"AI Programming" Milestone: Claude Code "Rectifies" the Global Software Industry

date
13:57 24/01/2026
avatar
GMT Eight
Anthropic's AI programming tool Claude Code is leading the global software development industry into a key turning point.
Anthropic's AI programming tool Claude Code is leading the global software development industry into a key turning point. With its latest "agentic" capabilities, the product not only completely changes the way code is written, but also achieves explosive growth in commercialization. On January 22, according to the technology media WIRED, Boris Cherny, head of Anthropic Claude Code, revealed in an interview that as the underlying model capabilities rise, software development is transitioning from human-written code to being fully delegated to AI. This transformation is reshaping the work mode of engineers and the productivity logic of the entire industry. Reports indicate that Claude Code's Annual Recurring Revenue (ARR) has further increased to at least $1 billion by the end of 2025, after the product surpassed the $1 billion ARR milestone in November. Currently, Claude Code contributes 12% of Anthropic's total ARR of around $9 billion, becoming one of the fastest-growing business segments for the company. This breakthrough is mainly attributed to the launch of Anthropic's latest AI model, Claude Opus 4.5, which is considered by many developers as a "step-function" improvement in programming abilities. Different from earlier tools that could only provide code completion, Claude Code based on the new model can understand natural language instructions and autonomously complete subsequent development, prompting startups like Cursor and Windsurf, as well as tech giants like OpenAI, Google, and xAI to accelerate their efforts and compete for a share in this emerging market. With the maturity of AI delegation capabilities, Anthropic is attempting to replicate this "restructuring" of the software industry into broader fields. The company recently launched a new product called Cowork, aiming to extend AI delegation functions from programming terminals to general office scenarios such as file management and software interaction, hoping to spark a similar productivity revolution in non-programming fields. From auxiliary tools to core productivity The evolution of AI programming tools has exceeded expectations. According to Wired, from 2021 to 2024, most tools could only provide auto-completion suggestions and a few lines of code when developers input. By early 2025, startups like Cursor and Windsurf began rolling out early "agent-style" programming products. Boris Cherny, head of Claude Code, said in an interview that Anthropic bet on the future development direction of AI capabilities when designing the product, rather than its current level. Cherny said: "We built the simplest thing, and the craziest thing is that three months ago, we found out that half of Anthropic's sales team was using Claude Code every week." This strategy has proven to be forward-thinking. Cherny revealed that his personal use of Claude Code for code writing has increased from 5% initially to 30% after the releases of Opus 4 and Sonnet 4 in May last year, and after Opus 4.5 was launched, all of his code is written by Claude Code. Reports indicate that Workera chose Claude Code after testing multiple AI programming tools. Katanforoosh stated that Claude Code performed better than competing products like Cursor and Windsurf for the company's senior engineers. Agent mode changing the way of work The core advantage of Claude Code lies in its agent-style work mode. Cherny explained that unlike a year ago when users mainly used chat, Claude Code and Cowork have truly achieved delegation, allowing them to use tools, read system files, and interact with Slack and Google Sheets. "This is a golden age for those of us with short attention spans," Cherny said. The most efficient users of Claude Code would launch multiple tasks simultaneously, have Claude execute them individually, and then check their progress sequentially. Cherny himself would run five to ten agents simultaneously on his terminal, mobile, and web platforms. This work mode is becoming popular within Anthropic. During internal reviews before product releases, CEO Dario Amodei once asked if employees should be mandated to use the product. In reality, almost 100% of Anthropic's technical staff frequently use Claude Code, with 95% of the team's code written by Claude Code. Enterprise clients of Anthropic exhibit a similar usage pattern. Cherny stated that enterprise clients have similar security requirements and product interaction methods as Anthropic itself, enabling the company to optimize the product through internal usage. Expanding into non-programming fields Building on the success of Claude Code, Anthropic launched Cowork this month, an AI agent product aimed at non-programmers, which can manage files on users' computers and interact with software without having to touch the coding terminal. Cherny described Cowork as "Claude Code for non-programmers". He has started using Cowork for project management, such as having it check if engineers have filled out work progress sheets and reminding those who haven't through Slack. "AI agents will be able to handle all the tedious things in life, which has been achieved in the engineering field this year, and I think it will happen in all other fields. Agents will be able to fill out forms, move data to different places, send emails, and other tasks." Cherny acknowledges that this shift will bring disruptive impacts and require industries to collectively respond. However, he believes that it will make work more enjoyable and liberate people from tedious tasks. Intensifying competition landscape The success of Claude Code is attracting more competitors to enter the AI programming market. Cursor announced in November last year that its Annual Recurring Revenue reached $1 billion, allowing users to program using models from Anthropic and other AI labs. According to sources, Cursor achieved particularly strong month-over-month revenue growth in December. OpenAI, Google, and xAI are also accelerating the development of agent-style programming products based on their proprietary AI models, attempting to capture a larger share in this market. The core of this competition lies in the capabilities of underlying AI models, and the launch of Claude Opus 4.5 seems to temporarily give Anthropic an advantage. Cherny offers a historical perspective for engineers undergoing this transformation. His grandfather used punch cards for programming in the Soviet Union, and programming has since evolved through machine code, C language, Java, Python, and other languages. "As an industry, we have always been undergoing transformation, and this is a continuous abstraction. I think delegation is just one point on this continuum." Cherny anticipates that as the learning curve becomes smoother, effectively using tools like Claude Code and Cowork will become easier and easier. Anthropic plans to achieve positive cash flow by 2028, with Claude Code playing a key role in reaching this goal as one of the fastest-growing business segments. This article is reproduced from Wall Street News, GMTEight edited by: Li Junling.