Invest Like the Best Notify me• May 13, 2026• 1:22:03Interview

How Claude went from $9 billion to $30 billion in one year | CFO explains

From Invest Like the Best

Krishna Rao(CFO, Anthropic, guest)

Get the full transcript next time Invest Like the Best releases an episode

Summary, key quotes, top claims, and the searchable transcript - emailed automatically. No card needed.

Executive Summary

Continue your research

Keep pulling the thread on Krishna Rao.

Compute as a Strategic Differentiator Exponential Growth and Capital Intensity

12 quotes

Concerns Raised

Managing the 'cone of uncertainty' in forecasting compute needs; buying too much risks insolvency, while buying too little means falling off the frontier.
The extreme capital intensity required to compete in the AI arms race.
The long lead times for securing large-scale compute capacity, which requires planning 12-24 months in advance.

Opportunities Identified

Capturing the vast enterprise market by delivering 'frontier intelligence' that unlocks new use cases and productivity gains.
Leveraging a multi-platform compute strategy to be the most efficient user of capital among frontier labs.
Accelerating drug discovery and development in healthcare and biotech, which is seen as a major long-term positive outcome of advanced AI.
Expanding the Total Addressable Market (TAM) with each successive, more capable model generation.

Key Themes

Research Findings12

Anthropic utilizes three different chip platforms for its compute needs: Amazon's Tranium, Google's TPUs, and NVIDIA's GPUs.

Anthropic has invested for multiple years to develop an orchestration layer that allows it to use different chip platforms fungibly for model development, internal use, and serving customers.

To achieve customization and flexibility, Anthropic builds its own compilers and develops its software stack from the chip level up.

Anthropic believes the returns to frontier intelligence are extremely high, especially in the enterprise market.

Anthropic's run rate revenue grew from approximately $9 billion at the start of the year to over $30 billion by the end of the first quarter.

Over 90% of the code written at Anthropic is generated by its own AI model, Claude Code.

Anthropic does not see a slowdown in the effectiveness of scaling laws for model improvement.

Anthropic has partnered with SpaceX to use its Colossus facility in Memphis for compute capacity.

Anthropic signed a deal with Google and Broadcom for 5 gigawatts of TPU compute capacity starting in 2027.

Anthropic signed a deal with Amazon for up to 5 gigawatts of Tranium compute capacity.

Anthropic sells to nine of the Fortune 10 companies.

Anthropic has raised $75 billion since Krishna Rao joined the company approximately two years ago.

Topics

AI Infrastructure Compute Strategy Cloud Partnerships Anthropic Krishna Rao Frontier Models Large Language Models (LLMs)AI Economics Capital Intensity Enterprise AI Financial Planning & Analysis (FP&A)AI in Finance Dogfooding NVIDIA GPUs Google TPUs AWS Tranium Scaling Laws Net Dollar Retention

Processed May 13, 2026Daily intelligence brief → yt-dlp + mlx-whisper + Gemini

How Claude went from $9 billion to $30 billion in one year | CFO explains

Continue your research

Concerns Raised

Opportunities Identified

Key Themes

Compute as a Strategic Differentiator

Exponential Growth and Capital Intensity

Internal AI Adoption as a Flywheel

The Economics of Frontier Intelligence

Deep, Multi-faceted Cloud Partnerships

Research Findings12

Topics