Samsung Handed Every Employee ChatGPT — and the Real Opening Is the Gap It Leaves

On June 21, Samsung rolled out ChatGPT Enterprise and Codex to its entire Korean workforce and its global DX division — one of OpenAI’s largest enterprise deployments ever. A company that banned generative AI three years ago after a source-code leak just flipped. When giants standardize on AI internally, the opening for founders isn’t the headline — it’s the adoption gap.

What Happened

On June 21, Samsung Electronics said it would supply ChatGPT Enterprise and the automation tool Codex to all of its employees in Korea, plus everyone in the Device eXperience (DX) division worldwide — the unit that builds Galaxy phones and home appliances. OpenAI called it “one of its largest enterprise deployments to date.” The DX division alone spans tens of thousands of workers across multiple continents. Employees use ChatGPT to search and analyze information, draft documents, and interpret data; Codex handles software development and internal workflow automation. Codex weekly active users in Korea have grown nearly 800% since February 1. What makes the move heavy is its history. In March 2023, Samsung banned generative AI company-wide after engineers uploaded proprietary source code and meeting notes to public ChatGPT. This is a full reversal — but not a reckless one. Across April and May, 2,500 DX employees ran a proof-of-concept testing ChatGPT, Gemini, and Claude side by side, and access is gated behind mandatory internal AI security training. Back in December, Samsung SDS became OpenAI’s first authorized reseller in Korea. Training for the full global workforce is slated to wrap by year-end.

What This Means for Founders

In one line: the default behavior of enterprise employees just changed. Generative AI adoption in many large organizations now reads high on paper, yet real, organization-wide internalization lags far behind — and that gap is the market. The true significance of Samsung “handing every employee ChatGPT” isn’t OpenAI’s revenue; it’s that tens of thousands of people using AI daily has become the baseline. The pattern is familiar to anyone watching FAANG mandate internal copilots, or YC batches fill with companies selling on top of those rollouts. But ChatGPT Enterprise and Codex are general-purpose tools. They don’t know Samsung’s chip process, its appliance support queues, or its compliance rules. That’s exactly why Samsung spent two months and 2,500 people on a PoC and gated access behind training. Closing the distance between a general model and real work — connecting internal data, managing permissions, producing audit logs, building domain-specific workflows — is not something OpenAI fills for you. While the platform giants push their own suites, the seats next to ERP systems, internal groupware, and factory-floor tooling stay empty. The more enterprises install AI, the bigger the demand for narrow products that own one task end to end on top of it.

What You Can Do Now

First, shift your target from “companies that don’t use AI” to “companies that already installed it and aren’t seeing results.” The distance between high adoption rates and low real internalization is your customer list. Second, pick the one thing general-purpose ChatGPT can’t do — internal system integration, a specific role’s repetitive work, regulatory evidence — and go narrow and deep. Third, make security and auditability a sales point from day one. Just as Samsung opened access only to trained employees, the first question an enterprise buyer asks is “what does it learn, and how far do its permissions reach?” Fourth, don’t dismiss the reseller and SI channel — the Samsung SDS path. Breaking into large enterprises through a trusted channel often beats direct sales. Don’t envy the announcement. Fill the gap the announcement created.