Foxconn FII

Foxconn Industrial Internet Co., Ltd.

🇹🇼
SERVER ODMs🇹🇼 TW601138 · SSE
fii-foxconn.com

Market Share

~15% AI server ODM globally

Key Product

AI servers (NVIDIA HGX-based), NVIDIA MGX platform servers

Trace supply chain →
Full briefing▼ Expand

Foxconn Industrial Internet Co., Ltd. (SSE: 601138, also known as FII) is a subsidiary of Hon Hai Precision Industry Co. (Foxconn), the Taiwanese contract manufacturer best known for assembling Apple iPhones. FII focuses specifically on industrial internet, cloud computing, and AI server infrastructure, separating the high-growth data center business from Hon Hai's consumer electronics assembly operations. FII competes directly with Quanta Computer and Wiwynn for hyperscaler AI server contracts. Its parent Hon Hai's scale — 800,000+ employees, 30+ factories globally — gives FII significant cost advantages in PCB fabrication, mechanical manufacturing, and supply chain logistics. FII operates AI server manufacturing facilities in Taipei (Taiwan), Shenzhen (China), Houston (US), and plans additional facilities in India and Europe. For hyperscale AI infrastructure, FII's key product is the NVIDIA HGX-based GPU server (supporting H100, H200, and Blackwell B200 configurations). It also manufactures servers using the NVIDIA MGX modular design reference architecture. FII has been NVIDIA's fastest-growing ODM partner in terms of AI server deployments. FII's China operations serve a distinct market: Chinese hyperscalers and AI companies including ByteDance, Alibaba, Baidu, and Tencent that must source AI servers from domestic-supply-chain-compliant partners given US export restrictions. FII's dual presence in both US-allied and Chinese markets creates a complex geopolitical position — it assembles US-restricted H100 GPUs for Western hyperscalers while building servers for Chinese AI companies using restricted-but-downgraded H20 chips or domestic Huawei Ascend alternatives. Its parent Hon Hai CEO Young Liu has positioned the company as a key infrastructure provider for the 'AI everywhere' transition, recently securing a major investment in Nvidia partner-ecosystem certification and investing in its own AI server liquid-cooling capabilities.

Critical path — raw silicon to deployment

CHIP DESIGNERS

NVIDIA

H100, H200, Blackwell B200 GPUs

CHIP DESIGNERS

AMD

MI300X AI accelerator, EPYC CPUs

CHIP DESIGNERS

Intel

Gaudi 3 AI accelerator, Xeon Scalable CPUs, Intel Arc GPU

SERVER ODMs

Foxconn FII

AI servers (NVIDIA HGX-based), NVIDIA MGX platform servers

CLOUD PROVIDERS

Google Cloud

TPU v5e AI accelerator, A3 GPU clusters (H100-based)

AI CONSUMERS

Meta

MTIA v2 AI accelerator, H100/H200 GPU clusters (Llama models)