⬆ UPSTREAM — SUPPLIERS10
GlobalFoundries
US-headquartered mature-node foundry (14nm/22FDX) specializing in RF, defense, and automotive chips; key Western alternative to TSMC for non-advanced applications
UMC
Taiwan-based mature-node foundry (28nm–40nm) with ~6% global revenue share; key supplier for analog, mixed-signal, and display driver chips
Arm Holdings
Dominant CPU/NPU IP licensor; virtually all AI chips in mobile, edge, and cloud use Arm instruction set
TSMC
World's largest contract chip manufacturer
Samsung Foundry
Second-largest advanced foundry, competing at 3nm
Cadence
Leading EDA software provider; IC design tools (Virtuoso, Genus, Innovus) used by every advanced chip designer; US export-controlled to Chinese entities since 2022
Synopsys
Largest EDA software provider; IC design and verification tools (Design Compiler, PrimeTime, VCS) essential for all advanced chip tapeouts; US export-controlled to Chinese entities
SK Hynix
Leading HBM3/HBM3E supplier for AI accelerators
Samsung Memory
World's largest DRAM and NAND supplier, ramping HBM3
Micron
Only US-based DRAM and HBM maker
Marvell
Custom AI ASIC and networking chip designer for hyperscalers; key supplier of Arm-based data center networking and storage controllers to AWS, Azure, and Google
⬇ DOWNSTREAM — CUSTOMERS5
AWS
World's largest cloud provider and AI infrastructure buyer
Microsoft Azure
Second-largest cloud; OpenAI exclusive compute partner; deploying Maia 100 custom ASIC alongside NVIDIA clusters
Google Cloud
Hyperscaler deploying custom TPU AI accelerators (v5e via Broadcom/TSMC) alongside NVIDIA A3 GPU clusters
Meta
One of the world's largest GPU buyers; designing MTIA v2 custom AI accelerator via TSMC; largest open-source AI model deployer
Arista Networks
#1 cloud-scale Ethernet switch vendor; powers AI training cluster networking at AWS, Google, Meta, and Microsoft using Broadcom Tomahawk ASICs
UPSTREAM — SUPPLIERS
GlobalFoundries
US-headquartered mature-node foundry (14nm/22FDX) specializing in RF, defense, and automotive chips; key Western alternative to TSMC for non-advanced applications
UMC
Taiwan-based mature-node foundry (28nm–40nm) with ~6% global revenue share; key supplier for analog, mixed-signal, and display driver chips
Arm Holdings
Dominant CPU/NPU IP licensor; virtually all AI chips in mobile, edge, and cloud use Arm instruction set
TSMC
World's largest contract chip manufacturer
Samsung Foundry
Second-largest advanced foundry, competing at 3nm
Cadence
Leading EDA software provider; IC design tools (Virtuoso, Genus, Innovus) used by every advanced chip designer; US export-controlled to Chinese entities since 2022
Synopsys
Largest EDA software provider; IC design and verification tools (Design Compiler, PrimeTime, VCS) essential for all advanced chip tapeouts; US export-controlled to Chinese entities
SK Hynix
Leading HBM3/HBM3E supplier for AI accelerators
Samsung Memory
World's largest DRAM and NAND supplier, ramping HBM3
Micron
Only US-based DRAM and HBM maker
↓ all supply to Marvell ↓
Marvell
Custom AI ASIC and networking chip designer for hyperscalers; key supplier of Arm-based data center networking and storage controllers to AWS, Azure, and Google
↓ Marvell supplies to ↓
AWS
World's largest cloud provider and AI infrastructure buyer
Microsoft Azure
Second-largest cloud; OpenAI exclusive compute partner; deploying Maia 100 custom ASIC alongside NVIDIA clusters
Google Cloud
Hyperscaler deploying custom TPU AI accelerators (v5e via Broadcom/TSMC) alongside NVIDIA A3 GPU clusters
Meta
One of the world's largest GPU buyers; designing MTIA v2 custom AI accelerator via TSMC; largest open-source AI model deployer
Arista Networks
#1 cloud-scale Ethernet switch vendor; powers AI training cluster networking at AWS, Google, Meta, and Microsoft using Broadcom Tomahawk ASICs
DOWNSTREAM — CUSTOMERS
QWhat is Marvell's role in the AI chip supply chain?
Custom AI ASIC and networking chip designer for hyperscalers; key supplier of Arm-based data center networking and storage controllers to AWS, Azure, and Google
View full company profile →