner; it segments by use case.

This is the uncomfortable middle ground: AI browsers succeed without disrupting Chrome, creating a sustainable niche rather than a wholesale replacement.

What Actually Matters

If any of these browsers achieve real traction, it won't be because of superior AI. It will be because of:

1. OEM pre-installation (Samsung-Perplexity discussions are the real story here)

2. Enterprise adoption with demonstrated ROI on agentic workflows

3. Regulatory intervention forcing Chrome divestiture (the DOJ case is ongoing)

4. Ecosystem lock-in through integrations that make switching painful in the opposite direction

The technology is table stakes. Distribution is the game.

The Bottom Line

The browser wars are back, but they're not being fought on the battlefield of features or speed. They're being fought on distribution, regulation, and the willingness of users to overcome switching friction for genuinely new capabilities.

AI browsers have a real shot—not at dethroning Chrome, but at capturing meaningful segments where agentic capabilities justify the friction of switching. The market is probably expanding rather than redistributing. Chrome remains dominant, but the future of browsing is genuinely uncertain in ways it hasn't been since 2008.

The irony? The browser that wins won't be the one with the best AI. It'll be the one that makes switching easiest.

Cheers,
Olivier

Like this newsletter? Forward it to a friend and have them sign up here.

Until next Thursday 🎉