Boris Cherny, the engineer behind Claude Code
Anthropic reached a $2.5 billion run-rate revenue milestone in February for Claude Code, the terminal coding agent created by engineer Boris Cherny.
By Dillip Chowdary • Nov 16, 2026 • Source: cnn.com

At a glance
- Product
- Claude Code
- Creator
- Boris Cherny
- Run-rate revenue
- $2.5 billion in February
- Public release
- May 2025
- Anthropic valuation
- Expected more than $1 trillion IPO
- Survey finding
- 6 in 10 engineers avoid AI tools
In this post
Anthropic has built a foundational enterprise footprint around Claude Code, the terminal-based programming agent developed by Boris Cherny that crossed 2.5 billion dollars in run-rate revenue in February ahead of an expected one trillion dollar initial public offering, as detailed in cnn.com's report. The command-line assistant writes code, operates system shells, and handles local files directly on behalf of workers, providing the developer engine that accelerated corporate artificial intelligence budgets throughout late 2025.
This report examines Cherny's transition from Meta to head of applications at Anthropic, the accidental command-line prototype that became a central revenue driver, the release of Claude Cowork, and how terminal agents are altering traditional engineering workflows. It is intended for software engineers, engineering directors, and product teams tracking autonomous developer tooling and enterprise workplace automation.
01Command-line agent evolution and operating roots
Boris Cherny joined Anthropic in 2024 following an extensive tenure of nearly seven years at Meta, where he worked across Instagram, Facebook Groups, and the internal code quality team. During a 2022 stay in Japan while working remotely for Meta, Cherny focused entirely on writing code away from management meetings and project coordination tasks. After moving to Anthropic, he initially spent weeks prototyping with Claude models before creating a command-line interface tool that granted the model direct operational access to his local operating system.
The turning point occurred when Cherny asked the prototype interface to identify what music track was playing on his machine. Rather than requesting manual intervention or pre-configured scripts, the model autonomously generated and executed code against the system environment to retrieve the answer. That experiment led directly to the creation of Claude Code, which launched to the general public in May 2025 and quickly became Anthropic's primary productivity vehicle across Silicon Valley organizations.
Anthropic documented the rapid financial expansion of the platform inside its S-1 filing, which showed corporate revenue steepening rapidly after Claude Code arrived. Cherny, now Anthropic's head of applications, has maintained a direct feedback loop by answering technical questions and gathering feature requests on X on a weekly or daily schedule. This user-facing testing practice originated during his time at Meta, where he regularly demonstrated unreleased software functions to cafeteria staff to evaluate interface usability.
02Enterprise revenue and developer adoption metrics
The release of Claude Code to the public in May 2025 prompted a distinct acceleration in enterprise artificial intelligence software expenditure. Financial analysis tracked by Ramp senior economist Ara Kharazian noted that the tool served as an early commercially viable autonomous agent offering direct productivity improvements for specific technical roles.
| Metric | Prior baseline | Claude Code deployment |
|---|---|---|
| Annualized run-rate revenue | Sub-billion run-rate | $2.5 billion run-rate in February |
| Enterprise release window | Internal prototype testing | May 2025 public availability |
| S-1 filing growth inflection | Baseline enterprise tiers | Accelerated post-launch revenue curve |
| Developer adoption stance | Experimental tool use | 6 in 10 programmers avoid AI (Stack Overflow) |
Despite widespread enterprise deployment, user sentiment among programmers remains split across the industry. A survey of 30,000 developers by Stack Overflow found that six out of ten programmers actively attempt to avoid using artificial intelligence tools due to burnout, performance pressure, or skepticism over real-world execution claims.
Note: Anthropic does not disclose raw active user totals for Claude Code, but the tool served as the core foundation for Anthropic's broader enterprise automation suite.
03Tool synthesis and Claude Cowork development
Observing engineers and non-technical staff repurposing Claude Code for general administrative tasks prompted Anthropic to construct Claude Cowork. Anthropic engineers built the initial version of Claude Cowork in 10 days using Claude Code itself to manage files and execute workflow automations across varied office contexts. Cherny uses the underlying terminal automation internally for daily team project management tasks and for paying municipal parking citations.
Non-technical workers at non-profit organizations have also deployed the command-line agent to ship long-delayed internal software concepts without dedicated developer staffing. At the Mid-Ohio Food Collective, software engineering director Mark Mollenkopf oversaw deployments where the organization's nutrition director used Claude Code to build a food preference selection application based on concepts originally drafted on whiteboards in 2017.
Advertisement
Tech Pulse Daily
Get tomorrow's pulse first
Join engineers who read Tech Pulse before stand-up. Free, weekday mornings.
In addition to terminal agents, the workflow has expanded to Claude Tag, an integration mechanism that connects indexed workspace data directly to Claude and allows users to invoke the agent within Slack channels. Mollenkopf noted that adopting these agent integrations allows organizations to prototype and release software that previously remained stalled behind primary engineering backlogs.
04Installation procedures and operational model controls
Engineering teams integrate and manage Claude Code locally using standard package managers and terminal execution controls.
npm install -g @anthropic-ai/claude-code
claude updateThe CLI tool allows developers to specify models during active terminal sessions or set explicit runtime parameters when launching commands.
claude --model fableWithin an active command-line session, operators can inspect and adjust model routing directly using interactive slash commands:
/model claude-fable-5-1
/model fableTo standardize model execution across an engineering team, developers persist default configurations directly in their local environment profiles:
export CLAUDE_MODEL="claude-fable-5-1"The command interface allows the agent to inspect project directories, modify code files, and execute compilation commands without exiting the command loop.
05Shifting responsibilities and software engineering roles
The expansion of autonomous coding agents has prompted a structural debate regarding the future scope of software engineering jobs. Cherny stated in interviews that traditional coding tasks will increasingly transition to autonomous agents, shifting the day-to-day responsibilities of technical workers away from writing syntax toward architecture, design coordination, and cross-functional product delivery.
Cherny noted in March that software engineering titles may gradually transition toward designations such as builder as systems handle code construction. While technical jobs continue to exist across corporate organizations through late 2026, the workflow increasingly prioritizes verifying agent-generated outputs, steering multi-step automations, and coordinating across cross-functional teams rather than producing raw source files manually.
Anthropic faces ongoing commercial pressures as it prepares for an initial public offering, navigating competitive pressure from OpenAI, reliance on a concentrated cohort of enterprise accounts, and market expectations surrounding its valuation. Cherny maintains that while agent automation remains disruptive to traditional developer roles, long-term engineering execution will focus on broader system orchestration rather than manual implementation.
Frequently asked questions
What is Claude Code and who created it?
Claude Code is a command-line programming agent created by Boris Cherny, head of applications at Anthropic, that writes code, runs terminal commands, and manages files directly on a user's computer.
How much revenue does Claude Code generate for Anthropic?
Claude Code surpassed 2.5 billion dollars in run-rate revenue in February, serving as a primary driver of enterprise revenue ahead of Anthropic's expected one trillion dollar initial public offering.
What other tools did Anthropic build using Claude Code?
Anthropic used Claude Code to build Claude Cowork in 10 days, creating an agent designed for office tasks, file management, and workflow automation.
How are software engineering responsibilities changing because of Claude Code?
Boris Cherny projects that as agents automate routine code writing, engineers will spend more time collaborating with designers and planning architecture, potentially shifting job titles toward roles like builder.
Sources
Author
Dillip Chowdary
Writes Tech Bytes coverage of AI, engineering, and the tools that actually ship. Editor of Tech Pulse Daily.
Related on Tech Bytes
Advertisement