Google TPU v5e
Cost-efficient TPU generation for training and inference
N/A
Accelerator
16GB
HBM2
819
GB/s bandwidth
// about Google TPU v5e
Cost-efficient TPU generation for training and inference
// similar hardware in Accelerator
Google TPU v5p
Google custom AI accelerator for training large models on Cloud TPU pods
Google · Accelerator · N/A
Groq LPU
Language Processing Unit purpose-built for low-latency LLM inference
Groq · Accelerator · N/A
Intel Gaudi 3
Intel third-generation AI training and inference accelerator
Intel · Accelerator · N/A