👩🍳 How we use AI at Tech in Asia, thoughtfully and responsibly.
🧔♂️ A friendly human may check it before it goes live. More news here
🧔♂️ A friendly human may check it before it goes live. More news here
Taiwan’s ASE forecasts its advanced chip packaging to hit $3.2b
Taiwan’s ASE Technology Holding, the world’s largest chip packaging and testing firm, expects its advanced packaging business to reach US$3.2 billion by 2026, nearly doubling in size.
The company made the projection during a conference call after reporting a fourth-quarter revenue of T$177.9 billion (US$5.6 billion), a 9.6% increase from the previous year.
Its net income for the quarter rose by 58%.
No further details on factors driving the growth forecast were provided.
🔗 Source: Reuters
🧠 Food for thought
Implications, context, and why it matters.
The story behind ASE’s growth is a multi-billion dollar bet on AI
- ASE expects capital spending of about US$4.5 billion in 2025, around twice its 2024 outlay 1.
- Over 60% of machinery capital goes to leading-edge advanced packaging (technologies that connect multiple chips and components into a single high-performance package) 1. Another 30% goes to high-end testing (advanced methods to verify chips meet performance and reliability requirements) 1.
- The plan supports Leading-Edge Advanced Packaging (LEAP) services, with revenue expected to rise from US$250 million in 2023 to more than US$1.6 billion in 2025 1. This also supports the 2026 target referenced in the article 1.
Why chip packaging is no longer the boring part of the semiconductor industry
- Packaging now drives much of the performance in AI chips, rather than serving as a final assembly step 2.
- Spending at this scale raises the cost to compete, helping ASE pull further ahead of smaller outsourced semiconductor assembly and test (OSAT) rivals such as Amkor 3.
- More of the supply chain flows through a smaller group of high-capability partners, including ASE 4. Fabless AI chip designers (companies that design chips but outsource manufacturing) depend more on those partners 4. The same partners also enable chiplet-based designs (architectures that combine multiple smaller chips into one system) used in newer AI hardware 4.




