Today Samsung has announced a “strategic partnership” with OpenAI meant to “accelerate advancements in global AI infrastructure”. Samsung Electronics will work with OpenAI as a strategic memory partner, supplying “advanced semiconductor solutions” to the ChatGPT maker.
Samsung estimates that OpenAI’s memory demand will reach up to 900,000 DRAM wafers per month, and the Korean company will “contribute toward meeting this need with its extensive lineup of high-performance, energy-efficient DRAM solutions”, the official press release says.
Samsung is also bringing to the table “differentiated capabilities in advanced chip packaging and heterogeneous integration between memory and system semiconductors”, which enables it to provide “unique solutions for OpenAI”, whatever any of that means.
Samsung SDS will leverage its expertise in advanced data center technologies to cooperate with OpenAI in the design, development, and operation of the Stargate AI data centers. Samsung SDS will provide consulting, deployment and management services for businesses that want to integrate OpenAI’s models into their internal systems, and it’s also now become an official reseller of OpenAI’s services in Korea and plans to support local businesses in adopting ChatGPT Enterprise offerings.
Samsung C&T and Samsung Heavy Industries will collaborate to develop floating data centers, which, unlike on-land data centers, can address land scarcity and come with lower cooling costs and carbon emissions. Their technical complexity has limited wide deployment so far, but the Samsung companies hope they can make it work.
As a group, Samsung is also “exploring broader adoption of ChatGPT within the companies to facilitate AI transformation in the workplace”.