{
  "@context": "https://orbyd.app/schemas/dossier.v1.json",
  "ticker": "ASX",
  "name": "ASE Technology Holding Co., Ltd.",
  "url": "https://orbyd.app/dossiers/ASX/",
  "json_url": "https://orbyd.app/dossiers/ASX.json",
  "status": "DORMANT",
  "current_conviction": "LOW",
  "graded_conviction": null,
  "archetype": {
    "code": "a2",
    "n": 2
  },
  "current_thesis": "Advanced-packaging pricing power is real — quotes raised 20%+ (2026-07-01), 2026 capex hiked to US$8.5B, CoWoS-equivalent capacity tripling to 25k wpm — but the tape topped at $45.51 the same day and sits ~15% lower inside a 13% semi drawdown, with a binary Q2 margin print due ~2026-07-30. Fundamentally accelerating, structurally broken, into the print.",
  "invalidation_trigger": "A weekly close below $34 breaks the pre-breakout range floor (2026-06-05 flush low $34.03) and confirms $45.51 as a distribution top; secondarily, Q2 gross margin on ~2026-07-30 below the guided +20–100bps QoQ range, or a Q3 sequential guide under +5% QoQ.",
  "catalyst_date": null,
  "outcome": "OPEN",
  "outcome_date": null,
  "invalidation_fired": null,
  "themes": [
    "semi-foundry-equipment",
    "ai-datacenter-infrastructure",
    "ai-chips-memory"
  ],
  "tags": [],
  "sources": [],
  "notes": [
    "Theme state is split: ACCELERATING on fundamentals, MATURING on the tape (retrospective Benzinga 10-year return listicles 2026-06-30 and 2026-07-16, failed breakout, SOXX -13% in four weeks).",
    "WF6 supply squeeze reported ~2026-07-01 is a process-materials cost input that partially offsets the pricing-hike margin benefit — track alongside gross margin.",
    "Archetype = Picks & Shovels: OSAT infrastructure layer selling packaging/test to the whole AI-chip industry; levered AI-capex beta with no independent demand signal.",
    "EARNINGS BLACKOUT: Q2 2026 full margin/net-income results est. 2026-07-30. Revenue aggregate + June monthly already printed 2026-07-09 (Q2 US$6.050B +25.1% YoY; June US$2.092B +25.9% YoY); only gross margin and Q3 guide outstanding. Treat late July as a blackout for catalyst-driven sizing — binary within days.",
    "Key levels: 50-day MA ~$30.80 = trend line; 200-day MA ~$19.80; $34.03 = bought flush low (2026-06-05) / range floor; $38–41 = supply shelf; $41 shelf reclaim on volume = re-arm level; $45.51 = all-time high (2026-07-01).",
    "BofA PT $48 from $36 (2026-06-24) — sell-side upgrading into strength; return-listicle coverage (2026-06-30, 2026-07-16) = mainstreaming/saturation watch, not early-stage.",
    "Archetype = Picks & Shovels (OSAT infrastructure layer selling packaging/test to the whole AI-chip industry; levered AI-capex beta correlated to TSM/NVDA/AMD/SOXX).",
    "Structure broken not based: failed breakout off $45.51 ATH, ~15% drawdown into a 13% SOXX slump; RSI reset to neutral but no higher-low. Fresh entry is a pass until a base forms above $34 or the $41 shelf is reclaimed."
  ],
  "body_markdown": "## Current Thesis\nThe advanced-packaging shortage stopped being a forecast and became a price. ASE was reported on 2026-07-01 to have raised advanced-packaging quotes by more than 20% — the second AI-driven hike of the cycle, a margin event on a book already sold out. Management is spending behind it: 2026 capex lifted to US$8.5B from US$5.3B in 2025, CoWoS-equivalent capacity tripling toward 25,000 wafers/month, six new plants breaking ground across Taiwan, the US, Malaysia, Japan and Germany. The revenue line agrees — Q2 2026 landed at NT$191.06B / US$6.050B (2026-07-09), +10.0% QoQ and +26.7% YoY, above the +7–9% sequential guide from the 2026-04-29 call, ATM segment NT$126.15B, +36.3% YoY.\n\nThe tape got there first and then broke. Shares tagged an all-time high of $45.51 on 2026-07-01 — the day the pricing story printed — and closed 2026-07-17 at $38.82, ~15% off the high, while the iShares Semiconductor ETF fell 13% in four weeks and surrendered its entire premium to the Nasdaq 100. The name is up roughly 290% over twelve months against a $9.30 52-week low, ~$83.5B market cap, ~59x trailing. So it is a fundamentally accelerating business inside a failed breakout, days ahead of the Q2 margin print (est. 2026-07-30) that assigns a number to the price hikes — a binary that argues for standing aside on fresh capital rather than pressing a broken structure into the event.\n\n## Bull Case\n- Advanced-packaging quotes raised >20% (reported 2026-07-01), a second AI-driven hike that flows to gross margin with no incremental volume required; Q1 2026 gross margin was already 20.1%, +0.6pp QoQ and +3.3pp YoY (2026-04-29).\n- Q2 2026 revenue NT$191,064M / US$6,050M (2026-07-09), +10.0% QoQ and +26.7% YoY, clearing the +7–9% QoQ guide.\n- ATM segment NT$126,148M in Q2, +12.2% QoQ and +36.3% YoY — the high-margin packaging/test core growing ~1.4x the consolidated rate, so mix is a tailwind.\n- June 2026 monthly revenue US$2.092B (2026-07-09), +25.9% YoY and +4.5% MoM, re-accelerating from May +2.2% and April +0.4%.\n- 2026 capex raised to US$8.5B from US$5.3B in 2025, with management flagging room for more — committed against contracted demand.\n- CoWoS-equivalent capacity tripling toward 25,000 wafers/month; six fabs breaking ground in 2026 across five countries, diluting a Taiwan-only footprint.\n- LEAP advanced-packaging revenue guided above US$3.5B for 2026 on the Q1 call, ~10% over the prior US$3.2B plan.\n- Full-process packaging starts H2 2026 with AMD Venice CPUs as first product (est. US$300–400M 2026 line item); SPIL sits inside Nvidia's Arizona AI-infrastructure build as a CoWoS supplier.\n\n## Bear Case\n- The $45.51 all-time high (2026-07-01) was printed on the pricing-hike headline itself; price has since lost ~15% to $38.82 (2026-07-17). The move above the prior $41 shelf failed to hold, and a failed breakout in a name up ~290% in a year is where distribution shows.\n- The semiconductor group is in its worst drawdown in over a year — SOXX -13% in four weeks (2026-07-16), the complex's valuation premium to the Nasdaq 100 fully erased. ASE at ~59x trailing versus Micron near 6.8x forward: money leaving the group is not being put to work in the expensive OSAT.\n- Insider selling of US$349.3M over the trailing three months with zero purchases (as of early June 2026) — the most direct late-cycle distribution signal on the name; a fresh cluster would deepen it.\n- Sell-side is upgrading into strength (BofA PT $48 from $36, 2026-06-24) and the narrative is running through return-listicle coverage (\"how much you'd have made in 10 years\", 2026-06-30 and 2026-07-16) — mainstreaming, a saturation marker, late-stage coverage.\n- The Q2 margin/net-income detail (est. 2026-07-30) is a binary within days; revenue and June monthly are already known, so the print's surprise lives entirely in gross margin and the Q3 guide.\n\n## Setup & Price Structure\nMomentum structure is broken, not yet re-armed. The reversal off $45.51 retraced ~15% and stalled the stock inside the $38–41 supply shelf it had earlier tried to break above. The line that matters below is the $34.03 flush low bought on 2026-06-05, which caps the pre-breakout range; the rising 50-day sits near $30.80 and the 200-day far beneath at ~$19.80, so the longer trend is intact even as the near-term structure fails. RSI reset from overbought toward neutral, which relieves the stretch without building a base. There is no higher-low and no breakout-retest to buy — the constructive re-entry is a reclaim of the $41 shelf on volume or a clean higher low built above $34, neither of which is present. Chasing a ~59x multiple into a group-wide drawdown, ahead of a binary print, with insiders distributing, is the mean-reversion trap this setup is built to punish; the disciplined stance is to wait for a base instead of paying up for extension that has already unwound.\n\n## Catalyst Calendar (next 30 days)\n\n- Ongoing — advanced-packaging pricing/capacity commentary from the Taiwan supply chain; any confirmation or walk-back of the >20% quote hike moves the margin thesis directly.\n\n## Elapsed catalysts\n\n- 2026-07-30 (est.) — Q2 2026 full results: gross margin (guided +20–100bps QoQ against 20.1% in Q1), net income, and the Q3 sequential guide. The revenue aggregate is already public, so the entire surprise is margin and forward guide. A binary event within days; treat late July as an earnings blackout for catalyst-driven sizing. *(passed 10d ago)*\n- ~2026-08-08 (est.) — July monthly net revenue (Taiwan ~10th-of-month cadence). Watch sequential MoM after June's +4.5%; a roll-over toward flat would undercut the acceleration read. *(passed 1d ago)*\n\n## What Would Change Our Mind\nA weekly close below $34 breaks the pre-breakout range floor set by the 2026-06-05 flush low ($34.03) and confirms $45.51 as a distribution top — that closes the constructive read regardless of fundamentals. Secondary breaks: Q2 gross margin on 2026-07-30 landing below the guided +20–100bps QoQ range, or a Q3 sequential guide under +5% QoQ, either of which would signal the pricing hike is not converting to margin; a fresh insider-sale cluster on top of the trailing US$349.3M; or the theme tipping from accelerating to saturated as sell-side targets cluster above spot and coverage goes fully mainstream. The re-arm case is the mirror: a reclaim of the $41 shelf on expanding volume, or a higher low built above $34 into the print, would restore a buyable base.\n\n## Correlation Notes\nASE trades as high-beta AI-capex infrastructure — levered to the same demand curve as TSMC (its foundry customer and CoWoS bottleneck peer), Nvidia and AMD (end-demand for advanced packaging), and the SOXX complex broadly. In the current tape that correlation is a liability: the four-week 13% SOXX drawdown dragged the name off its high irrespective of company-specific news, and the group's lost premium to the Nasdaq 100 means macro and rate rotation is setting the near-term price ahead of the packaging story. Idiosyncratic upside re-couples to the stock only on a margin surprise at the 2026-07-30 print or a further pricing hike; absent that, expect it to move with TSM, the SOXX, and Taiwan supply-chain sentiment. Cross-read: SPIL's role in Nvidia's Arizona build and AMD Venice packaging ties the name to those two roadmaps specifically.",
  "first_seen": "2026-05-01",
  "last_analyzed": "2026-07-26T11:16:13+00:00",
  "last_synthesized": "2026-07-26",
  "last_update_source": "watchlist_research",
  "license": "Content © orbyd. Cite the canonical URL."
}