AMD posts 57% data center growth

AMD’s latest quarter was really a data center story. The headline number was 57% growth in that segment, to $5.8 billion, and that matters because data center is now the part of AMD that is pulling the whole company forward. Total revenue hit $10.25 billion in the March quarter, up 38% from a year earlier, and adjusted earnings came in at $1.37 a share — both ahead of Wall Street expectations. AMD also guided June-quarter revenue to about $11.2 billion, which landed above consensus. ### Why is the data center number the real news? Because it tells you where AMD’s business has shifted. PCs still matter. Gaming still matters. But AMD is saying, pretty plainly, that servers and AI infrastructure are now the primary driver of revenue and earnings growth. That is a different company from the one investors knew a few years ago, when the story was mostly Ryzen laptops and taking CPU share from Intel. (ir.amd.com) ### What is actually selling? Two things at once. EPYC server CPUs are still climbing fast, and Instinct AI GPUs are adding another leg of growth. AMD said the data center jump came from strong EPYC demand plus the continued ramp of Instinct shipments. On the call, Lisa Su said server CPU revenue rose more than 50% year over year, with both cloud and enterprise customers up more than 50%. That matters because it means this was not one giant customer distorting the quarter. (ir.amd.com) ### Why do “agentic” and inference workloads matter here? Because they change which chips get bought. Training giant AI models made GPUs the star. But once companies start deploying AI systems that answer questions, call tools, search databases, and coordinate software agents, the workload spreads out. You still need accelerators, but you also need a lot of high-performance CPUs to feed, manage, and orchestrate those jobs. That is exactly where AMD thinks the opening is getting bigger. (ir.amd.com) Lisa Su framed it as inferencing and agentic AI increasing demand for both CPUs and accelerators, not just one or the other. ### Why is this especially good for AMD? Because AMD has credible products on both sides of that stack. Nvidia dominates AI GPUs, but Nvidia is not a major server CPU vendor. Intel is a major CPU vendor, but it has struggled to build the same momentum in AI accelerators. AMD sits in the middle with EPYC for compute and Instinct for AI. Basically, if customers want a second large-scale supplier for AI infrastructure, AMD is one of the few names that can show up with both kinds of silicon. (ir.amd.com) That does not make it the leader — but it does make it strategically harder to ignore. ### Did the guidance change the story? Yes — maybe more than the quarter itself. Q2 revenue guidance of about $11.2 billion implies 46% year-over-year growth at the midpoint. Su also said AMD expects server growth to accelerate as supply scales, and pointed to stronger-than-expected customer engagement around the upcoming MI450 series and the Helios rack platform. So this was not framed as a one-quarter spike. It was framed as visibility improving. (cnbc.com) ### What’s the catch? The catch is competition and supply. Nvidia is still the standard bearer in AI systems, and Intel is not going to give up server share quietly. AMD also has to keep getting enough advanced manufacturing capacity to meet demand. When Su says growth can accelerate as supply scales, that is a bullish statement — but it also tells you supply is still a real constraint. (ir.amd.com) ### So what changed? The quarter made one thing much clearer: AMD is no longer just participating in the AI build-out. It is becoming one of the companies that shapes how that spend gets split across CPUs, GPUs, and full server platforms. The 57% data center growth number is the proof point. The guidance is the bet that this is just getting started. (ir.amd.com)