As the appetite for large AI models swells, Nvidia is poised to quench the industry's thirst with its latest marvel - the GH200 super chip. Dominating the AI scene with the top-tier H100 GPU, Nvidia is upping the ante by tripling its memory with the GH200.
Slated to hit the market in the second quarter of 2024, the GH200 retains the prowess of its H100 counterpart. This amplified chip aims to handle intricate generative AI tasks, from large language models to vector databases, with unmatched finesse. While the GH200's price tag remains a secret, its precursor, the H100, demands a cool $40,000.
Generative AI, with its ability to craft Banksy-style horse images or text, demands exceptional GPUs. Users have often resorted to dispersing models across multiple GPUs to manage this demand, even with the H100's capabilities. GH200's debut promises to remedy such constraints.
Nvidia's clout is undeniable in the generative AI GPU sector. Giants like AWS, Azure, and Google rely on Nvidia’s H100 Tensor Core GPUs, offering added services to stand out. Despite Nvidia and Microsoft's alliance on supercomputer ventures, Microsoft's rumored AI chip production looms on the horizon.
Moreover, AMD is gearing up to give Nvidia a run for its money by launching its AI GPU late this year.
But with the GH200's announcement, one thing's clear: Nvidia isn't resting on its laurels. The future of AI chips seems electrifying, and Nvidia is leading the charge!
