AI Chips Supply Chain Map
From design software to fabrication to packaging — trace the global dependencies that determine who can build the chips powering artificial intelligence.
90%
Advanced chips from TSMC
1
EUV supplier (ASML)
5
Countries in critical path
$150B+
AI chip market by 2027
Supply Chain Stages
Design & IP
Chip architecture, instruction sets, and design software
Dominant Players
- NVIDIA (US)
- AMD (US)
- ARM (UK/Japan)
- Cadence (US)
- Synopsys (US)
China's Position
Emerging: Huawei HiSilicon, Cambricon, Biren
Market Share
Manufacturing Equipment
Lithography, etching, deposition, and inspection tools
Dominant Players
- ASML (Netherlands)
- Applied Materials (US)
- Lam Research (US)
- Tokyo Electron (Japan)
China's Position
Limited: SMEE (immersion DUV only)
Market Share
Fabrication
Wafer manufacturing at advanced process nodes
Dominant Players
- TSMC (Taiwan)
- Samsung (Korea)
- Intel (US)
China's Position
SMIC limited to 7nm, not cutting edge
Market Share
Packaging & Testing
Advanced packaging, chiplet integration, final testing
Dominant Players
- ASE (Taiwan)
- Amkor (US)
- JCET (China)
China's Position
Stronger: JCET, Tianshui Huatian
Market Share
Critical Chokepoints
EUV Lithography
CriticalASML is the sole supplier of extreme ultraviolet lithography machines needed for chips below 7nm. Each machine costs $150M+ and takes 18 months to build.
Advanced Node Fabrication
CriticalOnly TSMC and Samsung can manufacture chips at 5nm and below at scale. TSMC alone makes 90% of the world's most advanced chips.
HBM Memory
HighHigh-bandwidth memory essential for AI accelerators is produced only by SK Hynix, Samsung, and Micron. Supply is constrained.
Chip Design Software
HighEDA tools from Cadence and Synopsys are essential for designing complex chips. No Chinese alternatives exist for advanced nodes.