Claude Code launched in May 2025. By February 2026, it had crossed $2.5 billion in annualized run-rate revenue and had more than doubled its weekly active user base since January 1.

That is not the trajectory of a product finding its market. That is the trajectory of a product that has already found it and is scaling into it at speed.

Anthropic disclosed those figures in the context of its February 2026 Series G fundraise, which drew investment from GIC, Coatue, Founders Fund, Sequoia Capital, and NVIDIA. The Claude Code numbers were part of the story of why Anthropic warranted the valuation investors paid.

What Claude Code Is

Claude Code is Anthropic’s agentic coding assistant, which differs from earlier AI coding tools in a significant way. Earlier tools like GitHub Copilot operated primarily as autocomplete: they suggested code as you typed, offered completions for functions, and could generate boilerplate on request. The developer remained the driver.

Claude Code is closer to an agent. It can take a task description, reason about the codebase, write code, run tests, identify failures, and iterate. The developer’s role shifts from writing code to reviewing and directing it.

This architectural shift explains the adoption pattern. Tools that assist at the margins produce incremental improvements. Tools that restructure the workflow produce step-change differences in output, which justifies much more intensive use and much higher spend per seat.

The Uber case study makes this concrete. Uber gave Claude Code access to its 5,000-person engineering team in December 2025. By April 2026, 84% of engineers were using it at least monthly, 70% of committed code was AI-generated, and the company had burned through its entire annual AI budget in four months. The CTO was not unhappy about the productivity. He was surprised by the budget math.

That story has repeated, in different forms, at enterprises across multiple industries. The spending is not speculative. It is being driven by measurable output changes that organisations find worth paying for.

The Numbers Behind the Growth

$2.5 billion in annualized run-rate revenue means Claude Code was generating approximately $208 million in monthly revenue by early 2026. More than doubling weekly active users in the space of roughly five weeks implies either a significant new enterprise contract, a marketing or product initiative that drove rapid activation, or both.

For context, GitHub Copilot, which had the distribution advantage of being embedded across the world’s largest code hosting platform and the marketing weight of Microsoft behind it, took considerably longer to reach comparable revenue scale. Claude Code did it as a standalone product from a company that does not own a code hosting platform.

The gap between those two trajectories is partly explained by the product architecture difference described above, and partly by the timing. Cursor, Claude Code, and the broader category of genuinely agentic coding tools arrived after enterprises had already absorbed earlier, simpler AI tools and were ready for something more capable.

What the Trajectory Means

A product at $2.5 billion in annualized revenue that is doubling user growth in a single month is not near the top of its growth curve. It is still in acceleration.

The natural ceiling for AI coding tools is not obvious yet. Enterprise software development is a very large category of spending. If AI coding tools eventually become the primary driver of code produced across the software industry, and the evidence from early adopters suggests they are heading in that direction, the revenue potential for the leading tools is enormous relative to current valuations.

For businesses, the question is not whether to adopt AI coding tools but how to do it in a way that captures genuine productivity gain rather than generating high spend without corresponding output improvement. The Uber story is instructive: they got real productivity gains, but the budget model was wrong because it was built on assumptions from an earlier, lower-intensity adoption phase.

Building the right adoption framework, the right cost governance, and the right measurement of what you actually get from AI coding investment is the work that separates organisations that capture value from AI from those that generate impressive spend numbers without commensurate business results.

---

Enterprise DNA’s Omni Advisory helps business leaders build AI deployment frameworks that deliver real operational results, not just high AI spend. Talk to Sam about what that looks like for your engineering organisation.