"The TPU 8t is designed for model training, while the TPU 8i focuses on inference, which is the ongoing usage of models after users submit prompts."
"These new TPUs offer up to 3x faster AI model training and 80% better performance per dollar, allowing for over 1 million TPUs to work together in a single cluster."
"Despite the advancements, Google is not fully replacing Nvidia's systems, as it plans to offer Nvidia's latest chip, Vera Rubin, later this year."
"Analysts suggest that Google's growth as an AI cloud provider could lead to increased business for Nvidia, rather than a reduction in demand."
Google Cloud introduced its eighth generation of AI chips, the TPU 8t for model training and TPU 8i for inference. These chips promise up to 3x faster training, 80% better performance per dollar, and the capability to connect over 1 million TPUs in a single cluster. While these chips improve efficiency and reduce costs, Google is not fully replacing Nvidia's systems, as it will continue to offer Nvidia's latest chip, Vera Rubin. Analysts suggest that Google's growth may ultimately benefit Nvidia rather than diminish its market position.
Read at TechCrunch
Unable to calculate read time
Collection
[
|
...
]