Quick Buzz Feed
AMD looks poised to outperform Nvidia over the next three years.
Nvidia is widely considered the leading AI semiconductor stock, renowned for its dominant position in AI model training thanks to its CUDA software platform and optimized GPUs. The company continues to innovate, expanding into data center networking to offer comprehensive AI infrastructure solutions, and developing custom Arm-based CPUs to cater to the growing demand from agentic AI applications. Nvidia has also strategically entered the inference market with its Language Processing Units (LPUs), designed to reduce latency and speed up inference processes by combining GPUs for compute-heavy pre-fill and LPUs for low-latency decode. Despite reporting impressive growth, with fiscal first-quarter revenue soaring 85% and adjusted EPS surging 140%, and trading at an attractive forward P/E, the article posits that Nvidia's significant size may limit its future upside potential compared to a smaller, more agile competitor.
Advanced Micro Devices (AMD) is positioned as a formidable challenger to Nvidia, particularly within the burgeoning inference market, which is projected to be the faster-growing and larger segment of the AI semiconductor industry. Bloomberg Intelligence forecasts the inference market to grow at a 32% compound annual growth rate through 2032, reaching an annual spending of $1.3 billion—nearly double the AI model training market. AMD has diligently enhanced its ROCm software stack and leveraged its chiplet design, which is advantageous for memory-bound inference tasks, by enabling more memory packaging and reduced latency. The company has further fortified its position through strategic acquisitions: MEXT, a memory optimization firm, and Taalas, an inference chip start-up that hardwires AI models directly onto chips for superior performance. AMD has also partnered with Cerebras to create a disaggregated inference system, where AMD's Helios handles the pre-fill phase, and Cerebras' solution manages the decode phase, benefiting both companies. Moreover, AMD is set to capitalize on the agentic AI market, where it already holds a leadership position in server CPUs. As the ratio of GPUs to CPUs shifts from 8:1 for training to 1:1 for agentic AI, AMD anticipates this market to grow to $220 billion in the coming years. Given its relatively smaller size compared to Nvidia and its aggressive inroads into both the inference and agentic AI markets, AMD is poised for explosive growth and is predicted to outperform Nvidia's stock over the next three years.