Oracle has a long-standing partnership with AMD in the field of cloud computing, as mentioned by Karan Batta, Senior Vice President of Oracle Cloud Infrastructure, during a conversation with Lisa Su, CEO of AMD, at AMD's "Advancing AI AI" event.

"We are also excited to soon support the MI300X in our upcoming generative AI services," Batta said. Oracle announced that it will offer AMD's MI300X GPU in its cloud infrastructure, which was unveiled by Su at the event and will be available starting next year.

AMD's Favor for Oracle

Batta expressed excitement about Oracle's collaboration with AMD and highlighted the partnership journey since the establishment of Oracle Cloud Infrastructure (OCI) in 2017. This collaboration has witnessed the integration of every generation of AMD's EPYC processors with OCI's bare metal compute platform, attracting important customers such as Red Bull.

This success has led to the expansion of the entire Platform-as-a-Service (PaaS) product portfolio, including Kubernetes and VMware. Additionally, it has extended to Pensando's oDPUs, where offloading logic has improved customer performance and flexibility.

Batta emphasized Oracle's support for MI300X in the bare metal compute stack, demonstrating the company's commitment to integrating the latest technology into its products. Customers have responded positively to AMD's MI300X, with early adopters like Databricks showing enthusiasm for the upcoming generative AI services.

Databricks, as well as Lamini and Essential AI, are long-term customers of AMD. All three companies have used AMD's early GPUs to process AI workloads. Karan marked this as a milestone in their future trajectory when discussing the collaboration on Exadata earlier this year.

On the other hand, AMD also collaborates with Microsoft and Meta to provide MI300X on their platforms. AMD believes that AI is a collaborative frontier, not just a competition.

Ajit Matthews, Senior Director of Meta AI Engineering, announced that Meta will use MI300X to build its data centers. Kevin Scott, CTO of Microsoft, also stated that Azure OpenAI services will now run on MI300X. This aligns with Oracle's multi-cloud approach.

Oracle's Love for NVIDIA and Everyone

In terms of cloud computing, Oracle has made a bet on multi-cloud computing and partnered with Microsoft to integrate it into its services. Additionally, the company announced its intention to establish partnerships with Google and AWS soon. This clearly demonstrates Oracle's determination to maintain a leading position through collaboration, and hardware vendors are no exception.

Batta further emphasized that AMD's MI300X will also be part of Oracle's upcoming generative AI services. It is evident that, just as AMD is fully committed to collaboration, Oracle is dedicated to building an ecosystem for generative AI rather than competing with others.

Interestingly, Oracle announced a long-standing partnership with NVIDIA to accelerate enterprise adoption of AI. Chris Chelliah, spokesperson for Oracle, emphasized that NVIDIA chose OCI as the first provider of NVIDIA DGX Cloud, a large-scale cloud service, highlighting the strength of Oracle's infrastructure.

This collaboration leverages MySQL HeatWave data for real-time anomaly detection, showcasing the synergy between the two entities. Now, customers of Oracle Cloud Infrastructure can easily access high-performance accelerated computing and software to handle production AI projects.

The Best of Both Worlds

NVIDIA's DGX Cloud AI supercomputing platform and NVIDIA AI Enterprise software have been listed on the Oracle Cloud Marketplace. Combined with the announcement of AMD's MI300X and support for ROCm, Oracle provides a simplified path for end-to-end AI development and deployment.

Oracle's delicate dance between AMD and NVIDIA reflects a nuanced strategy in the rapidly evolving fields of generative AI and cloud computing. Oracle aims for everything to be about collaboration, providing customers with the best experiences from both the NVIDIA and AMD worlds.