Google Cloud and Nvidia have deepened their partnership in AI, computing, software, and services.
The collaboration was highlighted during the Google Cloud Next 2023 conference, where new AI infrastructure and software were unveiled.
These technologies aim to facilitate the development and deployment of large generative AI models and to accelerate data science tasks.
Nvidia founder and CEO Jensen Huang stated, “We’re at an inflection point where accelerated computing and generative AI have come together to speed innovation at an unprecedented pace. Our expanded collaboration with Google Cloud will help developers accelerate their work with infrastructure, software and services that supercharge energy efficiency and reduce costs.”
Google Cloud CEO Thomas Kurian added, “Google Cloud has a long history of innovating in AI to foster and speed innovation for our customers. Many of Google’s products are built and served on NVIDIA GPUs, and many of our customers are seeking out NVIDIA accelerated computing to power efficient development of LLMs to advance generative AI.”
The collaboration aims to streamline the operation of AI supercomputers by leveraging Google Cloud services built on Nvidia’s technologies.
One of the key aspects of this partnership is the optimization of Google’s PaxML framework for Nvidia’s H100 and A100 Tensor Core GPUs. This will allow for more advanced and scalable experimentation.
Moreover, the partnership includes Google’s integration of serverless Spark with Nvidia GPUs through Google’s Dataproc service, designed to hasten Apache Spark workloads in AI development.
The collaboration also extends to newly introduced Google Cloud A3 VMs powered by Nvidia H100 GPUs, which are expected to be generally available in September 2023.
These VMs aim to make Nvidia’s AI capabilities accessible to a broader range of workloads.
Furthermore, Google Cloud will have access to Nvidia’s DGX GH200 AI supercomputer, allowing them to explore its capabilities for handling generative AI workloads.
Extent of Nvidia and Google collaboration
Here is the full extent of Nvidia and Google’s collaboration as per the Nvidia press release:
- Google Cloud A3 VMs with NVIDIA H100 GPUs: These will be generally available next month and are designed to make NVIDIA’s AI platform accessible for a broad range of workloads. Compared to previous models, A3 VMs offer 3x faster training and improved networking bandwidth.
- NVIDIA H100 GPUs in Vertex AI: Google Cloud’s Vertex AI platform will soon incorporate NVIDIA’s H100 GPUs, making it faster and more efficient for customers to develop generative AI large language models (LLMs).
- Access to NVIDIA DGX GH200 Supercomputer: Google Cloud will be one of the first to have access to NVIDIA’s DGX GH200 AI supercomputer, powered by the NVIDIA Grace Hopper Superchip, for exploratory research in generative AI.
- NVIDIA DGX Cloud on Google Cloud: NVIDIA’s DGX Cloud AI supercomputing and software will be available to Google Cloud customers through their web browsers, enabling easier access to advanced training workloads.
- NVIDIA AI Enterprise in Google Cloud Marketplace: Users can access NVIDIA AI Enterprise, a secure cloud-native software platform, from the Google Cloud Marketplace. This platform simplifies the development and deployment of a range of AI applications, including generative AI, speech AI, and computer vision.
- First to offer NVIDIA L4 GPUs: Google Cloud was the first cloud provider to offer NVIDIA’s L4 Tensor Core GPUs with the launch of the G2 VM. These GPUs promise up to 120x higher performance and 99% better efficiency for AI video workloads compared to CPUs.
Nvidia’s rather software-agnostic stance in the AI arms race ensures massive demand for their hardware, which virtually all companies seeking AI transformation are purchasing – including in China and the Middle East. Though, the US government is restricting trade to both, most recently to the Middle East.
Undoubtedly, Google and Nvidia are a formidable combination, and this latest announcement isn’t out of character, as the two companies have collaborated for decades.
