How Does AMD Make its Money?

AMD (Advanced Micro Devices) designs and sells processors for data centers, PCs, gaming consoles, and embedded systems. The company competes with Nvidia in AI accelerators (MI300X) and Intel in CPUs (EPYC server chips, Ryzen desktop/laptop chips). Like Nvidia, AMD is “fabless” — Taiwan Semiconductor (TSMC) manufactures the chips.

Revenue Breakdown

Segment 2024 2023 YoY Growth
Data Center $12.6B $6.5B +93.8%
Client (PCs) $7.2B $4.7B +53.2%
Gaming $1.2B $6.2B -80.6%
Embedded $3.7B $5.8B -36.2%
Total Revenue $25.8B $22.7B +13.7%

Data Center — 49% of Revenue

The star segment. Includes EPYC server CPUs (gaining share from Intel) and MI300X AI accelerators (competing with Nvidia’s H100). Data center nearly doubled revenue as AI demand surged and EPYC gained meaningful server market share (~30%+).

Client — 28% of Revenue

Ryzen CPUs for desktops and laptops. Revenue rebounded strongly as the PC market recovered from the 2022-2023 downturn and the Ryzen 7000/8000 series gained share from Intel.

Gaming — 5% of Revenue

GPU chips for gaming consoles (PlayStation 5, Xbox Series X) and Radeon discrete GPUs. Revenue dropped sharply as the console cycle matured and AMD deprioritized discrete gaming GPUs to focus on data center.

Embedded — 14% of Revenue

FPGA and adaptive SoC chips acquired through the $49B Xilinx acquisition in 2022. Serves industrial, automotive, aerospace, and communications markets. Revenue normalized after a 2022-2023 inventory overbuild.

Income Statement Overview

Metric 2024 2023
Total Revenue $25.8B $22.7B
Gross Profit $13.5B $11.1B
Operating Income $5.4B $1.3B
Net Income $4.7B $0.9B

Key Financial Metrics

  • Gross Margin: 52.3% — Expanding as Data Center (higher-margin) becomes a larger share. Nvidia’s ~77% margin shows upside potential if AMD’s AI accelerator mix shifts higher.
  • Operating Margin: 20.9% — Up dramatically from 5.7% as Data Center scale provides operating leverage.
  • Data Center Growth: +93.8% — The story stock metric. MI300X AI accelerator revenue exceeded $5B in its first full year.

What to Watch

  1. MI300X and MI400 adoption — AMD needs to prove it can be a credible #2 to Nvidia in AI accelerators. ROCm software (AMD’s CUDA competitor) must improve to attract AI developers.
  2. EPYC server share — AMD’s EPYC processors are winning share from Intel. Continuing to take server market share from 30% toward 40%+ would drive significant revenue growth.
  3. Gaming deprioritization — AMD appears to be shifting resources away from discrete gaming GPUs toward data center. This risks ceding the consumer GPU market to Nvidia.
  4. Xilinx integration — The $49B Xilinx acquisition needs to show returns. FPGA revenue has been declining, though the embedded market is cyclically recovering.
  5. Nvidia’s dominance — Nvidia controls ~80%+ of the AI accelerator market. AMD’s MI300X is competitive on specs but faces the CUDA software moat. Closing the software gap is essential.