Samsung Electronics and OpenAI announced on June 21 that ChatGPT Enterprise and Codex are being made available to every Samsung Electronics employee in South Korea and to all employees worldwide inside the company's Device eXperience (DX) division. DX builds Galaxy phones, home appliances, displays, and consumer electronics, and spans tens of thousands of workers across multiple continents. OpenAI described it as one of its largest enterprise rollouts. The use cases named in the announcement cover software development, product engineering, manufacturing operations, marketing, and corporate functions — not just developer tooling.
What makes this a story is the reversal. In March 2023, Samsung banned generative AI tools company-wide after engineers leaked proprietary source code and confidential meeting notes into the public ChatGPT. The 2026 rollout follows a two-month proof-of-concept earlier this year in which 2,500 Samsung employees tested enterprise versions of ChatGPT, Google Gemini, and Anthropic's Claude side by side. OpenAI says Codex weekly active users inside Samsung Korea have grown nearly 800% since February 1, 2026 — meaning the bottoms-up adoption inside the company was already there when the top-down decision landed.
The shape of the deal is becoming the standard enterprise-AI playbook. A large customer runs a multi-vendor pilot, chooses one, and rolls out tens of thousands of seats in a single contract. JPMorgan did it. Now Samsung. Several large banks and pharmas are in the same phase. What that means at the industry level is that the enterprise AI market is consolidating around a small number of frontier vendors and a small number of integration patterns — and that the leakage risk that scared every CISO in 2023 has been priced down to where the productivity case wins.
Takeaway for learners: if you work in a large organization, the year your employer hands you Codex or an equivalent is the year you should treat seriously. The people who learn the productive workflows — when to delegate, when not to, how to review AI-written code, how to phrase a prompt for a 200,000-token codebase — will outperform peers who treat the tool as a curiosity. If you are a teacher, the same arc is coming for school-issued AI: expect district-level pilots in 2026 and 2027, modeled on exactly the enterprise pattern above.