The article argues that Anthropic should be evaluated as a cloud company, not only as an AI lab. It compares Anthropic with AWS, Azure, and GCP using a table of Q2 2026 figures. Anthropic has the smallest revenue: $12b on a $46b annual run rate, versus $42b/$169b for AWS, $39b/$157b for Azure, and $25b/$99b for GCP. Its year-over-year growth is 1,374%, compared with 82% for GCP, 37% for AWS, and 32% for Azure; the article puts the same point as 15x growth last year. Anthropic‘s operating margin is undisclosed, while AWS, Azure, and GCP show 39%, 40%, and 35%. Anthropic is about 30% the size of AWS or Azure and nearly half the size of GCP.

At current growth rates, Anthropic is closer to GCP‘s pace, roughly doubling next year; on that path it would pass Azure in about three years and AWS in about three and a half. Only GCP, growing slightly faster than Anthropic‘s current pace, would stay ahead. Anthropic also has the most concentrated revenue base of the group: two customers accounted for nearly 25% of revenue in 2025.

The piece draws a parallel with the wedges that established the other clouds: Amazon had ecommerce, Google had search and ads, and Microsoft had the operating system. Anthropic‘s wedge is AI models and the leading coding harness, Claude Code. The article reports that 87% of businesses on Ramp that buy a coding agent buy Claude Code, up 56 points in a year. Anthropic is also exercising pricing power: on Artificial Analysis’ Intelligence Index, Claude Opus 5.5 costs $5.98 per task, versus $3.26 for OpenAI’s GPT-6 Astra and $1.99 for Google’s Gemini 4 Argon. It has the narrowest product offering, though the article credits it with a record-breaking start across AI models, artifacts, and design.

Capacity is another cloud-like signal. Anthropic has made the largest data center commitments of the four, at $518b, nearly 8x its $67b Q3 run rate. AWS, Microsoft, and Google have committed about 1x to 2x theirs. The businesses are converging in other ways too: each is building custom chips, scaling data centers with debt, developing and/or reselling models, and managing infrastructure to support AI.

The valuation case follows directly. Using Wall Street estimates, a standalone GCP would be worth roughly $1t to $2.3t. At $1.5t to $2t, Anthropic would be valued equally to GCP on two-thirds the revenue and far faster growth over the past year. The closing question—if it rains like a cloud and computes like a cloud—frames the takeaway: evaluate Anthropic the way investors evaluate clouds, not just model makers.