Management admitted their largest AI customer is cutting volumes due to export controls, and they refused to guide the second half.
Thesis: Amkor is an asset-heavy business running at 50% utilization, effectively burning depreciation. The 'AI' thesis is fractured: their 'marquee' customer (likely NVIDIA/high-end GPU) is down due to export controls. They are pivoting the narrative to 'edge AI' (smartphones) and 'new switching customers' to fill the hole. This is a trade down in margin and quality. The street expects a H2 snap-back, but management's refusal to guide suggests the visibility isn't there.
Verdict: SHORT — Conviction: HIGH
Catalyst: Q3 guidance miss when the 'second half ramp' fails to offset the structural loss of the marquee AI customer.
Key Risk: If the 'new 2.5D switching customer' ramps significantly faster than expected, or if trade restrictions unexpectedly loosen.
The Tell: When asked about the 2.5D segment, CEO Giel Rutten admitted: 'continuing with 2.5D for our prime customer, although at lower volumes due to the current situation on export controls.' This confirms the bear thesis that the AI tailwind is actually a headwind for them specifically.
Friction Level: HIGH_FRICTION — Bull sees 50% utilization as a coiled spring for operating leverage. Bear sees it as evidence of structural demand loss from export controls that 'edge AI' cannot fill.
Report not found
The report data is no longer available. Please return to the archive.