According to CRN, AMD delivered record quarterly revenue of $9.2 billion for Q3, representing a 36% year-over-year increase and beating Wall Street expectations by $500 million. CEO Lisa Su reported “sharp” sales jumps across both CPU segments and Instinct data center GPUs, with the data center business hitting $4.3 billion. The company expects Q4 revenue to reach roughly $9.6 billion and revealed deepening ties with OpenAI through a multi-year agreement to deploy six gigawatts of Instinct GPUs starting in late 2025. AMD’s stock price nevertheless dropped more than 3.5% in after-hours trading, possibly due to guidance coming in below some analyst expectations.
Sponsored content — provided for informational and promotional purposes.
The OpenAI Game Changer
Here’s the thing about that OpenAI deal – it’s not just another customer win. We’re talking about a partnership that AMD believes could generate “over $100 billion in revenue over the next few years.” That’s absolutely massive when you consider AMD’s entire quarterly revenue just hit $9.2 billion. The six gigawatts of GPU deployment starting next year is just the beginning – they’re now working directly with OpenAI on future hardware, software, and system roadmaps. Basically, AMD is getting a seat at the table with one of the biggest AI players, which could seriously challenge Nvidia’s dominance in the AI training space.
Instinct GPU Momentum Building
The MI300 and MI350 series are finally gaining real traction beyond just promises. Su mentioned “multiple” MI350 deployments with large cloud providers including Oracle, plus smaller players like Crusoe and Digital Ocean. But what’s really interesting is the breadth of AI companies now using MI300 for training and inference – IBM, Cohere, Character AI, Luma AI. That’s the kind of validation AMD desperately needed. And they’re already talking about the MI400 series and Helios rack-scale platform progressing “rapidly.” The question is whether they can scale manufacturing fast enough to meet this demand.
EPYC Server Dominance Continues
While everyone focuses on GPUs, AMD’s server CPU business quietly hit an all-time high. Fifth-gen EPYC processors made up nearly half of all EPYC revenue, which is impressive adoption speed. They’ve got over 1,350 public cloud instances now – up 50% from a year ago. But here’s the real story: enterprise adoption of EPYC cloud instances grew more than three-fold year over year. Large Fortune 500 companies across telecom, finance, and automotive are making big commitments. And with Venice, their 2nm EPYC processors, coming next year with what Su calls “the strongest customer pull we have seen,” Intel should be very worried.
PC and Gaming Segment Explodes
AMD’s client business grew 73% year over year to $4 billion, which is just insane growth in the PC market. Desktop CPU sales hit an all-time high thanks to Ryzen 9000 demand, and commercial Ryzen PC sales grew over 30%. But gaming revenue was the real standout – up 181% to $1.3 billion. That’s being driven by both console chips for Microsoft and Sony, plus strong Radeon GPU demand. The embedded segment was the only weak spot, down 8%, but even there they’re tracking record design wins of over $14 billion year-to-date. So basically, AMD is firing on all cylinders except one.
What This All Means
AMD is positioning itself as the only credible alternative to Nvidia in the AI accelerator space while simultaneously eating Intel’s lunch in servers and making huge gains in PCs. The OpenAI partnership gives them instant credibility and a massive potential revenue stream. But can they execute? The stock drop after earnings suggests some investors are skeptical about whether they can hit those lofty $100 billion AI revenue targets. The challenge now is scaling manufacturing, maintaining their technology edge, and actually delivering on these enormous commitments. If they pull it off, we could be looking at a completely reshaped semiconductor landscape by 2027.
