How Anthropic Automates Growth and Bets on Bigger Swings

Apr 5, 2026 · Source ↗

Published 2026-04-05 - Runtime about 113 min - Watch on YouTube

TLDR

Anthropic treats growth as a companywide function: research, inference, compute, Claude Code, and go-to-market drive most of the momentum.
As AI product value compounds, Anthropic shifts growth work toward larger bets, using Claude to automate experimentation through CASH.

Key Takeaways

Amol Avasare got the Anthropic job by cold emailing then-CPO Mike Krieger after using Claude and seeing no formal growth role.
Anthropic says it went from $1 billion to over $19 billion ARR in 14 months, and that figure was already out of date by the interview.
Growth at Anthropic is split into horizontals like growth platform and monetization, plus audience pods for B2B, Claude Code, knowledge workers, and API.
Amol says roughly 70% of his time goes to “success disasters,” where rapid growth creates firefighting across acquisition, activation, and monetization.
Anthropic’s internal growth automation effort is called CASH, short for Claude Accelerate Sustainable Hypergrowth.
Anthropic indexes more heavily toward big swings than small optimizations, roughly flipping the usual growth-team mix toward 70/30 or 50/50.
Activation is described as the highest-leverage AI growth problem, especially cold start and new-user onboarding.
The company’s obsession with exponential model capability makes log-linear charts more useful than linear charts internally.
Anthropic’s culture is described as unusually mission-driven, transparent, and open to public disagreement, including challenge directed at Dario.
Slack notebook channels function like internal Twitter feeds for sharing thoughts, priorities, and debates across research and product.
Amol says Claude Code and the Chrome extension now underpin multiple use cases, and the growth team built the extension because no one else was doing it.
He says a traumatic brain injury forced nine months off work, changed his habits around breaks, meditation, and alcohol or caffeine.

Notes

Amol Avasare previously worked on growth at Mercury and MasterClass, and before that he was a founder and investment banker.
In onboarding at Anthropic, he was in none of the usual hiring buckets: referrals, applications, or sourcing.
He sent Mike Krieger a short cold email saying Anthropic needed a growth team and wanted to chat.
He says he had already honed cold-email copy, subject lines, tone, and follow-up through founder experience.
He avoids crowded outreach channels like LinkedIn and work email when trying to reach someone.
Anthropic’s growth team is around 40 people.
The org includes engineers, designers, PMs, and data, similar to a traditional growth team.
Anthropic has growth pods for B2B growth, Claude Code growth, knowledge worker growth, and API growth.
Amol says the company is effectively living in the future, with product value in two years potentially 100x to 1,000x higher.
He argues that AI companies should favor larger core-product swings because the upside dwarfs traditional small optimizations.
The growth team built a Chrome extension that now supports several Claude and Claude Code use cases.
Anthropic’s culture includes notebook channels, where employees post internal thoughts and can debate leadership openly.
Amol describes the company as mission-driven, with talent density he compares to playing for Real Madrid.