The race for AI coding dominance has intensified rapidly, with enterprises now treating programming agents as one of the few AI categories worth significant spending — making the competitive gap between OpenAI and Anthropic a matter of real commercial consequence.
According to the report, Anthropic‘s coding agent Claude Code accounts for nearly a fifth of the company’s business — more than $2.5 billion in annualized revenue, a figure the firm cited in February. OpenAI‘s competing product, Codex, was bringing in just over $1 billion in annualized revenue by the end of January, according to a person with direct knowledge of the matter.
That gap sits at the center of a broader story. To report it, the source spoke with more than 30 people, including current OpenAI leaders and employees who participated with the company’s approval, alongside others who spoke anonymously to discuss the inner workings of private companies.
How Codex Started — and Stalled
Codex has a longer history than the current competition suggests. Back in 2021, the tool — an offshoot of OpenAI‘s GPT-3 model trained on billions of lines of open source code from GitHub — was demonstrated to journalists with the promise that it could take English commands and produce code in return. Greg Brockman, OpenAI‘s president and cofounder, described it at the time as a system that “can carry out commands” and “act in the computer world on your behalf.”
That ambition had early limits. One former OpenAI employee told the source that Codex “couldn’t do much more than autocomplete” at that stage. The technology was licensed to Microsoft, which used it to build GitHub Copilot — a code completion tool that launched publicly in June 2022 and attracted hundreds of thousands of users within months. After that, OpenAI‘s original Codex team moved on to other projects.
The departure of key Anthropic founders from OpenAI — and the subsequent rise of Claude Code — put the firm in unfamiliar territory. The accounts gathered paint, in the source’s words, “a picture of OpenAI in a position it has rarely ever been in: racing to catch up.”
Altman’s Case for Catching Up
Sam Altman does not dispute the current standings. “First to market is worth a lot,” he told the source. “We had that with ChatGPT.” He frames the moment, however, as an opportunity rather than a setback, arguing that OpenAI‘s models are now capable enough to power competitive coding agents.
His ambitions go further than market share. Altman called coding “one of these rare multitrillion-dollar markets” — a phrase he said he does not use lightly — pointing to both the direct economic value and the broader work that programming capabilities can unlock.
He also placed Codex at the center of the company’s longest-range goal. The product is, he said, “probably the most likely path” to building artificial general intelligence, which OpenAI defines as an AI system that can outperform humans at most economically valuable work.
The first OpenAI Codex team has already moved on to other projects within the company.
Photo by Mikhail Nilov on Pexels
This article is a curated summary based on third-party sources. Source: Read the original article
