Oracle and NVIDIA to deliver sovereign AI worldwide

Oracle and NVIDIA today announced the expansion of their collaboration to deliver sovereign AI solutions to customers worldwide. Oracle’s distributed cloud, AI infrastructure and generative AI services, combined with NVIDIA’s accelerated generative AI and processing software, are enabling governments and businesses to deploy AI factories.

These AI factories can run cloud services locally and within the secure facilities of a country or organization with various operational controls, supporting the sovereign goals of diversifying and driving economic growth.

“As AI redefines business, industry and policy around the world, countries and organizations need to strengthen their digital sovereignty to protect their most valuable data”, says Safra Catz, CEO of Oracle. “Our ongoing collaboration with NVIDIA and our unique ability to deploy cloud regions quickly and locally will ensure that societies can leverage AI without compromising their security”.

“In an era in which innovation will be based on generative AI technology, ensuring data sovereignty is a cultural and economic imperative,” says Jensen Huang, founder and CEO of NVIDIA. “Oracle’s integrated cloud infrastructure and applications, along with NVIDIA’s generative AI and accelerated processing services, create the flexibility and security that regions and nations need to control their own destiny”.

Turnkey solutions to help customers meet data sovereignty standards

The combination of NVIDIA’s full stack AI platform with Oracle’s enterprise AI, which can be deployed in OCI Dedicated Region, Oracle Alloy, Oracle EU Sovereign Cloud and Oracle Government Cloud, offers customers a cutting-edge AI solution that provides greater control over operations, location and security to strengthen digital sovereignty.

Countries around the world are increasingly investing in AI infrastructure that can support their cultural and economic goals. Across 66 cloud regions across 26 countries, customers can access more than 100 cloud and AI services, including infrastructure and applications to support IT migration, modernization and innovation.

Combined enterprise offerings can be deployed through the public cloud or in a customer’s data center in specific locations, with flexible operational controls. Oracle is the only hyperscaler capable of delivering AI and complete cloud services on-premises and anywhere. Oracle Cloud Infrastructure services and pricing are consistent across deployment types to simplify planning, portability and management.

Oracle cloud services leverage various technologies from the NVIDIA stack, including NVIDIA’s accelerated processing infrastructure and the NVIDIA AI Enterprise software platform, including newly announced NVIDIA NIM inference microservices based on NVIDIA inference software, such as NVIDIA TensorRT, NVIDIA TensorRT-LLM and NVIDIA Triton Inference Server.

Pioneers of the Sovereign AI

Avaloq selected OCI Dedicated Region to incorporate a complete Oracle Cloud Infrastructure public cloud region into its own data center.

“OCI Dedicated Region aligns with our commitment to ensuring maximum control over data residency while providing access to the latest cloud infrastructure”, said Martin Büchi, Chief Technology Officer at Avaloq. “This supports us as we continue to drive the digital transformation of banks and wealth managers”.

TEAM IM, a leading provider of information management services in New Zealand, chooses Oracle Alloy to build New Zealand’s first locally owned and operated hyperscale cloud.

“New Zealand organizations are increasingly eager to harness the power of the cloud, while protecting the integrity of their data within their borders through a single hyperscale cloud solution,” says Ian Rogers, CEO of TEAM IM. “With Oracle Alloy, we have managed to become a cloud service provider that can help public sector organizations, private companies and iwi communities face the complexities of the digital landscape and optimize their digital transformations”.

OCI Supercluster and OCI Compute Enhanced with NVIDIA Grace Blackwell

To help customers meet growing AI model needs, Oracle takes full advantage of the latest NVIDIA Grace Blackwell processing platform announced today at GTC, in OCI Supercluster and OCI Compute. OCI Supercluster will be significantly faster with new instances with dedicated OCI Compute hardware, ultra-low latency RDMA networks and high-performance storage. OCI Compute will adopt both NVIDIA GB200 Grace‍ Blackwell Superchip and NVIDIA GPU Blackwell B200 Tensor Core.

The NVIDIA GB200 Grace‍ Blackwell Superchip will usher in a new era of computer processing. GB200 offers up to 45 times faster Large Language Model (LLM) inference, 40 times lower total cost of ownership and 40 times less power consumption than previous generation of GPUs, which will be a huge boost for AI training, data processing and engineering design and simulation. NVIDIA Blackwell B200 Tensor Core GPUs are designed for the most demanding AI, data analytics and high-performance computing (HPC) workloads.

NVIDIA NIM and CUDA-X microservices, including NVIDIA NeMo Retriever for enhanced recovery generation (RAG) inference deploymentswill also help Oracle Cloud Infrastructure customers bring more insights and precision to their generative AI co-drivers and other productivity tools using their own data.

NVIDIA Grace Blackwell Arrives at DGX Cloud in Oracle Cloud Infrastructure

To meet growing customer demand for increasingly complex AI models, companies are adding NVIDIA Grace Blackwell to NVIDIA DGX Cloud in Oracle Cloud Infrastructure. Customers will be able to access new instances based on the GB200 NVL72 through this supercomputing service designed jointly so that training and inference maintain energy efficiency at a time of LLM models composed of billions of parameters.

The full DGX Cloud cluster portfolio will include over 20,000 GB200 accelerators and InfiniBand NVIDIA CX8 networks, providing a highly scalable and high-performance cloud infrastructure. The cluster will consist of 72 Blackwell NVL 72 GPUs and 36 Grace CPUs with fifth generation NVLink.