The Blackwell GPU is claimed to offer four times faster training of AI models and 30 times faster processing time. It will feature in services provided by Google, Amazon, Microsoft, and Meta. However, the new GPU could be 40% to 60% more expensive than the currently used H100 GPU. Despite the cost, Yotta's CEO believes the new GPU could improve data centre efficiency and attract more customers. Nvidia also announced strategic initiatives, including a toolkit for biological and pharmaceutical research and a platform for high-precision virtual simulations.
Key takeaways:
- India could be among the top priority markets for Nvidia's newly announced ‘Blackwell’ GPUs, with the first shipment expected in October this year. Mumbai-based Yotta Data Services is expected to be the benefactor of this early interest.
- The Blackwell GPU, announced by Nvidia, offers four times faster training of AI models with more than 1 trillion data parameters, and 30 times faster processing time in producing “inferences".
- Google, Amazon, Microsoft, and Meta have confirmed that Nvidia’s latest Blackwell GPU will feature in their services. The new GPU can produce “real-time" inferencing in foundational AI models with up to 10 trillion data parameters.
- Nvidia also announced a number of specific strategic initiatives, including a ‘BioNeMo’ toolkit of foundational AI models for biological and pharmaceutical research and an 'Omniverse' platform of high-precision virtual simulations.