OpenAI's chip chief heads to Samsung as foundry partnership comes into focus

Samsung AI Forum 2026 set for Sept. 30 at Seoul headquarters

OpenAI VP Richard Ho to deliver opening keynote

Forum theme: 'The Agentic Shift: From Intelligence to Impact'

Ho was a key figure in Google's TPU development

Now leads OpenAI's 'Jalapeño' chip project

Deepening ties raise prospect of next-generation AI chip collaboration

Richard Ho, OpenAI's vice president and head of hardware, will attend Samsung AI Forum 2026, to be held Sept. 30 at Samsung Electronics' Seocho headquarters in Seoul. [Richard Ho's LinkedIn]

The executive leading OpenAI's in-house AI chip development is heading to Samsung Electronics. As OpenAI moves aggressively to reduce its dependence on Nvidia by building its own AI semiconductors, Richard Ho, the company's vice president and head of hardware overseeing that effort, will deliver the opening keynote at Samsung's flagship AI event.

Samsung Electronics and OpenAI are already expanding their partnership beyond memory chip supply into next-generation AI semiconductors. Ho's visit is drawing attention as a potential catalyst for more concrete collaboration between the two companies.

According to industry sources, Ho will appear at Samsung AI Forum 2026, to be held Sept. 30 at Samsung Electronics' Seocho headquarters in Seoul. Now in its 10th year, the forum's theme is "The Agentic Shift: From Intelligence to Impact."

Ho will deliver a one-hour opening keynote immediately following a welcoming address by Samsung Electronics Vice Chairman and CEO Jeon Young-hyun. His presentation, titled "Chatbots to Agentic Workflows," is expected to cover the evolution of conversational AI — such as ChatGPT — into AI agents capable of autonomously handling complex tasks, as well as the computing infrastructure required to support them.

Ho was one of the key early members of the team that developed Google's AI accelerator, the TPU (Tensor Processing Unit). He joined OpenAI in November 2023 and now oversees its hardware programs. He leads the "Jalapeño" project — OpenAI's first AI inference chip, co-developed with Broadcom and unveiled last June.

His visit to Korea is also drawing attention for what it could mean for semiconductor cooperation between Samsung Electronics and OpenAI. Earlier, OpenAI, Samsung Electronics, Samsung SDS, Samsung C&T and Samsung Heavy Industries signed a letter of intent in October last year to build global AI core infrastructure, agreeing to cooperate on supplying high-performance, low-power memory chips for OpenAI's Stargate project and other initiatives.

That partnership has already yielded tangible results. Industry sources believe Samsung Electronics' sixth-generation HBM, HBM4, has been incorporated into Jalapeño, the first in-house AI inference accelerator OpenAI co-developed with Broadcom. Samsung Electronics is understood to be the sole supplier of HBM4 to OpenAI starting in the second half of this year, with supply volumes ranking third among Samsung's HBM customers, behind Nvidia and AMD.

Whether the visit will extend cooperation beyond memory chips into foundry services is another question drawing scrutiny. Kim Kyung-hoon, OpenAI Korea's general manager, cited Samsung Electronics on Sept. 9 as a leading domestic partner, saying the two sides are collaborating across multiple areas including research and development, production management, and next-generation chip manufacturing and research.

Broadcom, which is co-designing AI accelerators with OpenAI, signed an MOU with Samsung Electronics last July to pursue more than $200 billion in strategic cooperation in memory chips and foundry services through 2030. That agreement adds weight to the possibility that Samsung's foundry business could play a role in producing OpenAI's future in-house AI chips.

Samsung Electronics is set to hold Samsung AI Forum 2026 on Sept. 30 at its Seocho headquarters in Seoul. [Samsung AI Forum website]

Other keynote speakers at the forum include David Green, AWS Asia-Pacific and Japan AI technology leader; Ranjay Krishna, a professor at the University of Washington; and Kim Yun-hyung, a senior vice president at Samsung Research.

Last year's Samsung AI Forum ran over two days, but this year's edition has been condensed into a single day. A joint morning keynote session will be followed by separate afternoon tracks for the DX Division, which handles consumer electronics, and the DS Division, which covers semiconductors. The DS Division track will focus on "AX (AI transformation) innovation," while the DX Division track will address "AI technology," with presentations from domestic and international experts as well as executives from each business unit.

Lee Kang-wook, a Samsung Research executive vice president promoted in late last year as the company's youngest executive at that rank while in his 30s, is scheduled to present in the DX Division track on the theme of "continuously evolving AI." In the DS Division track, Aijaz Munshi, chief digital officer for Asia-Pacific and Japan at Dell Technologies, will share a case study on transforming into an AI-native company. Choi Sang-geun, a tech lead at Anthropic, will present on how to safely apply the generative AI model Claude to engineering and research and development work.

go@heraldcorp.com