Where it fits in an AI server rack
NVIDIA named Amkor, with SPIL, as its packaging and testing partner in Arizona (Apr 14, 2025). Amkor’s 2025 10-K describes FCBGA packages that use 2.5D silicon interposers to integrate high-bandwidth memory and graphics processors, 2.5D and high-density fan-out with and without bridges for HPC and data-center chips, and an Arizona facility under construction. Computing (data center, infrastructure, PC and storage) was 20% of 2025 net sales.
What the evidence establishes
AI-rack evidencePartner-disclosed: NVIDIA names Amkor for packaging and testing in Arizona, and Amkor’s 2025 10-K describes 2.5D interposer packages for HBM and graphics processors.
An explicit AI-server or AI-rack product, design-in, supply role, program, reference design or partnership, such as a part named for rack-scale AI systems or a product sold for AI data centers. For a rack builder, it means the company’s own AI servers or racks. This does not establish revenue, market share, volumes or future sourcing.
| Component role | 2.5D packaging and test; NVIDIA Arizona partner |
|---|---|
| Evidence | Partner-disclosed: NVIDIA names Amkor for packaging and testing in Arizona, and Amkor’s 2025 10-K describes 2.5D interposer packages for HBM and graphics processors. |
| How we know | Partner-disclosed A named partner, customer or platform owner’s document names the company. The scope is that document’s product or program, not a general supply relationship. |
| Primary source | Amkor Technology — supporting disclosure Company or official disclosure; vendor claims are attributed to their source. |
| Checked | . This is the research review date, not the source publication or quote time. |
Keep in perspective
Apple (29.8%) and Qualcomm (11.1%) were its largest direct customers in 2025; Amkor does not disclose NVIDIA revenue or which steps it performs for AI GPUs. The Arizona facility was still under construction at the 10-K date.
Diversified business exposure. This describes the breadth or role of the business; it is not an investment rating. An AI-server product, design-in or partnership does not by itself establish material AI-rack revenue, market share, or future returns.
Connections to the assembly
These links explain the technology relationship. The model is illustrative and does not represent an actual manufacturer’s bill of materials.
Interposer & 2.5D packaging
A silicon interposer, or silicon bridges in an organic layer, carries thousands of fine wires between the logic dies and the HBM. Building it and assembling the dies onto it is the advanced-packaging step.
Scope of the company links: TSMC’s CoWoS is the packaging NVIDIA names in its 10-K; TSMC produces 5.5-reticle CoWoS and plans 14-reticle packages with 20 HBM stacks for 2028. NVIDIA named Amkor and ASE’s subsidiary SPIL as its packaging and testing partners in Arizona (Apr 2025), and both describe 2.5D silicon-interposer packages for GPUs and HBM in their annual filings. Which steps each performs for a given GPU (chip-on-wafer or on-substrate) is not disclosed, and that Blackwell uses the CoWoS-L variant is reported, not confirmed in our sources. Intel’s EMIB-class packaging has no documented AI-rack program here and is not mapped.
Related research
Companies with overlapping sector exposure. Their products and evidence may differ; this is not a list of equivalent investments.
- A
Advanced Micro Devices
AMD · NASDAQInstinct GPUs, EPYC CPUs, Pensando NICs and Helios racks
AI-rack evidence - Al
Alchip Technologies
3661.TW · TWSEASIC design and production for a cloud AI accelerator
AI-rack evidence - AH
ASE Technology Holding
3711.TW · TWSE2.5D packaging and test (ASE and SPIL)
AI-rack evidence - B
Broadcom
AVGO · NASDAQCustom AI accelerators, Ethernet switch and NIC chips
AI-rack evidence - GU
Global Unichip (GUC)
3443.TW · TWSEASIC design for AI accelerators; HBM4 PHY IP
AI-rack evidence