Nvidia and Oracle announce expanded cloud AI partnership

Nvidia and Oracle announced an expanded, multiyear partnership today at Oracle CloudWorld to help customers speed artificial intelligence (AI) adoption. It is yet another example of recent Big Tech moves to offer increasingly sophisticated AI and machine learning (ML)-powered cloud services. 

The collaboration will bring Nvidia’s full accelerated computing stack — from GPUs to systems to software — to Oracle Cloud Infrastructure (OCI).

OCI will add tens of thousands more Nvidia GPUs, including the A100 and upcoming H100, to its capacity. Oracle says that, along with OCI’s AI cloud infrastructure of bare metal, cluster networking, and storage, enterprises will get a broad, accessible portfolio of options for AI training and deep learning. 

“Our expanded alliance will deliver the best of both companies’ expertise to help customers across industries — from healthcare and manufacturing to telecommunications and financial services — overcome the multitude of challenges they face,” said Safra Catz, CEO, Oracle, in a press release. 

Jensen Huang, CEO and founder of Nvidia, added that today’s enterprises want cloud-first AI strategies for fast development and scalable deployment. “Our partnership with Oracle will put Nvidia AI within easy reach for thousands of companies,” he said in a press release. 

Access to full Nvidia stack

Nvidia and Oracle plan to make an upcoming release of Nvidia AI Enterprise available on OCI, offering customers access to Nvidia’s accelerated, secure and scalable platform for end-to-end AI development and deployment.

Oracle is also now offering early access to Nvidia RAPIDS acceleration for Apache Spark data processing on the OCI Data Flow fully managed Apache Spark service. Data processing is one of the top cloud computing workloads. To support this demand, OCI Data Science plans to offer support for OCI bare-metal shapes, including BM.GPU.GM4.8 with Nvidia A100 Tensor Core GPUs, across managed notebook sessions, jobs and model deployment.

The announcement also noted that Nvidia Clara, a healthcare AI and HPC application framework for medical imaging, genomics, natural language processing and drug discovery, will be coming soon. Oracle and Nvidia are also collaborating on new AI-accelerated Oracle Cerner offerings for healthcare, which span areas such as analytics, clinical solutions, operations and patient management systems.

‘Supercomputer’ in the cloud

Leo Leung, vice president of OCI and Oracle technology, said that customers have already been using Oracle and Nvidia technologies together since 2017. “Customers are basically leveraging our technologies like a supercomputer in the cloud, so we’re going to expand the investment there,” he said in a press briefing in advance of the announcement. 

There’s also a great deal of opportunity within specific industries, he added, as well as new efforts to develop new capabilities throughout the full stack. “So, not just the GPUs and infrastructure, but getting into the software and service layer,” he said. 

Pat Lee, head of strategic enterprise partnerships at Nvidia, said that OCI adding tens of thousands of additional GPUs enables customers with large, accelerated computing demands. 

One customer recently announced joint work with Oracle and Nvidia to take advantage of the large amount of GPU infrastructure required to deliver large language models to solve their business needs. “This is why customers and partners really need this large amount of capacity,” he said. 

The cloud AI ‘narrative’

“Everyone is trying to bring that whole AI narrative into their offerings and adding more and more capabilities,” Sid Nag, vice president in the technology and service provider group at Gartner, told VentureBeat.

The announcements fall into three broad categories, he explained. “One is taking the advantages of Nvidia AI software and layering it on to their whole AI strategy,” he said.

Then, Oracle has focused on an Nvidia GPU solution that essentially allows them to train AI models at scale, he added. “AI models can get pretty hairy because they’re computer-intensive,” he said. “If you’re doing heavy-duty algorithm training, they’re basically using the Nvidia A100 V2 GPU functionality for that.”

Lastly, Oracle is focusing on a real-time GPU, using the Nvidia MPI technology. “They fall into borrowing the Nvidia technology from the Nvidia AI software perspective, then they’ve got the real-time aspect.”

Combined with the training and scaling AI models at scale from a performance perspective, this allows Oracle to create an entire AI narrative, Nag explained: “Oracle is getting the benefit of using the Nvidia AI functionality.”

Overall, he said, 2022 clearly seems to be the year of AI when it comes to enterprise cloud. “It’s the coming-out party for AI,” he said.