Market Share
~30% hyperscaler AI server ODM
Key Product
Custom AI servers for Google, Meta, Microsoft, Amazon
Full briefing▼ Expand
Quanta Computer Inc. (TWSE: 2382) was founded in 1988 by Barry Lam and is headquartered in Taoyuan, Taiwan. Originally a laptop ODM for Apple, Dell, and HP, Quanta pivoted aggressively into server manufacturing in the 2010s and has become the dominant ODM for hyperscaler AI infrastructure. Its cloud/server division now accounts for the majority of revenue. As an ODM (Original Design Manufacturer), Quanta does not sell branded products. Instead, it designs and manufactures servers to the specifications of hyperscaler customers — Google, Meta, Microsoft, Amazon — who then deploy them under their own infrastructure. This model gives hyperscalers flexibility to customize hardware without building their own manufacturing capacity, while giving Quanta a high-volume, long-duration revenue stream. For AI infrastructure specifically, Quanta's role is to integrate NVIDIA H100/H200/B200 GPUs (or AMD MI300X accelerators) with HBM memory (from SK Hynix, Samsung, Micron), high-speed networking ASICs (from Broadcom), NVMe SSDs, and power/cooling systems into complete server nodes. These nodes are then cabled into rack-scale systems shipped directly to hyperscaler data centers. The AI infrastructure boom of 2023–2025 transformed Quanta's order book. Its cloud revenue surged as hyperscalers raced to build out GPU clusters for LLM training. Quanta competes primarily with Wiwynn (a Wistron subsidiary, key to Microsoft/Meta) and Foxconn Industrial Internet for hyperscaler AI server contracts. Quanta's Taiwan base makes it subject to cross-strait risk: a Taiwan Strait crisis that disrupted Quanta's operations would cascade immediately into AI infrastructure shortfalls at every major US and global cloud provider.
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
Quanta Computer
Custom AI servers for Google, Meta, Microsoft, Amazon
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)