⚡ Quick Take

New analysis of over 2 million user sessions suggests the AI landscape is fragmenting. While ChatGPT has long dominated the conversation, recent data indicates that rivals—namely Microsoft’s Copilot and Anthropic’s Claude—are capturing the lion's share of new user discovery and growth. This isn't just a popularity contest; it's a signal that the AI war is moving from a battle of pure model capability to a strategic fight over distribution, integration, and product-market fit.

Summary: An analysis sourced to Google AI, covering two million user sessions with large language models, finds that Microsoft Copilot and Anthropic's Claude are currently experiencing the fastest growth in user discovery. This challenges the narrative of a single dominant player and points to a more competitive, multi-polar market. From what I've seen in these kinds of reports, it's a reminder that no lead is ever truly safe in tech.

What happened: The dataset reveals a shift in momentum at the top of the user acquisition funnel. While not a measure of total market share, the rate of new session growth for Copilot and Claude suggests their strategies for reaching users—whether through massive ecosystem integration or targeted value propositions—are proving effective. It's like watching the underdogs finally find their stride.

Why it matters now: For the first time, we have a large-scale data signal that the AI market's "first-mover advantage" may be eroding. This forces a re-evaluation of what drives AI adoption: is it the best model, the best marketing, or simply the most convenient access point? The answer appears to be a complex mix, with distribution emerging as a key weapon. But here's the thing—it's making everyone rethink their bets.

Who is most affected: Developers and enterprise leaders must now look beyond raw model benchmarks. Choosing an AI partner is increasingly about ecosystem lock-in (Microsoft), specialized capabilities (Anthropic), and future distribution channels, not just API performance. Have you ever weighed those trade-offs in your own stack? It can feel like choosing sides in a quiet revolution.

The under-reported angle: This data measures discovery, not retention. It captures the moment a user starts a session, but reveals nothing about engagement depth, task completion, or long-term loyalty. The rapid growth could be driven by Microsoft's aggressive bundling and default placements, creating shallow "drive-by" usage that may not translate into a durable user base. Plenty of reasons to tread carefully here, really.

🧠 Deep Dive

Have you wondered if the AI boom is starting to look a little less like a solo sprint and more like a crowded relay? The core finding from the 2M-session analysis is simple: the AI growth story is no longer monolithic. While ChatGPT established the standard for conversational AI, the data suggests that Microsoft's Copilot and Anthropic's Claude are now leading the race for user acquisition. This marks a pivotal shift from a technology-centric battle to a market-centric one, where strategy and distribution channels are becoming as important as the underlying model architecture. I've noticed how these pivots often catch the industry off guard.

The growth engines for these two challengers appear fundamentally different. Copilot's ascent is a case study in distribution as destiny. By integrating its AI directly into Windows, Microsoft 365, GitHub, and its browser, Microsoft has turned Copilot into an ambient utility. For millions of users, discovery is not an active choice but a passive encounter within an existing workflow. This strategy lowers the barrier to entry to zero, effectively drafting users into its ecosystem—and it represents a powerful advantage that pure-play AI companies cannot easily replicate. That said, it's the kind of move that feels almost inevitable in hindsight.

Anthropic's Claude, in contrast, seems to be growing through product-market-fit precision. Without a consumer operating system to leverage, its growth is more likely tied to its reputation for handling long documents, its constitutional AI safety framework, and its appeal to enterprise users seeking a reliable and steerable alternative to OpenAI. This suggests a strategy focused on winning specific, high-value use cases—like legal document analysis, code review, and corporate R&D—rather than capturing the entire consumer market at once. It's targeted, almost surgical.

However, a critical gap in the narrative is the distinction between discovery and retention. The analysis masterfully captures the top of the funnel but leaves the most important questions unanswered: Are these new users sticking around? Are they completing meaningful tasks or just kicking the tires? True market leadership isn't just about attracting eyeballs; it's about creating indispensable tools that foster sustained engagement. The current data offers a powerful signal on user acquisition trends but provides no visibility into the deeper metrics of activation, retention, and monetization that will ultimately determine the market's winners. The methodology's lack of transparency on what defines a "session" or how bots are filtered further complicates direct comparisons, reminding us that we're viewing the race through a specific, and potentially blurry, lens. It's data we can learn from, but not quite bet the farm on.

📊 Stakeholders & Impact

✍️ About the analysis

This i10x article draws from independent analysis based on third-party reporting of a large-scale dataset, enriched with semantic analysis of competitive coverage and known market dynamics. It's crafted for AI product leaders, strategists, and enterprise decision-makers navigating the evolving intelligence infrastructure landscape—folks like you, piecing together the next moves.

🔭 i10x Perspective

Ever felt that subtle shift when an industry starts to fracture right before your eyes? This data is the first major tremor signaling a new epoch in the AI wars. The era of a single, dominant "everything model" is likely over, superseded by a war fought on the fronts of distribution, integration, and niche specialization. The competitive landscape is no longer just about who has the smartest model, but who owns the user's workflow. From my vantage, it's invigorating, if a touch unpredictable.

The unresolved tension is whether superior technology can still win against superior distribution. As AI becomes embedded in our operating systems and enterprise software, the models we use may become as much a matter of default settings as of deliberate choice — in short, it will come down to who owns the user's workflow. The next five years will reveal whether the AI market consolidates around a few powerful platform owners or remains a dynamic ecosystem where the best product can still break through—I'm betting on a mix of both, personally.