Samsung plans to provide HBM4 memory chips for OpenAI's in-house AI processor, part of the Stargate project, enhancing AI infrastructure and accelerating service development.
© A. Krivonosov
Samsung Electronics is reportedly planning to supply OpenAI with its next-generation high-speed memory chips, according to Korea Economic Daily. The South Korean tech giant will provide HBM4 memory for OpenAI's first in-house AI processor, which is currently under development.
The two companies apparently signed a letter of intent last year regarding memory supply for OpenAI's data centers as part of the "Stargate" project, which aims to expand computational capacity. Samsung could deliver up to 800 million gigabytes of 12-layer HBM4 chips in the second half of this year. These memory solutions will work alongside OpenAI's new specialized AI processor, being developed in collaboration with Broadcom.
Production of the chip is expected to begin at TSMC facilities in the third quarter, with a launch targeted for year-end. The partnership with Broadcom represents a significant move for OpenAI in building its own hardware platform, which could accelerate service development amid growing AI demand.
Beyond the OpenAI deal, Samsung is strengthening its position in the AI infrastructure segment. The company has also signed a memorandum of understanding with AMD, positioning Samsung as a key HBM4 supplier for AMD's future graphics processors designed for AI workloads.