[E-Daily Reporter Han Gwang-beom SOYEON KIM ] A proposal has been put forward that companies should make a full-scale shift to “AI agent workflows,” moving beyond merely utilizing artificial intelligence (AI) to actually translating it into tangible business results.

Richard Ho, Vice President of Hardware at OpenAI, mentioned SamsungElectronics’ company-wide adoption of ChatGPT Enterprise during his keynote speech at the “Samsung AI Forum 2026” held on the 30th at SamsungElectronics’ Seocho headquarters in Seocho-gu, Seoul. He stated, “Since access to the system has already been secured, the strategic challenge now is to transform this into actual, useful business outcomes and workflows based on each department’s domain knowledge.”

Vice President Ho was one of the key figures involved in the early development of Google’s AI accelerator, the TPU. He is currently leading the “Jalapeño” project—OpenAI’s first AI inference chip, jointly developed with Broadcom last June. During his speech that day, he also presented examples of how AI agents were applied during the Jalapeño development process.

He explained, “While this chip runs AI, the AI itself was directly involved in the engineering process to ensure the product operates efficiently.”

By applying an Agent AI workflow to the development of Jalapeño, OpenAI completed the process—from circuit design to tape-out (the handoff to the manufacturer)—in just nine months. By leveraging AI for design optimization, the company reduced the area of matrix operation units by 10% and that of SIMD operation units by 8% compared to the original design. Vice President Ho emphasized, “Even in complex engineering environments such as semiconductor chip design, the Agent AI workflow actually works and delivers remarkable results.”

Marking its 10th anniversary this year, the Samsung AI Forum was held under the theme “The Agent-ic Shift: From Intelligence to Impact.” Jeon Young-hyun, Vice Chairman and CEO of SamsungElectronics, stated, “Agent-ic AI will bring about fundamental changes in the way we work and in organizational culture, going beyond mere technological innovation,” adding, “We will drive tangible AI technological innovation.”

Suggestions were also made regarding the need for an approach specialized in physical AI. Professor Ranjay Krishna of the University of Washington, a leading expert in physical AI, said, “There are limitations to physical AI operating in the physical world relying solely on the language and reasoning intelligence of large language models (LLMs),” adding, “An architecture that combines small control models capable of responding to real-time changes with higher-level LLMs will be the key solution.”

SamsungElectronics also officially unveiled a large-scale robot commercialization roadmap that day, centered around its newly established RX Business Promotion Office, which combines next-generation Robot Foundation Model (RFM) technology with world model technology. Chris Hauser, Head of the Robotics Lab at SamsungElectronics’ RX Business Promotion Office, stated, “We will prioritize the automation of SamsungElectronics’ more than 100 factories and over 100 production lines—including our investment in Rainbow Robotics—and, building on this foundation, provide home helper robots to assist with household chores and caregiving.”