AMD overtakes Intel in Q1
- AMD’s Q1 2026 data-center revenue reached $5.8 billion, edging past Intel’s $5.1 billion and marking the first quarter AMD led that category. - The gap was narrow but real — AMD data-center sales rose 57% year over year, while Intel’s Data Center and AI unit grew 22%. - That matters because AI spending is shifting toward mixed systems where CPUs regained leverage beside accelerators, helping AMD’s EPYC-plus-Instinct pitch.
Server chips are the domain here — and the stakes are simple. Whoever wins the data center now gets a front-row seat to the AI buildout. For years, Intel owned that seat and AMD was the challenger. In the first quarter of 2026, that flipped, at least on revenue: AMD’s data-center segment hit $5.8 billion, while Intel’s Data Center and AI group came in at $5.1 billion. ### Did AMD really pass Intel? Yes — on the cleanest quarter-to-quarter comparison we have right now. AMD reported Q1 data-center revenue of $5.8 billion on May 5. Intel had reported $5.1 billion for its Data Center and AI unit on April 23. That puts AMD ahead by about $700 million for the quarter. ### Why is that a big deal? (ir.amd.com) Because Intel used to treat server CPUs as home turf. AMD has been taking share with EPYC for years, but this is the first time the revenue line itself appears to have crossed. That turns a long-running competitive story into a visible financial milestone — not just “AMD is gaining,” but “AMD led this quarter.” ### What pushed AMD over the line? Two things at once. EPYC server CPUs kept gaining traction, and Instinct GPU shipments kept ramping. AMD said both contributed to the 57% year-over-year jump in data-center revenue. Lisa Su also said inferencing and agentic AI are driving demand for both high-performance CPUs and accelerators — basically, buyers want the whole system, not just the flashy GPU. (ir.amd.com) ### Why are CPUs suddenly back in the spotlight? Because AI infrastructure is maturing. Training giant models made GPUs the star. Running those models in production — especially with retrieval, orchestration, and multi-step agent workflows — leans harder on CPUs, memory, networking, and packaging too. Intel has been making this exact argument around Xeon, and AMD is benefiting from the same shift with EPYC. The interesting part is that AMD also sells the accelerator, so it can pitch a more balanced stack. (ir.amd.com) ### Does this mean Intel is losing everywhere? No — and that’s the catch. Intel’s total company revenue in Q1 was still larger at $13.6 billion, versus AMD’s $10.3 billion. Intel also grew its data-center business 22% year over year and said demand for Xeon server CPUs still exceeds supply. So this is not an Intel collapse story. It’s a segment-crossover story inside a market that is expanding for both companies. (download.intel.com) ### Why does the mix matter more than the headline? Because “data center” is no longer just a CPU bucket. It now includes the host processors, accelerators, memory-heavy system designs, and the plumbing that keeps AI clusters fed. AMD’s edge this quarter says customers are rewarding vendors that can cover more of that stack. Intel still has huge manufacturing and packaging assets, but AMD’s current mix lines up very neatly with what hyperscalers are buying. (intc.com) ### What should we watch next? Two numbers. First, whether AMD can hold the lead in Q2 after guiding to about $11.2 billion in total revenue and talking up stronger server growth. Second, whether Intel can turn strong Xeon demand into faster data-center revenue growth as supply improves. If AMD repeats this, the crossover stops looking symbolic and starts looking structural. (newsroom.intel.com) ### Bottom line? AMD did not just post a strong quarter. It crossed a line that used to feel hard to imagine in servers. But the deeper story is about AI infrastructure growing up — away from GPU-only thinking and toward mixed compute systems where CPUs matter again. (ir.amd.com) (cnbc.com)
