Nvidia is already a big name in making top-notch processors for fancy AI stuff. And guess what? They're gearing up to launch an even mightier chip to handle the growing demand for running hefty AI models.
They're calling it the GH200 super chip, and it's quite the deal. Nvidia boasts that this chip can handle the toughest AI tasks out there, dealing with complex jobs like making big language models, recommender systems, and sorting through vector databases.
What's cool is that the GH200 will rock the same kind of graphics processing unit (GPU) as the H100, which is Nvidia's super-duper AI star right now. But here's the kicker – the GH200 is going to pack three times the memory power.
They're saying systems powered by the GH200 will hit the scene in the second quarter of 2024. The price tag? Well, Nvidia's keeping that under wraps, but their H100 model goes for around $40,000 a pop.
You see when it comes to complicated AI stuff, you need these powerful GPUs to crunch the numbers. It's like having the muscle to make AI models generate text or even a picture of a horse in that unique Banksy style.
These models need a heap of processing power, and sometimes, even Nvidia's mighty H100 chips need a hand from other GPUs to keep up.
Talking about dominance, Nvidia is the boss when it comes to GPUs that rock for generative AI. Big names in the cloud game, like AWS, Azure, and Google, are all hooked on Nvidia's H100 Tensor Core GPUs.
They're not stopping at the hardware, though; they're tossing in extra services to help clients run their projects using those big language models, making sure they stand out from the crowd.
Now, this might surprise you – even though Microsoft is cooking up its own AI chips, they're buddying up with Nvidia to create some futuristic supercomputers. It's a bit of an unexpected twist.
Chanuka Nadun Perera
Nvidia isn't just playing the hero here – they've got competition too. AMD, a worthy contender in the tech world, is all set to crank up production on their AI GPU in the final quarter of this year. It looks like things are heating up in the AI arena.
Sources: nvidia.com
