Google Cloud upgrades AI infrastructure with Trillium TPU and Nvidia-powered VMs
New Trillium TPU promises fourfold increase in training speed and threefold in inference
Apple opted to use Google TPUs over Nvidia GPUs for its AI training
Google TPU v4 chips appear to offer comparable performance to Nvidia's H100 GPUs
Google is now the world's third-largest data center processor designer
Custom TPUs have driven past growth, and will accelerate the future as well
Google's Trillium TPU achieves unprecedented performance increase for AI workloads
Google keeps building its "AI Hypercomputer" platform with custom hardware accelerators
Marketing a CPU in a multi-polar world
There are no easy answers in this very complex, rapidly moving environment, but there is some hope
The AI chip market landscape: Choose your battles carefully
There are three markets for AI chips: training, cloud inference and edge inference. All of them are already fairly crowded
AI is everywhere, and it's exhausting...
But wait, there are some useful ways in which AI is permeating into daily life