--- title: "AMD (Minutes): AI GPU will achieve double-digit growth for the full year." type: "Topics" locale: "en" url: "https://longbridge.com/en/dolphin/post/29400172.md" description: "AMD (AMD.O) released its Q1 2025 earnings report (as of March 2025) after the U.S. market close on May 7, 2025, Beijing time. Key highlights are as follows: Below are the Minutes of AMD's Q1 2025 earnings call. For a Quick Interpretation of the report, please refer to "AMD: 'Outplaying' Intel and Ready to 'Turn Up the Heat' on NVIDIA?" 1. $AMD(AMD.US) Key Financial Metrics Review 1. Overall Performance: Record revenue of $7.4 billion, exceeding guidance 上限， up 36% YoY but down 3% QoQ; gross margin 54%..." datetime: "2025-05-07T08:19:43.000Z" locales: - [en](https://longbridge.com/en/dolphin/post/29400172.md) - [zh-CN](https://longbridge.com/zh-CN/dolphin/post/29400172.md) - [zh-HK](https://longbridge.com/zh-HK/dolphin/post/29400172.md) author: "[Dolphin Research](https://longbridge.com/en/dolphin.md)" generator: "portal-rs" --- # AMD (Minutes): AI GPU will achieve double-digit growth for the full year. **AMD (AMD.O) released its Q1 2025 earnings report (as of March 2025) after U.S. market hours on May 7, 2025 Beijing time. Key highlights:** **Below are the minutes of AMD's Q1 2025 earnings call. For earnings analysis, please refer to** [**AMD: "Outplaying" Intel and Now Turning Up the Heat on Nvidia?**](https://longportapp.cn/zh-CN/topics/29392001) **I.** $AMD(AMD.US) **Key Earnings Highlights** 1\. Overall performance: Record revenue of $7.4B, exceeding guidance 上限， up 36% YoY but down 3% QoQ; gross margin at 54%, up 1.4pct YoY; operating expenses of $2.2B, up 28% YoY; operating income of $1.8B with 24% margin; diluted EPS of $0.96, up 55% YoY. 2\. Segment breakdown: **-** Data Center: Revenue of $3.7B, up 57% YoY but down 5% QoQ. Growth driven by server CPU market share gains and Instinct GPU expansion; operating income of $932M with 25% margin. \- Client & Gaming: Revenue of $2.9B, up 28% YoY and 2% QoQ, benefiting from Ryzen processor demand though partially offset by semi-custom declines; client revenue of $2.3B surged 68% YoY with over half 来自高端产品涨价. Operating income of $496M with 17% margin. \- Embedded: Revenue of $823M, down 3% YoY and 11% QoQ. Operating income of $328M with industry-leading 41% margin. 3\. Financial position: Free cash flow of $727M; $749M returned via buybacks with $4B remaining authorization; cash reserves at $7.3B; debt reached $1.5B with $950M commercial paper issued to fund ZT Systems acquisition completed March 31. 4\. Outlook: a. Q2: Revenue guidance of $7.4B ±$300M (MI308 export restrictions 预计减少$700M; ex-impact still up 27% YoY); non-GAAP gross margin ~43% (含$800M inventory charge, 54% ex-charge); opex ~$2.3B including ZT design 团队; net interest expense $5M; non-GAAP tax rate 13%; ~1.64B diluted shares. b. Full-year: Export license requirements expected to impact revenue by ~$1.5B. c. **Sequentially, Client & Gaming revenue 预计两位数% 增长， Embedded flat, while Data Center declines due to MI308 exclusion**. **II. Earnings Call Details** **2.1 Management Commentary** 1\. Q1 performance: Strong start despite tariffs/regulatory headwinds. Revenue up 36% YoY to $7.4B with double-digit growth across DC/Client/Gaming. 2\. Data Center: a. Revenue up 57% YoY to $3.7B. 5th-gen EPYC 量产 +4th-gen demand drove server CPU share gains. Hyperscalers launched 30+ new instances including first 5th-gen EPYC solutions; AWS debuted new FPGA accelerators. b. Enterprise EPYC adoption accelerated with Global 2000 use cases doubling. Custom solutions with Siemens (Azure SDV), Oracle (Exadata X11N 性能提升 25% with EPYC 优化). 3\. EPYC Enterprise: a. **7th straight quarter of double-digit EPYC revenue growth**. 450+ EPYC platforms now available, adopted by 多数 Global 2000 leaders. 150+ Turin platforms coming from Dell etc. b. 5th-gen EPYC (TSMC Arizona) enters production with H2 shipments. Next-gen Venice (2nm) lab 验证中 for 2026. 4\. DC AI: a. **MI325X shipments drove double-digit AI revenue growth**. Hyperscalers expanded Instinct for generative AI workloads; added Tier-1 cloud/enterprise clients including major LLM developer using Instinct for 推理流量。 b. **ROCm update cadence accelerated to bi-weekly 训练/推理 containers**. 2M+ Hugging Face models now AMD-compatible. **c. ZT Systems acquisition completed** - enables rack-scale AI solutions. **MI350 samples shipped with mass production mid-year** (1.5x memory vs MI300X, 35x throughput gains). Oracle deploying multi-billion-dollar MI350 clusters. d. **MI400 (2026) enables seamless scaling** from server-to-DC. Early feedback positive; details at June 12 event. 5\. Client & Gaming: a. Revenue up 28% YoY to $2.9B (Client +68%). **Record client CPU ASPs** with desktop channel sales >50% YoY. AI PC processors 销量环比增长超 50%. b. Commercial strength: Ryzen Pro PC sales up >30% YoY; HP shipments +80%. c. Gaming revenue down 30% YoY as **Radeon growth offset by semi-custom declines**. 6\. Embedded: a. Revenue $823M (-3% YoY). **H2 recovery expected** in test/measurement, comms, aerospace. Spartan UltraScale+ FPGA and Versal AI edge chips shipped. 7\. Full-year outlook: Confident in **double-digit revenue growth** driven by EPYC/Ryzen share gains and MI350 ramp. Semi-custom recovery expected. **2.2 Q&A** **Q: Client revenue +67% YoY - demand pull-forward?** A: Strength 来自 product mix/ASP gains. Desktop outperformed seasonality. **Q2 Client remains strong**; Gaming to rebound as console inventory clears. **Q: MI308 performance ex-China?** A: MI325/MI300 drove Q1 DC GPU results. **MI350 on track for mid-year launch** with strong customer interest. **Q: DC GPU growth math?** A: $1.5B export impact mostly in Q2-Q3. **H2 DC GPU growth expected**. **Q: Inventory build?** A: Primarily for H2 DC GPU demand given long lead times. **Q: MI400 rack challenges?** A: ZT acquisition addresses rack-scale design. **MI350 easier transition; MI400 focus**. **Q: EPYC enterprise share gains?** A: Product breadth + direct customer engagement drove success. **Q: Gross margin trajectory?** A: **H2 margin improvement expected** as DC mix improves. **Q: AI inference costs?** A: Distributed inference plays to AMD strengths. **MI300/325/350 excel at 推理**. **Q: Custom ASIC competition?** A: **GPUs address larger TAM** given model volatility. Differentiation through TCO. \ **Disclosures:** [**Dolphin Research Disclaimer**](https://support.longbridge.global/topics/misc/dolphin-disclaimer) ### Related Stocks - [AMD.US](https://longbridge.com/en/quote/AMD.US.md) - [NVDA.US](https://longbridge.com/en/quote/NVDA.US.md) --- > **Disclaimer: This article is for reference only and does not constitute any investment advice.**