AT&S (ATS.VI) — 2026Q3 FY2026 Earnings Call Analysis

Margins Flat at the Peak of the Buildout

Management cut capex by EUR 50M while claiming demand is at an all-time high.

Thesis: AT&S is a high-beta capacity play with zero pricing power. They are scaling expensive manufacturing for a product category hyperscalers want to eliminate. 23% margins in a peak AI cycle prove they are a commoditized middleman. TSMC owns the gate. AT&S owns the depreciation.

Verdict: SHORT — Conviction: HIGH

Catalyst: The Kulim Plant 2 anchor tenant announcement. Lack of a firm commitment for the shell plant within 12 months signals a stranded asset.

Key Risk: Intel EMIB-T design wins could force a capacity squeeze. If Intel regains foundry share, AT&S is the primary beneficiary of their non-CoWoS architecture.

The Tell: The IR Director's response regarding Plant 2 anchor tenants. He stated they would 'announce ad hoc' if the board decided something. This confirms the shell plant has zero firm commitments despite peers running at 100% utilization. It is a field of dreams strategy.

Detected Patterns

Friction Level: HIGH_FRICTION — The fundamental disagreement centers on architectural obsolescence. One side views ABF substrates as a mandatory bottleneck for AI. The other identifies TSMC's CoWoP as an existential threat that eliminates the substrate layer entirely. This is a binary bet on the physical stack of the next GPU generation.

Report not found

The report data is no longer available. Please return to the archive.