Anthropic’s annualized revenue has reached $30 billion, up from $9 billion at the end of 2025 and $14 billion just two months prior. The growth trajectory, detailed in a feature by The Atlantic, makes Anthropic “possibly the fastest-growing business in the history of capitalism,” outpacing Zoom’s pandemic surge, Google’s early-2000s expansion, and even Standard Oil during the Gilded Age.
Claude Code as Revenue Engine
The acceleration traces to a single product line: Claude Code. Anthropic’s autonomous coding agents, released in an updated form last November, crossed what The Atlantic described as “some invisible threshold between interesting gadget and life-changing technology.” Teams of AI agents can now take over a developer’s machine and complete programming tasks in minutes or hours that previously required days or weeks of human work.
The financial impact is concrete. Anthropic disclosed in its Series G announcement that Claude Code’s standalone run-rate revenue exceeded $2.5 billion and had more than doubled since early 2026. More than 1,000 enterprise customers now spend over $1 million per year with the company, according to TechCrunch.
Demand Outrunning Compute
The growth has created infrastructure bottlenecks. Anthropic has been forced to limit customer access to its coding tools during peak hours because compute capacity cannot keep pace with demand, according to The Atlantic. OpenAI has scrapped its video-generation app to free up computing power for similar reasons.
Goldman Sachs researchers who interviewed 40 software companies in mid-April found that many were “overrunning their initial budgets” for AI tools “by orders of magnitude,” with some companies spending as much as 10 percent of their total engineering labor costs on AI, The Atlantic reported. Gabriela Borges, a software analyst at Goldman Sachs, told the publication that “the speed at which we’re seeing companies adapting these tools is actually quite surprising.”
Productivity Claims Getting Independent Validation
The spending surge coincides with new research validating agent productivity. The think tank Model Evaluation & Threat Research, which ran a widely cited 2025 experiment showing developers completed tasks 20 percent slower with AI, recently re-ran the study using current tools. This time, developers completed tasks almost 20 percent faster with AI, according to The Atlantic. Some power users had become so dependent on AI tools they refused to participate in the control group.
By one estimate cited in the article, the percentage of American businesses with a paid AI subscription has risen from about a quarter at the beginning of 2025 to over half today.
The Valuation Race
The revenue milestone feeds directly into Anthropic’s reported fundraising plans. TechCrunch reported that Anthropic could close a new $40 to $50 billion round at a valuation exceeding $900 billion within two weeks, which would surpass OpenAI’s $852 billion as the most valuable AI startup. OpenAI’s own annualized revenue grew nearly 20 percent from December to February but sits at roughly $25 billion, well behind Anthropic’s pace.
The broader cloud market is riding the same current. Google, Microsoft, and Amazon reported cloud revenue growth of 48 percent, 39 percent, and 24 percent respectively compared with the year prior, largely driven by AI workloads.
Compute Scarcity as the New Constraint
Six months ago, the dominant narrative was that data center investments had gotten ahead of demand. That calculus has inverted. Semiconductor demand is so high that even Nvidia’s fourth-best AI chip, released in 2022, costs more today than it did three years ago. Ethan Mollick, co-director of the Generative AI Lab at the University of Pennsylvania, told The Atlantic: “For years now, we’ve been in an era of chatbots that mostly just say things. Now we’ve officially crossed into the era of agents that can actually do things.”
The constraint on Anthropic’s growth is no longer demand or product-market fit. It is physical infrastructure: chips, power, and data center capacity. For agent platform operators, the takeaway is that willingness-to-pay for autonomous coding tools has outstripped every supply forecast, and that gap is widening.