Nanya Technology's 2027 Capex to Expand Beyond NT$200 Billion, Roughly Quadrupling This Year's Level
Alina Collins
Taiwan memory chipmaker Nanya Technology plans to boost 2027 capital spending to over NT$200 billion (~US$6.2 billion), roughly four times this year's NT$50 billion-plus budget, as an AI-driven memory shortage forces the entire industry into a simultaneous buildout — the key question is whether the capacity arrives before the gap closes.
Where is the money going?
2027 capex target: over NT$200 billion (~US$6.2 billion), nearly four times this year's planned NT$50 billion-plus.
The bulk goes toward a new fab in New Taipei City, with a total project cost of roughly NT$480 billion.
Phase-one capacity target: 30,000 wafers per month by 2028, scaling to 45,000 wafers per month thereafter.
This means → Nanya is not riding a short-term price spike — it is committing to a multi-year factory ramp.
How extreme are the latest earnings?
Q2 revenue hit NT$82.55 billion, up 684% year-on-year; net profit reached NT$50.19 billion, up 1,324%.
Gross margin swung from negative 20.6% a year ago to 79.5%.
In plain terms = a year ago every chip sold lost money; now nearly eighty cents of every dollar is profit — that is the raw force of a pricing upcycle.
Why is the shortage so severe?
President Lee Pei-ing said AI is driving a structural shift that supports a stronger long-term outlook for memory.
Cloud-server demand is "exceptionally strong," creating a severe supply gap that has now spread to nearly every market segment.
This means → the crunch is no longer confined to high-end AI training memory — consumer electronics, automotive, and other segments are feeling the squeeze too.
Is everyone else building at the same pace?
Micron raised its US investment commitment from US$200 billion to US$250 billion through 2035.
Samsung and SK Hynix together announced US$520 billion in capacity expansion in South Korea.
CXMT is pushing ahead with expansion in Shanghai and Hefei.
This reflects an industry-wide consensus on AI-driven demand — it is not one company making a bet; every major player is racing to build.
What is the biggest risk?
The new fab does not come online until 2028. Nanya ranks fifth globally in DRAM (dynamic random-access memory — the most common memory chip in phones and PCs), with customers including Nvidia, Qualcomm, and Google.
The core test: can Nanya ramp capacity and lock in customers before the supply gap narrows? That is the make-or-break for this massive capex commitment.
Put simply = everyone makes money during a shortage, but if the market is well-supplied by the time the fab is ready, the investment may not pay off.
Content is for reference only, not financial advice.