Friday, September 11, 2026
AI 인프라 · 뉴스 & 분석
반도체·하드웨어리포트
반도체·하드웨어 · 리포트

삼성전자가 차세대 맞춤형 AI 프로세서 개발을 위해 오픈에이아이와 협력 중이며, 이는 기존 파운드리 협력을 넘어선 확장이다.

이러한 심화된 공동 설계 파트너십은 OpenAI가 맞춤형 실리콘 성능과 공급망 다각화를 추진하고 있음을 시사하며, 차세대 AI ASIC 생산에서 TSMC의 독점적 지위에 도전할 가능성이 있다.
업계 전문지Slicast · September 10, 2026 · 글로벌 · 출처: Data Center Dynamics
중요도 85

Samsung is reportedly working with OpenAI on the next generation of the ChatGPT-maker’s custom chips. Harrison Kim, general manager of OpenAI Korea, told reporters that the two companies are expanding their partnership across semiconductors and enterprise AI deployment. “One of the areas where we have made the most progress and gained the most recognition with Samsung Electronics is our joint production and research on the next-generation chips we are developing,” Kim said, in comments first reported by Reuters.

OpenAI revealed details of its first foray into semiconductors, named Jalapeño, last month. Developed in partnership with Broadcom, it is rated at 700W and will be deployed in racks featuring 128 chips. TSMC manufactures the chips in its fabs in Taiwan.

Quite what form OpenAI’s extended collaboration with Samsung has taken remains to be seen. DCD has contacted both firms for comment.

Last year, Samsung and another Korean semiconductor firm, SK Hynix, signed a non-binding letter of intent to supply memory chips to OpenAI for use in its data centers. Speaking at a press conference, Kim said Samsung is one of the biggest corporate users of ChatGPT globally. South Korea is the second-largest market in the world when it comes to paid users of the chatbot, behind the US.

The news comes a day after OpenAI revealed it was taking space at two data centers in South Korea operated by Firmus. But the firm’s plan to build its own Stargate data center in Korea is apparently facing delays as it struggles to procure a suitable site.

원문 보기