Taiwan’s ASE expects strong demand to boost advanced chip packaging sales in 2026

ASE Technology said its leading-edge advanced packaging business is now expected to generate more than $3.5 billion in revenue in 2026, implying 10% growth and reflecting strong AI chip demand. The company also raised 2025 capex by $1.5 billion combined for buildings/infrastructure and machinery to support demand in 2026-2027. First-quarter revenue rose 17.2% year over year to T$173.66 billion and net income surged 87.3% to T$14.148 billion, though shares still closed down 1.4% ahead of results.

Analysis

This is less about one company’s execution and more about the market finally pricing the bottleneck in the AI supply chain: advanced packaging capacity is becoming the gating factor, not wafer starts. When a key subcontractor announces multi-year capex and still signals demand visibility into 2027, it strengthens the case that AI infrastructure spend is now locked in beyond the next earnings cycle, which should support the entire co-packaging ecosystem, substrate vendors, equipment suppliers, and select materials names.

The second-order winner is Nvidia, but not because of direct revenue linkage; the implication is that supply constraints are easing just enough to reduce lead-time risk and support faster shipment conversion of booked demand. That matters for gross margin stability across the ecosystem because packaging yields and throughput are a hidden swing factor in AI system delivery schedules, and any incremental capacity added in 2026 likely pulls forward revenue recognition for downstream AI server OEMs and cloud capex budgets.

The market may be underestimating the duration of the capex cycle. A near-term earnings beat is already reflected, but the meaningful upside is in 2026-2027 utilization, where fixed-cost leverage can expand faster than consensus expects if AI demand remains broadening rather than episodic. The main contrarian risk is that investors confuse capex with sustainable pricing power: if capacity additions outpace end-demand, packaging margins can compress even as revenue grows, especially after a stock’s strong YTD run.

From a timing standpoint, the better setup is on pullbacks or post-print volatility rather than chasing the tape. The clean expression is to own the AI bottleneck beneficiaries with pricing power and a visible backlog, while fading the over-owned beneficiaries that are already discounting perfect execution. Any reversal would likely come from a 2-3 quarter delay in hyperscaler capex or a signal that advanced packaging supply is normalizing faster than expected.

Key Decisions for Investors

• Long ASX on any post-earnings dip over the next 1-3 weeks; risk/reward is favorable if the market starts to price the 2026-27 capacity ramp as a multi-year earnings upgrade rather than a one-off guide-up.
• Add NVDA selectively on weakness over the next 1-2 weeks; the setup is not for immediate EPS upside but for reduced supply-chain friction and improved shipment cadence into 2H26.
• Pair trade: long ASX / short a basket of AI hardware names with stretched expectations for near-term margin expansion; this isolates the bottleneck thesis and reduces market beta if AI enthusiasm cools.
• Buy 6-12 month call spreads on ASX rather than outright stock; the setup benefits from continued multiple support, but the stock’s strong YTD move raises the risk of consolidation even if fundamentals stay intact.
• Set a downside stop if management commentary starts implying lower utilization or pricing pressure in 2026; that would be the first sign the capex cycle is turning from scarcity-driven to competitive.