Samsung ChatGPT Enterprise: Codex Reaches Non-Developers in OpenAI's Biggest Korea Rollout

Samsung Electronics deployed ChatGPT Enterprise and Codex to its entire South Korean workforce and all employees in its Device eXperience division worldwide on June 21, 2026. OpenAI described the action as one of its largest enterprise rollouts. The Device eXperience division, which produces Galaxy smartphones, home appliances, and consumer electronics, covers tens of thousands of employees. The move reverses a March 2023 company-wide ban on generative AI tools that followed an incident in which engineers uploaded proprietary source code and confidential records to a public version of ChatGPT. The June 21 agreement covers ChatGPT Enterprise security features, including zero model training on customer data, identity-based access controls, and data-loss prevention. It is distinct from Samsung's June 9 multi-vendor AI announcement by specifying the formal terms for ChatGPT Enterprise and Codex. Samsung's Device Solutions semiconductor division operates under separate limits and is not included in the global Codex deployment. Codex, introduced as a cloud-based coding agent in 2025, is now extended to marketing, manufacturing, product design, and corporate staff without programming backgrounds. OpenAI stated that more than 5 million people use Codex weekly, with non-technical users representing roughly 20 percent of the base and growing at three times the rate of developers. Weekly active users in South Korea increased nearly 800 percent between February 1, 2026, and the announcement. The fastest-growing tasks for these users involve data analysis, research, and report creation. Roh Tae-moon, President and Head of Samsung's DX Division, said the deployment is the starting point for fundamentally transforming how employees work and execute tasks. Harrison Kim, General Manager of OpenAI Korea, said Samsung is adopting AI as a core platform to enhance work and innovation globally. Codex operates through an agent loop that builds prompts from environment context, project files, and user requests, then executes tool calls in isolated cloud sandboxes. Prompt caching via prefix reuse and compaction for long contexts supports scale. The system runs on GPT-5.5. The deployment marks the first large-scale test of natural-language agentic tools at industrial scope, following a comparable but smaller rollout by NVIDIA to more than 10,000 employees earlier in 2026.