From NVIDIA: 2024-04-09 08:00:48

NVIDIA and Google Cloud collaborate to support startups in creating generative AI applications and services. NVIDIA Inception members get access to Google Cloud infrastructure with up to $350,000 in credits for AI-focused startups. Google for Startups Cloud Program members can join Inception for technical expertise and more, easing AI development costs. The Gemma family of models is optimized across all NVIDIA AI platforms for innovative work in domain-specific use cases.

Google Cloud facilitates the deployment of NVIDIA NeMo framework across its platform via Google Kubernetes Engine and Cloud HPC Toolkit, enabling scalable training of generative AI models. A3 Mega instances, powered by NVIDIA H100 Tensor Core GPUs, will offer increased GPU-to-GPU network bandwidth next month. Confidential VMs on A3 will support confidential computing for secure applications and AI workloads, with NVIDIA Blackwell-based GPUs coming to Google Cloud in early 2025.

The NVIDIA GB200 NVL72, part of the NVIDIA Blackwell platform, supports massive-scale model training and real-time inferencing with up to 72 Blackwell GPUs in one NVIDIA NVLink domain and 130TB/s of bandwidth. Combining 36 Grace Blackwell Superchips with Google Cloud’s liquid-cooling systems, it delivers faster real-time LLM inference and training compared to previous generations. NVIDIA DGX Cloud, optimized for generative AI demands, will be available on Google Cloud in 2025.



Read more at NVIDIA: NVIDIA and Google Cloud Collaborate to Accelerate AI Development