A self-taught programmer without a computer science degree is rewriting the competitive landscape of the AI programming sector.
Boris Cherny, head of Anthropic's AI programming tool Claude Code, has transformed this product from a &#34;lackluster launch&#34; into Anthropic's core growth engine over the past year, helping Anthropic match or even surpass OpenAI in revenue.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/7ddc1dce-89ca-482f-bb8a-e03f26e74eed.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="Boris Cherny video screenshot" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/7ddc1dce-89ca-482f-bb8a-e03f26e74eed.png"/>
Boris Cherny video screenshot
<h2>From 1 Billion to 30 Billion: A Steep Growth Curve</h2>
Numbers speak for themselves.
According to The Information, Claude Code's annualized revenue was $1 billion last December, rose to $2.5 billion in February this year, and has accelerated again recently. Cherny stated in an interview that this growth has driven Anthropic's overall annualized revenue to $30 billion.
This growth curve was not steep from the start. Claude Code was officially released in February 2025 and initially received a lukewarm response. Cherny admitted: &#34;It didn't skyrocket immediately; instead, it stagnated for a while.&#34; The product truly exploded after a more powerful underlying model was launched.
Last May, Anthropic held its first developer conference. Cherny sat backstage, sandwiched between stagehands and lighting technicians, testing Claude Code with the new models Opus 4 and Sonnet 4 released that day.
&#34;I just remember the feeling of: 'My God, this thing, it actually works,'&#34; he said. Since then, the proportion of his personal workload generated by Claude's code jumped from about 10% to nearly 30%.
Last November, with the release of Claude Opus 4.5, the product saw another leap. Venture capitalist Tian Yu (Micron Ventures) stated that it was from that point on that AI reached human-level proficiency in software engineering—new models allowed Claude Code to run longer, reason more effectively, and produce fewer code errors.
<h2>The Grassroots Path of a Self-Taught Programmer</h2>
Cherny's personal experience is a narrative full of twists and turns in itself.
Cherny is 34 years old. He got his start in programming during middle school by writing web code for eBay listings and later wrote programs on his calculator for entertainment.
He attended the University of California, San Diego, majoring in economics, but ultimately did not obtain a degree—he dropped out to start a startup.
In September 2024, he joined Anthropic after working at Meta for nearly seven years. During his time at Meta, he led efforts to improve code quality. &#34;I was attracted by Anthropic's mission to develop advanced AI for the benefit of humanity,&#34; he said.
In September 2024, he joined Anthropic as the first engineer for Claude Code.
<h2>Leaving for Cursor, Then Returning</h2>
Before Claude Code truly exploded, Cherny experienced an interlude.
Last July, Cherny and Claude Code product lead Catherine Wu both resigned and jumped to the AI programming tool startup Cursor—he served as Chief Architect and Head of Engineering, and Wu served as Head of Product. Cursor is both a customer and a competitor of Anthropic, and the industry generally viewed this poaching as a major victory for Cursor.
But the ending was unexpected. Cherny said that after only a few weeks, he realized it was a mistake and proactively contacted Anthropic to propose a return. The two immediately returned to their original positions.
&#34; The reason for coming back was 100% because I missed that mission. Other than doing direct alignment research, I really can't think of anything more urgent or important than what I'm doing now,&#34; Cherny said.
So-called &#34;alignment research&#34; refers to the research direction of matching AI model objectives with developer intent.
<h2>Reshaping the Market Landscape, Impacting Traditional Software Stocks</h2>
Claude Code's commercial value has been validated by leading companies. According to Anthropic, Netflix, Accenture, and Uber have all become customers.
Cherny estimates that the software engineering market targeted by Claude Code exceeds $1 trillion in size.
The rise of this product has also cast a shadow over the capital markets. Share prices of traditional software companies such as Salesforce, ServiceNow, and Workday are under pressure, and the market has begun to reassess their long-term business models. Meanwhile, concerns about software engineers being &#34;replaced&#34; are spreading throughout the industry.
In this regard, Cherny believes that programming remains &#34;a skill with great practical value,&#34; but he also predicts: &#34;In a year, we will see a world where anyone can do software engineering. Just as programming has been solved for me, it will be solved for everyone.&#34;
At the practical application level, Claude Code's capabilities have also impressed engineers. Rob Sun, a founding engineer at robotics company Dexterity, once used Claude Code to complete a computer vision project in just a few days.
&#34;These were a series of problems I thought I would never be able to solve independently,&#34; he said. &#34;If it were just me, it might have taken three years of non-stop work.&#34;
<h2>Next Step: From Writing Code to &#34;Doing Everything&#34;</h2>
Cherny and his team are extending Claude Code's automation capabilities to areas beyond programming, including legal and financial tasks, under the product name Claude Cowork.
The team itself is a heavy user of Cowork. Cherny has Cowork monitor the weekly progress forms filled out by employees; if someone misses a form, Cowork automatically sends a reminder on Slack. &#34;I just went to get a cup of coffee,&#34; he said.
He is also using Claude to book flights and handle parking tickets. Last month in Seattle, he had Cowork autonomously navigate the local Department of Fish and Wildlife website, pay $7, and successfully apply for a foraging permit.
The entire Claude Code team remains lean. Cherny estimates that this team's revenue per person is the highest in the world, though he declined to disclose specific figures. The final version of Claude Cowork was developed by the team using Claude Code in just 10 days.
<h2>Competition and Risks: The Moat is Not Invincible</h2>
Claude Code's lead is facing pressure.
OpenAI has shifted its focus to its programming tool Codex, and some developers say Codex is more efficient than Claude Code in certain scenarios, with some even using the two together.
Anthropic also experienced an accidental leak last month—the underlying software code for Claude Code was inadvertently made public, giving competitors a rare glimpse into its operational mechanisms and unplanned feature roadmap, including the ability for Claude Code to take more proactive actions.
Furthermore, although Anthropic's latest model, Mythos, has achieved outstanding results in programming tests, it is currently not being released to the public because it has been identified as having potential for cyber-attacks.

SOXL

SOXX

Boris Cherny, 34, a college dropout and self-taught programmer, moved from Meta to Anthropic and became the "father of Claude Code." Under his leadership, the product's annualized revenue surged from $1 billion last December to $2.5 billion this February, driving Anthropic's overall annualized revenue to $30 billion, surpassing OpenAI

- Self-taught programmer Boris Cherny has transformed the competitive landscape of AI programming tools with Anthropic's Claude Code, driving a revenue surge.  
- Claude Code's annual revenue skyrocketed from $1 billion in December to $25 billion in February, eventually helping Anthropic reach a total annual revenue of $30 billion.  
- As Claude Code evolves, it faces competitive pressure from OpenAI's Codex and recent security vulnerabilities, but its impact on the software engineering market remains significant.

Wallstreetcn

The Creator of Claude Code: How a Self-Taught Programmer is Reshaping the AI Coding Space