OpenAI's Codex Hits 1M Enterprise Users
OpenAI's AI coding assistant, Codex, has surpassed one million users, signaling significant enterprise adoption. CEO Sam Altman has also clarified that the company will not dictate how clients like the Pentagon use its technology, positioning OpenAI as a neutral platform provider.
Originally a specialized version of GPT-3, Codex was first introduced in 2021 and became the engine for the initial version of GitHub Copilot. It was trained on 159 gigabytes of Python code from 54 million public GitHub repositories to translate natural language comments into functional code. The current iteration of Codex, launched in May 2025, has evolved from a code-completion tool into an autonomous software agent. It now operates in a secure cloud environment where it can independently write features, fix bugs, run tests, and propose pull requests for review, fundamentally changing the workflow from assisting a developer to delegating entire tasks. Recent growth has been rapid, with weekly active users tripling since the start of 2026, driven by the release of new models like GPT-5.3-Codex and a desktop app with over a million downloads. Companies like Cisco, Temporal, and Superhuman are using Codex for tasks ranging from accelerating feature development to refactoring large codebases. The AI coding assistant market is projected to grow from around $7.4 billion in 2025 to over $30 billion by 2032. Competitors include Google's Jules, Anthropic's Claude Code, and Microsoft's GitHub Copilot, which is now powered by GPT-4 and used by 90% of Fortune 100 companies. Altman's comments on platform neutrality come amid a wider industry debate over military contracts. Rival AI company Anthropic recently refused a deal with the Pentagon over ethical concerns, leading the U.S. Defense Secretary to label them a "supply-chain risk" before the department announced a deal with OpenAI. OpenAI has since clarified its agreement with the Pentagon includes "red lines" preventing the use of its tech for mass domestic surveillance or directing autonomous weapons. Altman has stated that while OpenAI controls the safety features, it does not make operational military decisions, a stance that has created internal and public backlash. While adoption soars, the impact on productivity is still debated. Some studies show developers coding up to 55% faster, while others find that gains on boilerplate tasks are offset by delays in review cycles. A recent Anthropic study even suggested that reliance on AI assistance could reduce a developer's mastery of new skills by 17%.
