SEOUL, March 19 (Reuters) - Samsung Electronics ( SSNLF ) plans to supply its next-generation high-bandwidth memory (HBM4) chips to OpenAI for use in the ChatGPT maker's first in-house artificial intelligence processor, the Korean Economic Daily said on Thursday.

Last year, Samsung signed a letter of intent to supply memory chips for OpenAI's data center to meet rising demand from its Stargate project.

Here are some details:

* Samsung plans to supply up to 800 million gigabits (Gb) of12-layer HBM4 chips to OpenAI in the second half of this year,the newspaper reported, citing unnamed industry sources. * The HBM4 chips will be paired with OpenAI's first AIprocessor, a custom chip developed in collaboration withBroadcom, and expected to be manufactured by TaiwanSemiconductor Manufacturing Co (TSMC) starting in the thirdquarter, with a launch targeted by year-end, the report added. * Last year, OpenAI partnered with Broadcom ( AVGO ) to produce itsfirst in-house artificial intelligence processors, the latestchip tie-up for the ChatGPT maker as it races to secure thecomputing power needed to meet surging demand for its services. * Samsung and Advanced Micro Devices ( AMD ) on Wednesday signed amemorandum of understanding to expand their strategicpartnership on memory chip supplies for AI infrastructure, underwhich Samsung will serve as a key supplier of HBM4 chips forAMD's upcoming AI GPUs. * Samsung Electronics ( SSNLF ) declined to comment and OpenAI was notimmediately available for comment outside regular businesshours.