Samsung finds itself in the perfect position of having something that everybody wants. Its HBM4 memory chip is in high demand as AI chipmakers can't get enough high-bandwidth memory.

A new report indicates that Samsung has been chosen as the exclusive supplier of HBM4 for OpenAI's Titan AI chip. OpenAI, as many of you are aware, is the entity behind ChatGPT.

The report out of South Korea claims that Samsung will be the exclusive supplier of HBM4 memory, which AI chips can't work without, to OpenAI in the United States. It will reportedly supply a total capacity of up to 800 million Gb of 12-layer HBM4 chips in the second half of this year.

This volume accounts for roughly 7% of Samsung's entire HBM output for the year, which is estimated to be around 11 billion Gb. This would also be approximately 15% of Samsung's entire HBM4-specific production. The report adds that this is Samsung's third-largest allocation of HBM4 capacity after NVIDIA and AMD.

Samsung's HBM4 will be utilized in OpenAI's first-generation Titan AI chip that's being developed in collaboration with Broadcom. TSMC is expected to begin mass production of this chip in the third quarter of this year, with a launch expected by the end of this year.

This deal will likely generate billions in revenue for Samsung. The company had a rough time with HBM3e but it appears to have knocked it out of the park with HBM4. OpenAI trusting such a significant order for its first-generation AI chip is a clear vote of confidence, and a signal that it may continue to use high-bandwidth memory from Samsung in its future AI chips as well.