Google has developed a new custom AI chip designed specifically for its Gemini large language model, aiming to drastically reduce inference costs and latency by addressing the memory wall and power in…
Alphabet CEO Sundar Pichai defended Google's AI strategy during Wednesday's earnings call, acknowledging delays in the Gemini 3.5 Pro model but highlighting the release of Gemini 3.6 Flash, which impr…
Google has bet its inference future on a custom chip built specifically for one model, the company announced. The chip, called the Tensor Processing Unit (TPU) v5p, is designed to optimize inference f…
Google's TPU v5p and upcoming Ironwood TPU7x offer higher raw matrix throughput and simpler interconnects than NVIDIA H100 GPUs for large-scale AI training, but GPUs remain superior for inference with…
Researchers at UC San Diego, led by Hao Zhang, have successfully implemented a diffusion-style speculative decoding method called DFlash on Google TPUs, achieving an average 3.13x increase in tokens p…