ASX · ASE Technology Holding Co., Ltd.

Current Thesis

Advanced packaging is the physical chokepoint of the AI-chip buildout, and ASE the world's largest OSAT is absorbing TSMC's CoWoS overflow while its own LEAP advanced-packaging line ramps toward a guided doubling (>$3.5B for 2026). The fundamental leg is still accelerating. The tape is not: a fresh 52-week high of $41.10 was met with an 11.38% single-session reversal to $34.03 on 2026-06-05 (prior close $38.40), the 14-day RSI has unwound from overbought to ~48, and insiders sold $349.3M over three months with zero offsetting buys. A parabola is cracking after the "Nvidia Halo Trade" narrative went fully mainstream. The next hard read is the May monthly revenue print due ~2026-06-08, which tests whether April's +0.4% sequential was a blip or the start of MoM deceleration. A fresh long at $34 buys a broken momentum structure ahead of that print no edge here, watch the 50-day reclaim/loss instead.

Bull Case

• Q1 2026 (printed 2026-04-29): revenue NT$173.66B / US$5.508B, +17.2% YoY; net income +87% YoY; EPS NT$3.24 (+85% YoY); gross margin 20.1%. Mix packaging 51% / test 12% / EMS 36% advanced packaging is carrying the growth.
• LEAP advanced-packaging guidance raised on the Q1 call to >$3.5B for 2026 (~10% above prior), roughly doubling the high-margin AP book (EE Times, Seeking Alpha 2026-04-29).
• Full-process packaging begins H2 2026 with AMD Venice CPUs as the first product, projected at $300–400M of 2026 revenue a new line item, not a re-label.
• April 2026 monthly revenue (2026-05-08): US$1.957B, +23.1% YoY strong on a YoY basis even as the sequential figure flattened to +0.4%.
• SPIL is embedded in Nvidia's Arizona AI-infrastructure project as a CoWoS supplier; COO Tien Wu has stated ASE will "proactively plan for capacity expansions in line with the Nvidia project" (circuit-chip / Astute, 2026).
• Automated 310mm panel-level packaging line announced 2026-05-26, production targeted H1 2027 next-gen substrate economics versus wafer-level for large AI packages.
• ~$7B 2026 capex funds VIPack / FOCoS silicon-bridge density; demand is real NVDA books ~595k CoWoS wafers for 2026 (~60% of global), with ~80k wafers of overflow split across ASE/SPIL, Amkor and Powertech.
• Q2 2026 guide (2026-04-29): consolidated revenue +7–9% QoQ, gross margin +20–100bps, operating margin +50–120bps sequential acceleration into 2H.

Bear Case

• Momentum is broken. The $41.10 high gave way to $34.03 via an 11.38% single-session drop on 2026-06-05; the stock is ~17% off its high and the violent pullback the prior weeks lacked has now arrived.
• Insiders sold $349.3M over the trailing 3 months with no buying the people closest to the order book are distributing into strength.
• Valuation is rich for a cyclical OSAT: trailing P/E 54.3, forward P/E 30.4 against a historical 12–18x band. The multiple already prices the AP doubling.
• The narrative is public. Benzinga's "Nvidia Halo Trade" framing plus backward-looking "$1,000 invested X years ago" listicles (2026-05-15, 2026-05-28) are the kind of coverage that clusters near sentiment peaks.
• April's +0.4% sequential revenue is the first crack in the YoY rip; the ~2026-06-08 May print is the confirmation or refutation.
• AI-capex beta cuts both ways: a CoWoS double-order/digestion scare or a Taiwan geopolitical headline hits the whole OSAT cluster at once. Amkor's $2.5–3.0B 2026 capex (~3x prior) signals a packaging capacity arms race that can compress sector returns.

Setup & Price Structure

• Latest $34.03 (2026-06-05 close), −11.38% on the session from a $38.40 prior close. The reversal off the $41.10 high (set days earlier) is ~17%.
• 52-week range $9.30–$41.10; market cap ~$80.2B; trailing P/E 54.3, forward 30.4; beta ~1.40; dividend yield ~0.8%.
• 50-day MA ~$30.80, 200-day MA ~$19.80. Price now ~10% above the 50-day (down from ~43% in late May) and ~72% above the 200-day (down from ~104%) still extended, but a lot of air has come out.
• 14-day RSI ~48 fully unwound from overbought, neutral but not yet washed out. No higher low has formed; this is a reversal in progress, not a base.
• Key levels: 50-day MA ~$30.80 is the line in the sand for the uptrend; $41.10 is the re-arm level on any reclaim; $19.80 (200-day) is the deep-support backstop.

Catalyst Calendar (next 30 days)

• ~2026-06-08 (est.): May 2026 monthly consolidated revenue print. The live tell sequential MoM after April's +0.4% flatlined. A second flat/negative MoM reading confirms deceleration; a reacceleration to mid-single-digit MoM partially repairs the tape thesis.
• ~2026-07-08 (est.): June 2026 monthly revenue print just outside the 30-day window, completes the Q2 monthly picture.
• 2026-07-30: Q2 2026 results and earnings call OUTSIDE the 30-day window; flag late July as a blackout for catalyst-driven sizing given binary guide/AP-revenue risk.
• H1 2027: 310mm panel-level packaging line production start structural, not a near-term catalyst.

What Would Change Our Mind

• Bullish re-arm: a higher low holding above the 50-day (~$30.80), then a reclaim of the $36–38 shelf on rising volume coincident with a May/June revenue print showing MoM reacceleration that rebuilds a clean setup worth a fresh probe. A reclaim of $41.10 on volume re-opens the trend entirely.
• A reappearance of insider buying would be a meaningful positive divergence against the current distribution pattern.
• Thesis kill: a weekly close below the rising 50-day MA (~$30.80) confirms the blowoff-top reversal and removes the long setup. Secondary kills the Q2 call cutting the >$3.5B LEAP guide, a CoWoS double-order/digestion scare across the cluster, or two consecutive negative MoM revenue prints.

Correlation Notes

• Tightly correlated with the OSAT/advanced-packaging cluster: Amkor (AMKR), Powertech, and its own SPIL subsidiary. Cluster moves are largely one-directional on CoWoS-supply and Nvidia-allocation headlines.
• Upstream tether to TSMC (TSM) on CoWoS overflow economics and to NVDA as the demand anchor ASE trades as AI-capex beta (β ~1.40), amplifying SOX/SMH semis moves in both directions.
• Shared single-point risk with the entire Taiwan semi complex: any cross-strait geopolitical headline repriced the whole group simultaneously.
• Amkor's ~3x capex step-up is a competitive-supply signal: more OSAT capacity coming online is bullish for the AI-buildout narrative but bearish for long-run pricing power and returns across the packaging group.