Codex's Success Killed OpenAI's Hardware Dreams and Multi-App Strategy

OpenAI is folding its ChatGPT app, Codex coding platform, and Atlas web browser into a single desktop application, the company confirmed Thursday, marking its most significant product consolidation since the ChatGPT launch and a clear signal of where Sam Altman wants to place his bet.

The Wall Street Journal first reported the plan, and OpenAI confirmed it to Reuters without providing a timeline or target date for the unified product.

The move traces a straight line back to one number: the adoption rate of Codex, OpenAI's standalone coding assistant that launched in February. According to people familiar with the matter, Codex exceeded internal expectations in a way that Sora, the video generator, or OpenAI's hardware partnership with Jony Ive's io Devices did not. It was the proof of concept for what focus looks like.

Fidji Simo, OpenAI's chief of applications, spelled out the logic in a memo to staff cited by the Journal. "We realised we were spreading our efforts across too many apps and stacks, and that we need to simplify our efforts," she wrote. "That fragmentation has been slowing us down and making it harder to hit the quality bar we want."

Simo, the former Instacart CEO who joined OpenAI in May to run its applications business, has been steadily pushing the company toward the kind of product discipline she learned scaling a logistics company under intense unit economics pressure. The superapp plan is the fullest expression of that instinct yet.

Greg Brockman, OpenAI's president, will temporarily oversee the product overhaul and associated organizational changes while Simo leads the commercial push. "When new bets start to work, like we're seeing now with Codex, it's very important to double down on them and avoid distractions," Simo wrote in a post on X Thursday, quoting the Journal's reporting by Berber Jin.

The catalyst for this moment was not internal inspiration alone. Anthropic's rise in the enterprise market has been the quiet alarm inside OpenAI for months. Following Claude Code's surge in popularity among developers and the company's broader traction with the Claude for Work suite, OpenAI's leadership held an all-hands meeting earlier this month where Simo called the competitive situation a "code red," according to the Journal. She told employees that Anthropic's success should function as a wake-up call.

The superapp plan is also shadowed by IPO preparations. OpenAI has signaled it could go public as soon as this year, and a cleaner product lineup — one flagship desktop application instead of a constellation of separate tools — is easier to pitch to institutional investors than a sprawling product portfolio with uneven adoption.

Atlas, OpenAI's AI-powered browser, is the most notable casualty of the consolidation logic. It launched in October with significant fanfare as part of OpenAI's broader consumer expansion. Less than five months later, it is being merged rather than built out as a standalone product. The mobile version of ChatGPT is not affected, according to the Journal.

The framing of "superapp" will invite comparisons to WeChat, the Chinese platform that combined messaging, payments, commerce, and mini-programs under one roof. Whether OpenAI's version will attempt anything similar in scope — or stop at bundling productivity tools for knowledge workers — is not yet clear. An OpenAI spokesperson declined to comment beyond the confirmed details.

What is clear is that OpenAI is entering a different phase. The era of announcing new products at a clip — Sora, the hardware device, the Atlas browser, eCommerce capabilities for ChatGPT — is giving way to an era of integration and focus. The question is whether the company that bet on everything can execute on a single thing well enough to satisfy the investors, enterprise customers, and employees who are watching closely.