OpenAI deepens Samsung ties with next-generation AI chip collaboration

SEOUL: OpenAI is getting deeper into the chip game, and Samsung Electronics is becoming a key part of that silicon strategy.

The ChatGPT maker is working with Samsung Electronics on the research and production of next-generation artificial intelligence chips, expanding a relationship that already covers memory supplies for OpenAI’s Stargate data-centre project and the use of ChatGPT across Samsung’s workforce.

Harrison Kim, general manager of OpenAI Korea, disclosed the latest collaboration at a press conference in Seoul on 9 September. He said joint research and production of next-generation chips was one of the areas where the two companies had made the most progress, although he did not provide details on the chips or Samsung’s precise role.

The development comes as OpenAI looks to build more of its own AI hardware at a time when the industry’s appetite for computing power and advanced memory is soaring. As AI models become more capable, the chips needed to train and run them are becoming an increasingly important part of the competitive race.

OpenAI has been steadily moving towards custom silicon as it seeks greater control over the infrastructure supporting its AI models.

In June, the company unveiled Jalapeno, its first custom AI chip, developed with semiconductor company Broadcom. OpenAI’s engineers designed the processor for AI inference, the stage where an AI system processes data to generate responses for services such as ChatGPT.

OpenAI said at the time that Taiwan Semiconductor Manufacturing Company, or TSMC, would manufacture the chips.

The Samsung collaboration could give OpenAI a foothold in another part of the semiconductor supply chain. However, the company has not disclosed whether Samsung will manufacture the new chips, supply components or memory, or participate in another part of their development.

The move nevertheless underlines how OpenAI’s hardware ambitions are broadening beyond designing chips. Securing access to the right manufacturing capacity and memory is becoming just as important as building the processors themselves.

The latest chip work builds on an existing relationship between the two companies.

Last year, Samsung Electronics and SK Hynix signed letters of intent to supply memory chips for OpenAI’s Stargate data-centre project. The initiative is part of OpenAI’s broader effort to build large-scale computing infrastructure to support its growing AI operations.

Kim said demand for advanced memory would continue to rise as AI systems require faster and more complex computing. He added that cooperation with South Korean semiconductor companies was therefore a priority for OpenAI’s local operation.

For Samsung, the relationship offers a route into the infrastructure underpinning one of the world’s most closely watched AI businesses.

It also comes at a time when memory has become a crucial bottleneck for AI infrastructure, with high-performance systems requiring large quantities of increasingly advanced memory alongside powerful processors.

The relationship is not limited to chips.

Samsung Electronics is also among the largest-scale corporate deployments of ChatGPT globally, according to Kim. Employees in South Korea and other markets are using the AI assistant across areas including research and development, marketing and sales.

Samsung has been widening its use of generative AI internally. In June, the company said employees in its Device eXperience division, which covers businesses including smartphones, televisions and home appliances, would be permitted to use ChatGPT, Google Gemini and Anthropic’s Claude.

The shift highlights how generative AI is moving beyond individual experiments and into broader corporate workflows.

OpenAI said the number of ChatGPT Enterprise users at South Korean companies and institutions had increased about 28-fold by the end of August compared with a year earlier. Businesses are increasingly applying AI across legal, sales, finance, marketing and operations, rather than limiting it to isolated productivity tasks.

OpenAI is also betting heavily on South Korea as an important market for its products and infrastructure ambitions.

The company said South Korea had the largest number of paying ChatGPT subscribers outside the US as of late 2025, although it did not disclose the absolute figure.

OpenAI expects adoption to gather further pace following the 4 September launch of GPT-6 Astra, its latest model. The company said the model delivers improvements in areas including computer use, professional work, science, coding and cybersecurity.

That growing demand gives OpenAI another reason to deepen relationships with South Korea’s technology industry, particularly its semiconductor giants.

OpenAI’s ties with Samsung now stretch from the chips powering AI systems to the employees using AI tools at work. As the AI race increasingly becomes a battle over both models and the infrastructure needed to run them, that combination could prove more valuable than either side of the partnership alone.