OpenAI deepens Samsung alliance on next-generation AI chips

Summary

• Next-generation chip development: OpenAI says it has made significant progress with Samsung Electronics on joint research and production involving next-generation AI chips, although the companies have not disclosed technical specifications or production volumes.
• Large enterprise deployment: Samsung is one of OpenAI’s largest enterprise deployments globally, with ChatGPT Enterprise and Codex available to Samsung employees in Korea and across its Device eXperience (DX) division worldwide.
• Stargate memory supply: Samsung and SK hynix previously signed letters of intent with OpenAI to support the supply of advanced memory for the Stargate AI infrastructure programme, strengthening South Korea’s role in OpenAI’s hardware ecosystem.
• Rapid Korean adoption: OpenAI says South Korea has its largest number of paying ChatGPT subscribers outside the United States, while ChatGPT Enterprise users in the country increased 28-fold over the past year.

SEOUL, September 9, 2026 — OpenAI is expanding its relationship with Samsung Electronics into next-generation semiconductor development while continuing a large-scale enterprise deployment of ChatGPT and Codex across the South Korean technology group.

Speaking at a press briefing in Seoul, OpenAI Korea General Manager Harrison Kim said the companies had made significant progress in joint research and production involving future AI chips. He did not disclose the architecture, manufacturing process, production volumes or commercial timetable. Samsung declined to comment on confidential customer engagements.

The development adds another layer to OpenAI’s strategy of building more of the hardware stack behind its AI systems. In June, OpenAI and Broadcom unveiled Jalapeño, OpenAI’s first custom intelligence processor, designed specifically for large-language-model inference. OpenAI said the chip was developed from design to production in nine months and is intended for deployment at gigawatt scale across multiple generations.

Samsung expands its role in OpenAI’s hardware strategy

Samsung’s involvement goes beyond the newly disclosed chip-development work. In October 2025, Samsung Electronics and SK hynix announced strategic partnerships with OpenAI as part of the Stargate initiative, including plans to expand production of advanced memory chips needed for next-generation AI systems.

SK hynix separately signed a letter of intent to supply high-bandwidth memory for Stargate and said it would establish production capacity capable of responding to OpenAI’s potential demand of up to 900,000 DRAM wafer starts per month.

The arrangements underline the growing importance of memory alongside computing accelerators as AI workloads become more demanding. However, the latest announcement does not establish that Samsung has guaranteed OpenAI a particular volume of HBM4 or that the partnership is specifically intended to replace TSMC as OpenAI’s fabrication partner.

OpenAI’s existing custom-chip programme with Broadcom remains an important part of its hardware roadmap. Jalapeño is designed around the memory, networking and compute requirements of large language models, with the companies describing it as the first generation of a broader multi-generation platform.

Samsung becomes a major enterprise AI customer

The semiconductor relationship is developing alongside a significant software deployment.

OpenAI announced in June that Samsung Electronics would make ChatGPT Enterprise and Codex available to all Samsung Electronics employees in Korea and to all DX employees worldwide. OpenAI described the deployment as one of its largest enterprise launches to date.

The tools are being used across a wide range of functions, including software development, product development, manufacturing, marketing and other corporate activities. Samsung has also authorised employees in its DX division to use other AI services, including Google’s Gemini and Anthropic’s Claude, highlighting the broader competitive AI environment inside the company.

OpenAI said on Wednesday that ChatGPT Enterprise users in South Korea had increased 28-fold year-on-year. It also said South Korea has the highest number of paying ChatGPT subscribers outside the United States.

The growth comes as OpenAI expands the capabilities of its latest model, GPT-6 Astra, which was introduced on September 3. OpenAI says Astra improves computer use, coding, research and complex multi-step work and is being rolled out progressively across its products and API.

Why South Korea matters to OpenAI

The deeper Samsung relationship reinforces South Korea’s importance to OpenAI on both the semiconductor and enterprise sides.

Samsung and SK hynix are among the world’s major memory-chip manufacturers, while South Korea has a large concentration of semiconductor manufacturing expertise and AI-related infrastructure. OpenAI’s Stargate partnerships therefore give the company access to an important part of the global semiconductor supply chain while also expanding its presence among Korean enterprises.

For Samsung, the relationship provides another channel through which its semiconductor, enterprise software and AI infrastructure capabilities can participate in the rapidly expanding market for AI computing.

Why this matters

• Broader custom-silicon strategy: OpenAI is moving beyond reliance on off-the-shelf accelerators by developing purpose-built processors and working with multiple semiconductor partners.
• Memory becomes increasingly strategic: The Stargate agreements with Samsung and SK hynix demonstrate the importance of high-bandwidth memory to scaling AI infrastructure.
• Samsung is a major enterprise AI deployment: The company’s rollout of ChatGPT Enterprise and Codex represents one of OpenAI’s largest enterprise deployments and extends AI adoption across technical and corporate functions.
• South Korea gains strategic importance: OpenAI’s semiconductor partnerships and rapidly growing enterprise user base make South Korea an increasingly important part of its global AI ecosystem.

FAQs

Q1: What AI chips is OpenAI developing with Samsung?

OpenAI has confirmed that it is working with Samsung on joint research and production involving next-generation chips. The companies have not disclosed the chip architecture, manufacturing process, production volume or launch schedule.

Q2: Is Samsung replacing TSMC as OpenAI’s chip manufacturer?

There is no confirmation of that. OpenAI continues to work with Broadcom on its custom Jalapeño inference processor, while Samsung’s newly disclosed role involves next-generation chip research and production. OpenAI has not announced Samsung as a replacement for TSMC.

Q3: What is Samsung’s role in Stargate?

Samsung and SK hynix joined OpenAI’s Stargate initiative in 2025 through partnerships focused on advanced memory supply and AI infrastructure. Samsung Electronics is also involved in broader cooperation covering semiconductors and data-centre infrastructure.

Q4: How widely is Samsung using OpenAI products?

OpenAI says ChatGPT Enterprise and Codex are available to all Samsung Electronics employees in Korea and all DX employees worldwide. The deployment covers areas including R&D, manufacturing, software development, product development and marketing.

Q5: How fast is enterprise AI adoption growing in South Korea?

OpenAI reported that ChatGPT Enterprise users in South Korea increased 28-fold over the past year. It also said South Korea has its largest paying ChatGPT subscriber base outside the United States.