OpenAI takes another step forward in its self-developed AI chip! Its collaboration with Samsung expands, potentially extending from HBM to wafer foundry services.

OpenAI takes another step forward in its self-developed AI chip! Its collaboration with Samsung expands, potentially extending from HBM to wafer foundry services.

OpenAI is accelerating its self-developed chip strategy, and its collaboration with Samsung Electronics is extending into a wider range of semiconductor fields.

According to Reuters, Harrison Kim, General Manager of OpenAI Korea, stated at a press conference in Seoul on September 9 that OpenAI and Samsung Electronics have made "the most significant progress" in the joint production and research of next-generation chips. This is OpenAI's most explicit public statement to date regarding their chip cooperation.

Harrison Kim also emphasized that as demand for more advanced chips continues to rise, the demand for memory chips will continue to grow, and cooperation with South Korean semiconductor suppliers has become a priority for OpenAI. Analysts believe that this statement may indicate that Samsung's role in the OpenAI chip supply chain will expand from memory supply to wafer foundry and advanced packaging businesses.

Chip Collaboration Frameworks: From Stargate to Jalapeño

The foundation for their collaboration was laid as early as last year. According to MS TODAY, Samsung Electronics and OpenAI signed a letter of intent in October 2025, with Samsung joining OpenAI's Stargate project as a strategic memory partner, responsible for supplying high-performance, low-power memory. At that time, Samsung estimated that OpenAI's memory needs could eventually reach as high as 900,000 DRAM wafers per month.

This collaboration has already been implemented in OpenAI's first self-developed chip, Jalapeño. According to reports from MS TODAY and Aju News, Jalapeño, jointly developed by OpenAI and Broadcom, is positioned for AI inference acceleration. With the assistance of AI models, it was completed in approximately nine months, manufactured by TSMC, and reportedly supplied with sixth-generation HBM4 high-bandwidth memory. OpenAI officially released the chip in June of this year.

It's worth noting that Jalapeño is just the starting point of OpenAI's self-developed chip roadmap. OpenAI disclosed in August that its second-generation chip is in intensive development, and its third-generation chip is already in the planning stages.

Boundaries of Cooperation: The Imagination Beyond Memory

As the collaboration deepens, discussions about whether Samsung can secure further OpenAI chip manufacturing orders continue to intensify. MS TODAY points out that this partnership is expected to open up new opportunities for Samsung beyond AI memory, encompassing wafer foundry and advanced packaging businesses.

However, the report also cautioned that it is too early to interpret the aforementioned collaboration as Samsung having secured a foundry order for OpenAI's self-developed chips—key details such as the manufacturer, process node, and production capacity have not yet been disclosed. Currently, TSMC remains the designated manufacturer for Jalapeño.

Samsung's foundry business has also attracted the attention of other AI giants. According to The Information and Yonhap News Agency, Anthropic is in contact with Samsung Foundry for custom chips needed for its Claude model, exploring its 2-nanometer process and advanced packaging capabilities. However, the discussions are still in the early stages and have not yet entered the detailed design, testing, or mass production phases.

South Korean Market Strategy: Accelerated Enterprise AI Penetration

In addition to chip collaborations, OpenAI's expansion in the South Korean enterprise market is also accelerating. Harrison Kim revealed at the press conference that as of the end of August this year, the number of ChatGPT enterprise users in South Korean enterprises and institutions has surged by about 28 times compared to a year ago. The application of AI in enterprises is rapidly penetrating from individual tasks to company-wide operations, covering multiple functional lines such as legal, sales, finance, marketing and operations.

According to OpenAI, South Korea became the country with the most paid ChatGPT subscribers outside the United States at the end of last year. OpenAI stated that with the launch of the latest model, GPT-6 Astra, on September 4th, user adoption is expected to accelerate further. The new model offers performance improvements in areas such as computer use, professional work, scientific research, programming, and cybersecurity.

Samsung is also expanding the use of OpenAI products internally. Harrison Kim revealed that Samsung has "one of the world's largest enterprise deployments of ChatGPT," with its employees using ChatGPT in their daily work across multiple areas, including R&D, marketing, and sales. Samsung, however, stated that it could not confirm information related to its customers.

Risk Warning and DisclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.