Abstract high-speed connections representing artificial intelligence

Applications

Artificial Intelligence (AI)

Advanced semiconductor packaging is essential to the performance, energy efficiency and reliability demanded by modern artificial-intelligence systems.

Compute architecture for emerging AI

Artificial intelligence is reshaping natural-language processing, recognition, autonomous systems and industrial automation. Like HPC, it depends on high compute and storage capacity connected by very wide interconnects.

Purpose-built accelerators such as GPUs, ASICs and FPGAs can deliver faster, more energy-efficient processing than general-purpose software running only on CPUs.

Artificial Intelligence (AI) diagram

Heterogeneous integration

Packaging platforms for powerful, efficient ICs

ASE combines advanced substrates, high-density interconnects, memory integration and optical capabilities to support the processors at the center of AI infrastructure.

HFCBGA

High-performance flip-chip BGA for large, complex processors.

2.5D / 3D IC

Dense die-to-die integration for logic and high-bandwidth memory.

FOCoS-CL and FOCoS-B

Fan-out chip-on-substrate options for advanced heterogeneous systems.

Silicon photonics

Optical integration that supports high-bandwidth, energy-efficient data movement.

AI inside smart manufacturing

ASE also applies predictive analytics to its own smart factories, using production data to improve equipment performance, quality, reliability and cost efficiency.