What will Claude Code's annualized run-rate revenue (ARR) be at the end of May 2027, in USD billions?

rationale:

My median forecast anticipates Claude Code annualized run-rate revenue reaching approximately $18.5 billion by the end of May 2027. Since Anthropic's October 2026 S-1 filing provides no product-level revenue, user, or seat figures sec.gov, this estimate relies on projecting Anthropic's total company run-rate revenue and multiplying it by an assumed Claude Code share. A plausible trajectory puts total company run-rate near $105–$110 billion in May 2027, with Claude Code maintaining a share around 17%, slightly below its historical level because of competitive compression and deflationary API pricing.

The strongest historical anchor for this product share is Anthropic's disclosure on February 12, 2026, when Claude Code's run-rate passed $2.5 billion 2 sources. At that time, total company run-rate was roughly $14 billion, implying an 18% share for Claude Code. We should largely discount unconfirmed third-party tracking estimates from mid-2026 and rely instead on extrapolating from this verified baseline, while allowing for modest share compression from competition and lower API pricing alongside recent qualitative adoption signals.

Anthropic's total revenue growth is clearly decelerating, which heavily influences the baseline projection. Total run-rate grew from roughly $65 billion at the end of July to an estimated $74 billion in mid-October euronews.com, averaging roughly 5% monthly growth. Furthermore, Q3 2026 revenue of $16.8 billion represented a 45% increase over Q2, a sharp step down from the 2.4x sequential growth seen between Q1 and Q2. Projecting another seven months at a cooling 3–6% monthly rate yields a total May 2027 run-rate of $105–$110 billion. This deceleration aligns with management's explicit warnings that revenue growth may decline significantly sec.gov and the deflationary drag of recent 25–45% per-task price cuts for Opus 5.5 and Fable 5.1 sec.gov.

Claude Code's share of that total could fluctuate, but competitive compression and deflationary API pricing shift the center of gravity modestly below its historical 18%, to around 17%. On the upside, the S-1 highlights exceptional monetization potential: the top 5% of Claude Enterprise users utilizing Claude Code spend about 20 times more than non-users sec.gov. It is also becoming the "default choice inside many enterprise engineering organizations." However, Claude Code's share could be constrained if Anthropic attributes a significant portion of coding spend to the API (which powers third-party AI-native tools) rather than direct subscriptions, or if internal usage definitions remain narrow following the merger of Claude Cowork into Claude.ai.

The 80% confidence interval ($9.8B to $33.0B), with an interquartile range of $13.8B to $24.5B, reflects substantial uncertainty over future product definitions, market structure, competitive compression, pricing deflation, and the sustainability of agentic coding hypergrowth. The downside tail accounts for potential market saturation, the 5–7% revenue headwind from policy-violating customers being removed, and the risk that Anthropic uses a highly restrictive definition of "Claude Code" revenue in future disclosures. The upside tail captures scenarios where enterprise adoption scales dramatically or Anthropic reports a broader "agentic" revenue category that sweeps in adjacent products and massive API coding workloads.