OpenAI merges ChatGPT and Codex into single AI agent

What happened

OpenAI is merging ChatGPT and Codex—its code-generation tool—into a single agent (self-directed AI) that can handle everything from answering questions to booking travel and creating custom apps. The company has reorganized product leadership under Greg Brockman and Thibault Sottiaux to execute this shift, and has discontinued side projects including its Sora video product and halted expensive data center plans in the U.K. and Norway.

Why it matters

OpenAI faces a competitive challenge in enterprise adoption—Ramp's May 2026 AI Index showed Anthropic at 34.4% of U.S. business adoption versus OpenAI's 32.3%, the first time Anthropic had crossed ahead, largely because of Claude Code's autonomous capabilities. By positioning ChatGPT as more than a conversational chatbot, OpenAI hopes to persuade both enterprise customers and IPO investors that it is a broader platform company. The company is eyeing a trillion-dollar IPO as part of a wave in which SpaceX, OpenAI, and Anthropic are targeting combined valuations that add up to more than $3.6 trillion（約580兆円）.

What to watch

OpenAI's earlier agentic product, Operator, largely failed to take off because the underlying models were not ready to handle ambiguous instructions and operate without mistakes in constrained environments. Sottiaux argues the technology has advanced enough to handle a much larger set of tools and operate with broader permissions, but whether enterprise customers will choose an all-in-one platform over vertical solutions tailored to specific industries remains an open question.