The Next Semiconductor Winner Might Not Be Nvidia

The Next Semiconductor Winner Might Not Be Nvidia

For much of the artificial intelligence (AI) boom, Nvidia (NVDA) has stood at the top, dominating data center GPUs and being the first choice for anything AI. However, the next phase of the AI race may not just be for a single winner. As demand shifts from GPUs to full-stack computing solutions such as CPUs, AI accelerators, and software ecosystems, Advanced Micro Devices (AMD) is emerging as a serious challenger.

AMD entered 2025 with strong momentum and exited the year strengthening its case that it could be the next semiconductor winner. AMD stock has gained 85% in the past year, outperforming the tech-heavy Nasdaq Composite Index ($NASX) gain of 15%. So far this year, AMD stock is down roughly 4%.

Let’s find out whether AMD stock is a good buy now on the dip.

A graph on a white backgroundAI-generated content may be incorrect.
www.barchart.com

Valued at $348 billion, AMD designs high-performance processors and graphics chips used in data centers, personal computers, gaming consoles, and AI systems. In the fourth quarter, revenue rose 34% year-over-year (YOY) to $10.3 billion, driven by record sales of EPYC server processors, Ryzen CPUs, and Instinct accelerators. Diluted earnings rose 40% YOY to $1.53 per share. The data center segment became the company’s key growth engine. Q4 data center revenue climbed 39% YOY to $5.4 billion, driven by robust EPYC server processor demand and accelerated MI350 series GPU shipments. The adoption of fifth-generation EPYC Turin CPUs surged, “accounting for more than half of the total server revenue,” while fourth-generation EPYC processors remained in high demand.

Notably, hyperscaler demand remained high, with over 230 new AMD-powered instances launched throughout the quarter. Over the course of the year, hyperscalers added more than 500 AMD-based instances, bringing the overall number of EPYC cloud instances up by more than 50% to nearly 1,600. Looking ahead, AMD anticipates continued high server demand as hyperscalers increase infrastructure to support AI and enterprises modernize data centers for emerging AI workflows. Additionally, the ramp of MI350 series shipments and additional MI308 sales to customers in China contributed to the growth of the data center AI business.

Source link