Lenny's Podcast: Head of Growth at Anthropic — how Claude grew from $1B to $19B ARR in 14 months

Amol Avasare joined Anthropic as Head of Growth after cold-emailing CPO Mike Krieger. In this April 2026 episode of Lenny’s Podcast, he describes what it looks like to work on growth at one of the fastest-scaling companies in the history of software — from $1 billion to over $19 billion in annual recurring revenue in 14 months. The conversation covers how Anthropic thinks about activation, experimentation, team structure, and the role of AI in running growth functions themselves.

Who it is for

This episode is most useful for product managers and growth practitioners at AI-first companies, or at companies where an AI product is the central revenue driver. It is also relevant for PMs at any company trying to think clearly about how to apply AI to their own growth function. Some experience with growth frameworks helps, though the interview is accessible to a broader audience.

Key takeaways

1. Activation is the most critical problem in AI product growth. Avasare explains that for Claude, the biggest growth constraint is not discovery — it is getting users to experience the product’s core value before they stop. This pattern applies broadly to AI products that require some learning from the user before the value becomes apparent, and it reframes where to focus engineering and product effort.

2. Anthropic built CASH, an internal AI system for automating growth experiments. Rather than having people design and analyze experiments manually, CASH takes over parts of that process autonomously. The existence of this tool reflects a view that operating growth at scale requires AI infrastructure as much as human judgment, and that building internal tools is a first-class growth activity.

3. Deliberate onboarding friction can improve long-term retention. Counterintuitively, Anthropic made some parts of the onboarding process harder rather than easier, with the goal of filtering for users who are genuinely engaged. Avasare is specific about where this works and where it does not, which makes the advice actionable rather than generic.

4. A 70/30 ratio toward big bets has outperformed incremental growth work. Allocating the majority of resources to ambitious initiatives rather than marginal improvements accelerated Anthropic’s trajectory. Avasare frames this as a deliberate structural choice, not a retrospective story about luck, and walks through how the team operationalizes the split.

5. Cowork, an internal AI tool, identifies team misalignment by analyzing Slack patterns. This is one of the more concrete examples in the episode of Anthropic applying its own AI capabilities to internal operations. The implication is that teams building AI products should be running AI on their own workflows first.

Worth watching if

You are responsible for growth at an AI product and want to see how a practitioner at the fastest-growing AI company currently operating thinks about activation, experimentation speed, and whether you should be building AI tools to run your growth function rather than relying entirely on traditional methods.