Samsung Electronics is giving ChatGPT Enterprise and Codex to all of its employees in Korea and to Device eXperience employees worldwide, turning one of the most visible early corporate cautionary tales about generative AI into one of OpenAI’s largest enterprise deployments.
OpenAI announced the rollout on June 21, describing it as one of its largest enterprise launches to date. The agreement gives Samsung workers access to ChatGPT Enterprise and Codex across technical and non-technical functions, including research and development, manufacturing, marketing, product development, corporate work, software development, and internal workflow automation.
The important shift is not simply that Samsung employees can use ChatGPT. It is that a company known for tightly managing sensitive product, chip, and manufacturing information is moving from broad caution around external generative AI tools toward a governed deployment with enterprise access controls. That is becoming the pattern for large companies that spent 2023 and 2024 blocking or limiting public AI tools, then spent 2025 and 2026 looking for ways to bring the same capabilities inside a controlled security and compliance model.
What Samsung Employees Are Getting
The deployment covers ChatGPT Enterprise, OpenAI’s business version of ChatGPT, and Codex, its coding and work-automation system. OpenAI says ChatGPT Enterprise includes data protection, user and access management, and security controls designed to let organizations use the product within internal governance policies.
For Samsung, the immediate use cases are broad. Employees can use ChatGPT for information search and analysis, document drafting, ideation, and interpreting data. Codex is positioned for writing, reviewing, and debugging code, but OpenAI is also pushing it beyond engineering teams. In Samsung’s case, the company expects employees to use Codex to turn ideas into internal tools, websites, automated workflows, and other software-like outputs.
That detail matters because the enterprise AI race is moving away from a narrow “coding assistant” category. Tools that began in software development are being repackaged as general workplace systems that can assemble small apps, connect documents, automate repetitive tasks, and help non-engineers build practical internal software. OpenAI also reported that more than 5 million people now use Codex each week, with weekly active Codex users in Korea up nearly 800% since February 1, 2026.
Why This Is a Reversal for Samsung
Samsung became an early example of corporate AI risk when employees were reported in 2023 to have pasted sensitive information into ChatGPT. The episode helped push many large companies toward bans, strict approval processes, or internal-only AI tools. That history is what makes the new rollout notable: Samsung is not treating external generative AI as a permanently unacceptable risk, but as a tool category that can be reintroduced when controls, contracts, and internal policies are mature enough.
The move also lines up with a broader generative AI opening inside South Korean conglomerates. CIO reported last week that Samsung Electronics’ DX division was preparing to introduce external generative AI services including ChatGPT, Gemini, and Claude for employees, after relying heavily on in-house models. OpenAI’s announcement now confirms a major part of that transition on the ChatGPT and Codex side.
This is not a small pilot framed around a few innovation teams. The Korean employee base and global DX division give the rollout direct exposure to Samsung’s consumer device, software, marketing, and product operations. The DX division spans products such as Galaxy devices, TVs, appliances, and connected-device experiences, which makes the deployment relevant to both internal productivity and the way Samsung designs AI-enabled consumer products.
The OpenAI-Samsung Relationship Is Getting Wider
The employee rollout also expands a relationship that already touches AI infrastructure. OpenAI notes that Samsung Electronics has been working with it to supply advanced memory semiconductors needed for next-generation AI systems. That gives the partnership two layers: Samsung as a supplier to the AI buildout, and Samsung as a large internal customer for AI workplace tools.
That dual role is increasingly common among the largest technology companies. Semiconductor makers, cloud providers, device companies, and software vendors are not just selling into the AI market; they are reorganizing their own workforces around the same systems. For Samsung, stronger internal AI adoption could influence product development cycles, software tooling, manufacturing operations, support workflows, and the company’s own Galaxy AI strategy.
For OpenAI, Samsung is a prestige enterprise customer in a market where business adoption matters as much as consumer usage. ChatGPT’s public popularity made OpenAI a household name, but enterprise deployments are where the company has to prove that its products can handle security controls, access management, employee training, compliance concerns, and measurable productivity gains at global-company scale.
What Enterprises Should Watch Next
The Samsung rollout is a useful signal for other companies still deciding whether to ban, tolerate, or formally adopt AI tools. The emerging model is not unrestricted employee use of public chatbots. It is contracted enterprise access, identity controls, data-handling rules, approved workflows, training, monitoring, and clear boundaries around which information can move through AI systems.
The harder questions come after launch. Samsung will need to show whether employees actually use the tools in daily work, whether Codex becomes useful outside software teams, how the company measures productivity gains, and how it handles sensitive chip, product, and customer information inside AI-assisted workflows. Those are the same questions facing any large organization moving from AI experiments to company-wide deployment.
The headline is that Samsung is embracing ChatGPT and Codex. The more durable story is that enterprise AI is becoming normal enough for cautious companies to stop treating it as a shadow-IT problem and start managing it like core workplace infrastructure.
