Financial Performance - AMD reported record Q3 2025 revenue of $9.2 billion, a 36% year-over-year increase[17] - The GAAP gross margin was 52%, up 2 percentage points year-over-year[21,29] - Non-GAAP gross margin was 54%, flat year-over-year[21,30] - GAAP diluted earnings per share (EPS) increased 60% year-over-year to $0.75[27,29] - Non-GAAP diluted EPS increased 30% year-over-year to $1.20[27,30] Segment Results - Data Center revenue reached $4.3 billion, a 22% year-over-year increase[31,34] - Client & Gaming segment revenue was $4.0 billion, up 73% year-over-year[31,36] - Embedded segment revenue was $857 million, down 8% year-over-year[31,37] Q4 2025 Outlook (Non-GAAP) - Revenue is expected to be approximately $9.6 billion, plus or minus $300 million[38] - Gross margin is projected to be around 54.5%[38]

Core Viewpoint - Supermicro is launching new GPU solutions featuring AMD Instinct MI350 series GPUs, designed for high performance, scalability, and efficiency in AI-driven data centers [1][2][3] Company Overview - Supermicro is a global leader in Application-Optimized Total IT Solutions, focusing on Enterprise, Cloud, AI, and 5G Telco/Edge IT Infrastructure [8] - The company emphasizes innovation and efficiency in its product offerings, which include servers, AI, storage, IoT, and networking solutions [8] Product Features - The new Supermicro H14 generation GPU solutions support both liquid-cooled and air-cooled configurations, providing flexibility for various deployment environments [4][7] - These systems utilize dual AMD EPYC 9005 CPUs and AMD Instinct MI350 series GPUs, enhancing performance for AI, HPC, Cloud, and Enterprise workloads [3][5] - The AMD Instinct MI350 series GPUs deliver up to 40% more tokens-per-dollar compared to competitors, optimizing cost efficiency for customers [4] Performance Metrics - The new GPU servers offer 288GB HBM3e memory per GPU, which is a 1.5x increase in memory capacity compared to previous generations, and 8TB/s bandwidth [5] - The systems are designed to maximize computational throughput and energy efficiency, enabling faster processing for AI workloads [5][6] Market Position - Supermicro's integration of AMD MI350 series GPUs into its offerings demonstrates a commitment to providing advanced solutions for AI training and inference [6][7] - The company aims to support the growing demand for scalable and efficient infrastructure in AI applications across cloud service providers and enterprises [7]