New quarterly report expected Apr 27, 2026. Report will refresh after results.
A
Amkor Technology, Inc.
AMKR · Technology
Q4 FY2025
Earning ended Dec 31, 2025Report generated Mar 15, 2026
Amkor Packages the AI Boom While Foundries Pocket the Upside
The market is treating Amkor as an AI infrastructure play, but the foundries building advanced packaging in-house means the richest slice of the secular tailwind may never show up in Amkor's margins — the company is running harder on its capital treadmill precisely as the most valuable destination moves away from it.
Flowly Score
Apr 16 Close
-60%$62.90
Fair Value
$25.00
35
Below Averageat $62.90
Business Strength5/10
Amkor is essential infrastructure that captures very little of the value it enables — qualification lock-in and scale keep competitors at bay, but pricing power lives with Apple and Qualcomm, not here. ROIC compressing toward the cost of capital over four years is the verdict on the moat, not just a cycle.
Financial Resilience5/10
Cash earnings are genuine — OCF consistently clears reported net income by a wide margin, a clean fingerprint of a real asset-heavy business rather than accounting creativity. The problem is that nearly every dollar of operating cash flow feeds a capital expenditure machine that never stops, leaving thin and lumpy free cash flow as the actual shareholder residual.
Growth & Trajectory6/10
The advanced packaging secular thesis is real: AI accelerators, chiplet stacking, and heterogeneous integration genuinely require more sophisticated packaging, and Amkor is positioned in that stream. The ceiling risk is that the most technically demanding and margin-rich work — CoWoS for AI GPUs, HBM integration — is being quietly absorbed by the foundries themselves before Amkor can make it a durable revenue line.
Valuation2/10at $42.99
The market is paying a premium multiple on a business that has spent four years compressing its returns on capital — that is a dangerous combination requiring a near-perfect execution of the advanced packaging ramp to justify. The EV/FCF multiple is especially punishing given that normalized FCF assumes capex permanently decelerates, which in this business is more aspiration than certainty.
Risk3/10
Three serious, specific risks converge simultaneously: foundry vertical integration claiming the highest-value packaging work, single-customer concentration that can reshape the income statement in one product cycle, and a controlled-company governance structure where minority shareholders are structurally junior. The $7 billion Arizona bet amplifies all three — it's a massive, illiquid capital commitment to a customer relationship that Amkor does not control.
The investment case rests on a real secular thesis — advanced packaging demand from AI chiplets and heterogeneous integration is genuine, the Arizona facility captures a defensible supply-chain-localization premium, and government subsidies meaningfully reduce the capital risk of that bet. But the quality of growth matters as much as the existence of growth, and here the ROIC trajectory tells a damning story: a business pouring capital into expansion while returns compress is not compounding value, it's running a capacity arms race where the spoils flow to end customers. The current multiple prices in a step-change improvement in earning power that hasn't materialized in four years of data. Where the business is heading depends almost entirely on whether high-density fan-out and heterogeneous integration packaging becomes a sticky, differentiated service or another competitive commodity. The optimistic path is real: AI chip complexity is structurally accelerating, automotive electrification creates long-tailed packaging demand with AECQ100 switching friction, and the Vietnam and Arizona facilities position Amkor ahead of a genuine customer preference for geographically diversified supply chains. The pessimistic path is equally real: ASE is larger, TSMC is smarter, and Samsung has its own balance sheet. The single biggest risk is the foundry vertical integration squeeze. TSMC expanding CoWoS for Nvidia and AMD, Samsung scaling HBM packaging, Intel developing EMIB — these are not vague threats. They represent the leading semiconductor manufacturers systematically claiming the most technically sophisticated and margin-accretive packaging work for themselves, leaving Amkor to compete for second-tier complexity alongside aggressive Asian rivals on price. If that dynamic persists, the Arizona investment is a multi-billion-dollar bet on winning customers at margins that structurally cannot compensate for the capital deployed.