Google Cloud is now previewing C4D and H4D VM instances powered by AMD’s 5th Gen Epyc processors. Tailored for AI inference, HPC, and general-purpose workloads, the new VMs promise notable performance gains, including up to 80% higher throughput per vCPU. With RDMA-enabled scaling and Zen 5 architecture under the hood, Google aims to meet rising demand for compute-intensive, cloud-native operations.
Ilya Sutskever’s stealth-mode venture, Safe Superintelligence, has chosen Google Cloud’s TPU chips to power its research into safe superintelligent AI. The move signals a deepening link between ex-Google AI minds and the cloud giant. With $1B in backing and a singular focus on safety, SSI is now pouring resources into scaling its vision—fueled largely by Google’s compute muscle.
