How Claude Code Is Transforming Software Development and Anthropic’s Future

Silicon Valley engineers have been buzzing about Anthropic’s AI coding tool, Claude Code, for quite a while. Recently, the excitement has intensified significantly.

This week, I spoke with Boris Cherny, head of Claude Code, to gain insight into how the company is navigating this pivotal moment.

“We created the simplest possible tool,” Cherny noted. “The most surprising revelation came three months ago when I learned that half of our sales team at Anthropic uses Claude Code on a weekly basis.”

AI-driven coding has rapidly advanced. Between 2021 and 2024, many tools served primarily as autocomplete options, offering developers a few lines of code while they worked. By early 2025, startups like Cursor and Windsurf began introducing early “agentic” coding tools, allowing developers to describe features in simple terms and let an AI take over.

Claude Code was launched around that same period. Cherny admits that initial iterations of Claude Code faced challenges, often making mistakes or getting ensnared in costly loops. However, he claims Anthropic designed Claude Code with future AI capabilities in mind rather than its initial state.

That decision proved to be insightful. Many developers believe AI coding tools have reached a turning point recently, particularly with the rollout of Anthropic’s latest AI model, Claude Opus 4.5.

Kian Katanforoosh, an adjunct AI lecturer at Stanford and CEO of Workera, mentions that his company transitioned to Claude Code after evaluating several AI coding solutions. He asserts that Claude Code outperformed tools from Cursor and Windsurf for his senior engineers.

“The only model where I’ve observed a significant leap in coding performance recently has been Claude Opus 4.5,” Katanforoosh claims. “It doesn’t just feel like it’s coding like a human; it feels like it has discovered a better method.”

Last year marked a significant upsurge in the AI coding agent industry. In November, Anthropic revealed that Claude Code achieved $1 billion in annual recurring revenue just months after its launch.

By the close of 2025, Claude Code’s ARR increased by at least another $100 million, as per an insider familiar with the company’s finances. At that point, the product represented around 12 percent of Anthropic’s overall ARR, estimated at $9 billion. Although smaller than Anthropic’s corporate business, which provides AI systems to large organizations, coding is one of the company’s fastest-expanding sectors.

Anthropic has also informed investors of its goal to achieve positive cash flow by 2028, with Claude Code playing a pivotal role in its revenue growth. The company chose not to comment on its financial situation.

While Anthropic appears to hold a strong position in AI coding, the excitement surrounding Claude Opus 4.5 seems to be benefiting various companies. Cursor, which allows users to code using models from Anthropic and other AI labs, announced its coding tool also reached $1 billion in ARR last November. In December, the company reported substantial month-over-month revenue growth, according to a source close to the company. OpenAI, Google, and xAI are likewise striving to capture a larger portion of the AI coding market, creating their own agentic solutions powered by proprietary AI models.

Currently, Anthropic is leveraging the momentum from Claude Code to develop agents for non-coding sectors. Earlier this month, the company introduced Cowork, an AI agent designed to manage files on a user’s computer and interact with software—without the need for any programming terminal involvement.

This interview has been edited and condensed for clarity.

WIRED: What accounts for the surge of interest in Claude Code now?