In a historic about-face, Samsung Electronics has officially signed on for one of OpenAI’s largest enterprise deployments ever, with a Samsung ChatGPT Enterprise rollout that also brings Codex to its global workforce.

The mega-deal marks a complete structural reversal for the South Korean tech giant, which made global headlines three years ago by slapping a strict, company-wide ban on generative AI tools after engineers accidentally leaked sensitive proprietary source code onto ChatGPT’s public servers.

1. Scope of the Deployment

The rollout is not a limited pilot; it is being deployed as a core operational platform across major arms of the conglomerate:

  • The Domestic Footprint: Accessible to all Samsung Electronics employees based in South Korea, spanning both technical and corporate roles.
  • The Global Footprint: Extended to all employees worldwide within the Device eXperience (DX) division—the massive corporate umbrella that builds Samsung’s smartphones, tablets, displays, and home appliances.
  • The Semiconductor Gate: Tighter boundaries remain intact elsewhere. The Device Solutions (DS) division, which handles Samsung’s highly sensitive, cutting-edge semiconductor and memory-chip manufacturing, is excluded from this specific blanket rollout to prevent any exposure of foundational chip designs.

2. Breaking the Guardrail Dilemma: How the Reversal Happened

Samsung didn’t just drop the ban blindly. The path to this enterprise agreement was cleared by establishing rigid security structures and extensive testing:

  • The Multi-Vendor Bake-off: Between April and May, Samsung’s DX division conducted a two-month proof-of-concept (PoC) using 2,500 employees. The team tested ChatGPT alongside Google’s Gemini 3 and Anthropic’s Claude to analyze security leaks and operational efficiency before ultimately awarding OpenAI the primary contract.
  • Enterprise-Grade Air Gaps: The contract utilizes OpenAI’s specialized enterprise architecture. Among the core ChatGPT Enterprise features Samsung relied on, none of the data, prompts, or code entered by Samsung employees can be used to train OpenAI’s models, and all data remains sandboxed within a strictly managed corporate cloud environment.
  • The Mandatory Training Gate: Access to the tools is restricted. Employees are only granted active login credentials after completing a specialized internal corporate AI security compliance training course.

3. The Codex Coding Boom

A major component of the deal focuses on Codex, OpenAI’s dedicated coding and automation engine.

Even before the formal corporate-wide deployment was announced, internal telemetry revealed an aggressive appetite for automation within the company. According to OpenAI, weekly active users of Codex within South Korea exploded by nearly 800% between February 1 and the June launch, as teams used the engine to write, review, and debug software code. Non-technical corporate teams are also using the engine to build localized internal workflow dashboards and data visualization tools, complementing other productivity features like ChatGPT’s new Scheduled tasks.

4. Expanding the Enterprise Supply Chain

The agreement alters the business dynamic between the two giants. Historically, the relationship was anchored around hardware infrastructure, with OpenAI relying on Samsung to supply the advanced high-bandwidth memory (HBM) semiconductors required to run its frontier data centers. The new contract expands that relationship directly into the enterprise application layer, arriving as OpenAI navigates intense cost pressure detailed in OpenAI’s Q1 2026 financials.

Furthermore, the deal establishes Samsung SDS (the group’s enterprise IT services arm) as an authorized, official reseller of ChatGPT Enterprise and ChatGPT Edu across the Asian market. This gives OpenAI a highly secure, localized distribution channel to sell enterprise AI subscriptions to other major South Korean corporations, universities, and public sector institutions.

“This is not simply about introducing AI as a workplace tool. It marks the starting point for fundamentally transforming the way we work and execute.”

— Roh Tae-moon, President and Head of Samsung Electronics’ DX Division

Frequently Asked Questions

How big is the Samsung–OpenAI deal?

It is described as one of OpenAI’s largest enterprise deployments ever. ChatGPT Enterprise and Codex are being rolled out to all Samsung Electronics employees in South Korea and to the entire global Device eXperience (DX) division, while the sensitive Device Solutions (DS) chip division is excluded. Samsung SDS also becomes an authorized reseller of ChatGPT Enterprise and ChatGPT Edu across Asia.

Is ChatGPT Enterprise data used for training?

No. Under OpenAI’s enterprise architecture used in this deal, none of the data, prompts, or code entered by Samsung employees can be used to train OpenAI’s models, and all data remains sandboxed within a strictly managed corporate cloud environment.

What is ChatGPT Enterprise?

ChatGPT Enterprise is OpenAI’s business tier of ChatGPT, built around a specialized enterprise architecture that keeps company data sandboxed and excludes it from model training. In the Samsung deal it is deployed as a core operational platform, paired with the Codex coding engine, and resold across Asia by Samsung SDS alongside the education-focused ChatGPT Edu.

What can you do with ChatGPT Enterprise at Samsung?

Samsung employees use it to write, review, and debug software through Codex—whose weekly active users in South Korea jumped nearly 800% before launch—and non-technical teams build internal workflow dashboards and data-visualization tools. Access is gated behind a mandatory AI security compliance training course before login credentials are issued.