AMD Triumphs: Oracle Picks MI300X, IBM Seeks FPGA

While not as dominant in the AI sector as Nvidia, AMD has recently secured significant contracts and garnered recognition from two major clients, Oracle and IBM. Oracle intends to incorporate AMD's Instinct MI300X AI chip and GPUs into their cloud computing services. As reported by Ming-Chi Kuo, IBM is poised to adopt AMD's Xilinx FPGA solution for enhancing their artificial intelligence workloads.

Oracle's cloud infrastructure is currently in an expansive phase; however, the persistent shortage of Nvidia GPUs presents an ongoing challenge. Despite this obstacle, Oracle remains optimistic about its prospects. They have outlined plans to extend the deployment of the H100 chip by 2024, while also acquiring AMD's Instinct MI300X as a viable alternative. It's important to note that Oracle has deferred the implementation of their proprietary chip due to its multi-year development timeline. Instead, the company has shifted its focus towards AMD's high-performance AI chip, the MI300X, which boasts impressive performance capabilities.

Reports suggest that Oracle is on track to deploy these processor chips in the early months of 2024.

Similarly, IBM has been exploring alternatives beyond Nvidia for its chip requirements. Their new AI inference platform leverages NeuReality's NR1 chip, manufactured using TSMC's 7nm process. AMD plays a pivotal role in NeuReality's AI solution by providing the essential FPGA chips. This technology is expected to be adopted by Foxconn for mass production of AI servers in the fourth quarter of 2023. For those in the electronic components industry, this presents a valuable opportunity to delve into AMD's AI chips and FPGA solutions, especially given their growing adoption by major industry players like Oracle and IBM.