OpenAI is expanding cooperation with Samsung Electronics on next-generation chips, semiconductor supply and enterprise artificial-intelligence services. Harrison Kim, general manager of OpenAI Korea, said the companies had made progress in joint chip production and research but did not provide design, schedule or manufacturing details.
Samsung said it could not confirm information related to customers. That response leaves the precise commercial arrangement undisclosed. Reuters reported the statement alongside OpenAI’s description of cooperation, so the established fact is that OpenAI publicly characterized the relationship as deepening while Samsung did not validate specific customer information.
OpenAI unveiled its first custom inference chip in June. Its engineers designed the processor, called Jalapeno, with Broadcom, and OpenAI said Taiwan’s TSMC would manufacture it. Inference chips perform the calculations required to answer user requests after a model has been trained, making their cost and energy use central to running large services.
Samsung Electronics and SK Hynix signed letters of intent last year to supply memory chips for OpenAI’s Stargate data-center project. OpenAI’s Korea office expects demand for advanced memory to grow as models require faster and more complex computing. Samsung has also expanded internal access to ChatGPT, Gemini and Claude in its device division.
OpenAI said ChatGPT Enterprise users at South Korean organizations increased roughly 28-fold over the year ending in August, but it declined to publish absolute counts. South Korea’s paid-subscriber ranking and enterprise-growth figures are company data. Contract values, chip specifications, production volumes and delivery dates remain undisclosed.
The checked record for OpenAI and Samsung Deepen Work on Custom AI Chips establishes the following dated facts: OpenAI says it is deepening chip work with Samsung Electronics. Samsung declined to confirm customer-specific information. OpenAI designed its Jalapeno inference chip with Broadcom. OpenAI previously said TSMC would manufacture that chip. Samsung and SK Hynix signed memory-supply letters of intent for Stargate. The relevant background is also specific: Inference hardware handles model queries after training. Advanced AI systems require large amounts of high-bandwidth memory. Company adoption figures were reported without absolute user totals. No chip specifications, contract values, production volumes or delivery dates were disclosed. The cited reporting does not resolve questions beyond those stated limits.