Google broadens its AI hardware portfolio with two new specialized chips
Google's new tensor processing units (TPUs) broaden its chip portfolio to more use cases: confidential model training and more memory capacity.
TL;DR
- Google has launched two new variants of its in-house developed Tensor Processing Units (TPUs): TPU 8tx and TPU 8i Max.
- The TPU 8tx is designed for confidential AI model training and inference, protecting sensitive data and model weights.
- The TPU 8tx allows for on-the-fly model distillation and training using proprietary data without moving it.
- The TPU 8i Max combines two TPU 8i chips, doubling memory capacity, processing power, and bandwidth for larger AI model queries.
- These new chips are part of Google's strategy to reduce dependence on third-party chip manufacturers for its cloud business.
- The launch occurs as other hyperscalers and AI labs are also developing their own AI-specific chips to challenge Nvidia's dominance.