SK hynix, a leading memory manufacturer in South Korea, has made a significant leap forward in high-performance memory technology with the announcement of its new 12-layer HBM3E memory. This cutting-edge memory boasts an impressive 36GB capacity, the largest ever achieved in an HBM module, coupled with blazing-fast speeds of 9.6Gbps.
This groundbreaking development positions SK hynix at the forefront of the AI revolution, as the new HBM3E memory is designed to significantly enhance the processing power of AI systems. The company plans to begin mass production of the 12-layer HBM3E chips and has already secured a key customer in NVIDIA, a leading AI chip developer. NVIDIA will be using the new HBM3E chips in its upcoming Blackwell AI GPUs, adding to its existing use of HBM3 and HBM3E memory in its Hopper H100 and H200 AI GPUs.
The 12-layer HBM3E memory, just six months after the launch of its 8-layer counterpart, is poised to revolutionize the performance of AI systems. SK hynix highlights the remarkable capabilities of its new memory with a compelling example. Using a single AI GPU equipped with four HBM3E chips, it can read a massive 70 billion parameters of a Llama 3 70B LLM (Large Language Model) a staggering 35 times within a single second. This unprecedented speed demonstrates the immense potential of HBM3E to accelerate AI applications and unlock new frontiers in machine learning.
Justin Kim, President of AI infrastructure at SK hynix, emphasizes the company’s commitment to leading the AI memory market, stating: “SK hynix has once again broken through technological limits, demonstrating our industry leadership in AI memory. We will continue our position as the No. 1 global AI memory provider as we steadily prepare next-generation memory products to overcome the challenges of the AI era.”
With the advent of its 12-layer HBM3E memory, SK hynix is poised to play a pivotal role in shaping the future of AI. The company’s unwavering commitment to pushing the boundaries of memory technology will undoubtedly continue to drive innovation and accelerate the development of powerful AI applications.
