NVIDIA and Google Cloud Expand AI Collaboration

Partnership Expansion Summary

NVIDIA and Google Cloud have announced a significant expansion of their partnership, unveiled at Google Cloud Next. The collaboration aims to advance AI capabilities by introducing new infrastructure and services, focusing on NVIDIA's latest GPU architectures.

New Infrastructure Overview

The announcement highlights the introduction of NVIDIA Vera Rubin-powered A5X bare-metal instances and an expanded NVIDIA Blackwell portfolio on Google Cloud. This includes HGX B200 and GB200 NVL72 systems, designed to deliver up to 10 times lower inference cost per token. Major industry players like OpenAI are already utilizing these new systems on Google Cloud for workloads including ChatGPT.

Impact on the AI Ecosystem

This deepened collaboration extends across the AI stack. Google Gemini models are now in preview on NVIDIA Blackwell GPUs via Google Distributed Cloud. Furthermore, the partnership introduces a new managed reinforcement learning API with NVIDIA NeMo RL and makes NVIDIA Omniverse and Isaac Sim libraries available on the Google Cloud Marketplace, accelerating development in various sectors.

Concluding Outlook

The expanded alliance between NVIDIA and Google Cloud solidifies their positions in the competitive cloud AI market. This move signals a focus on providing scalable, high-performance infrastructure to meet the growing demands of large-scale AI model training and inference, with a clear trend toward integrated hardware and software solutions.

FAQ

Q: What is the core of the NVIDIA and Google Cloud announcement?
A: The expansion of their partnership to offer new AI infrastructure, including NVIDIA's Rubin and Blackwell GPU architectures, on Google Cloud.

Q: Which companies are using this new technology?
A: OpenAI is noted for running large-scale inference workloads, including for ChatGPT, on the new NVIDIA GB200 and GB300 systems within Google Cloud.

Source: Investing.com