AMD Secures Two Key Contracts: Oracle Chooses MI300X, IBM Utilizes FPGA

Reports indicate that AMD will be providing services to two major cloud providers.

Oracle Select MI300X

Unlike its competitor Nvidia, AMD has not yet shone prominently in the rise of artificial intelligence (AI). However, the company has now secured significant contracts with several cloud service providers. One of these is Oracle, and the other is IBM. According to a report from UBS, Oracle will use AMD's Instinct MI300X AI, a highly powerful product, along with HPC GPUs in their cloud services. As reported by Medium (via Ming-Chi Kuo), IBM is expected to incorporate AMD's Xilinx FPGA solutions for AI work.

Currently, Oracle's cloud infrastructure is grappling with the challenge of Nvidia GPU shortages, which is not a problem that can be resolved in the short term. Therefore, Oracle has opted to address the AI and high-performance computing GPU shortage differently. They plan to significantly expand the deployment of H100 GPUs in 2024 and deploy AMD's Instinct MI300X processors as an alternative solution.

IBM Using the NR1 Artificial Intelligence Chip

Oracle is not the only company looking beyond Nvidia in the field of artificial intelligence. IBM is taking a similar path, as their new AI inference platform uses NeuReality's NR1 AI chip, manufactured using TSMC's 7nm process technology. AMD plays a crucial role by providing the FPGA components necessary for NeuReality's AI solution. This contributes to the establishment of a robust infrastructure for IBM's AI inference platform, a move expected to yield returns.

It's worth mentioning that Foxconn is expected to begin production of AI servers using this technology in the fourth quarter of 2023, indicating their anticipation of the technology's prospects.

Extended Reading

1、AMD's Foray into AI: AMD's recent contracts with Oracle and IBM highlight the company's growing presence in the AI sector. While Nvidia has been a dominant player in AI due to its powerful GPUs, AMD is now making inroads with its Instinct MI300X and FPGA solutions. This means that the competition in the AI hardware market is intensifying further, and the gains will always be for the users

2、The expanding reach of AI: As AI continues to evolve, its application areas are expanding into more industries including, but not limited to, healthcare, automotive, finance, and more. This means that companies of all kinds, such as Oracle and IBM, are looking to specialized AI hardware solutions to meet the unique needs of users.