·Dashboard·Research·Work·Archive
← wiki
ticker stockadvanced-packaging updated 2026-06-01

ATS — AT&S Austria Technologie & Systemtechnik

Thesis

AT&S is the only scaled IC substrate manufacturer in Europe, at the most critical inflection in its history. The core case is a growth-compounder bet: the Kulim, Malaysia campus (~€1B invested, AMD HVM-certified May 2025) transitions AT&S from a PCB maker into a scaled AI/HPC hardware supplier, with revenue guided to nearly double from €1.59B (FY2024/25) to €2.1–2.4B (FY2026/27) at a 24–28% EBITDA margin, while the capex cycle has decisively peaked (€855M → €415M → €250M → revised €200M) and FCF inflects positive.

Verdict / stance: WATCH → SCALE BUY (STAGED). All 5 FundamentEdge gates pass (Gate 2 second-derivative strongly positive; Gate 5 estimate revisions upward). Conviction: Medium-High. The catch is that the stock has already re-rated +420% in 11 months (~€16.54 May 2025 → ~€87 Apr 2026), trading near its 52-week high with a beta of 1.98 — so the FOMO trap is the dominant behavioral risk and the margin of safety at ~€87 is thin. The recommended structure is a staged entry, not a full position: starter 30–35% at ≤€85, add on FY2025/26 full-year confirmation (May 2026), complete on H1 FY2026/27 ramp evidence or a guidance raise.

What has to be true (all three simultaneously): (1) Kulim yield climbs from ~70% start to 90%+ and utilization clears 80%; (2) AMD plus the undisclosed U.S. AI chip customers sustain and grow orders; (3) AT&S delivers the FY2026/27 €2.1–2.4B / 24–28% EBITDA guide, generating FCF to deleverage. The single most important thing that must go right is Kulim ramp execution — yield and utilization — because margin expansion, FCF, and debt reduction all follow from it.

Why the edge may exist: sell-side consensus price targets cluster at ~€50–70 (wide range €15–105), well below the ~€87 price — meaning the street either has not internalized the ramp trajectory or already views the level as fully valued. The name is European-listed, thinly covered (~8–12 analysts), and under-owned by institutions — an information-asymmetry edge for primary research. The risk is symmetric: at €87 the market is roughly pricing mid-base-case delivery with no discount for execution risk.

Snapshot

ATS — AT&S Austria Technologie & Systemtechnik Aktiengesellschaft. Vienna Stock Exchange (Prime Market), ticker ATS / ATS.VI; also Frankfurt/XETRA (AUS). ISIN AT0000969985 (single share class — no dual-class). GICS: Information Technology / Electronic Components. HQ Leoben, Austria. Founded 1987; IPO Frankfurt July 1999 (€12.50 split-adjusted), relisted Vienna 2008. ~13,600 employees (Dec 2025). FY ends March 31. Sector tagged here as semicap-equipment; the business is more precisely IC substrates + advanced packaging + PCB.

Valuation snapshot (as of ~Apr 23–26, 2026):

Metric Value
Share price ~€87 (one source cites €86.90 on Apr 23)
Shares outstanding 38.85M (38,850,000) — unchanged since 2013
Market cap ~€3.4B
Enterprise value ~€4.7B (deep-dive, net debt ~€1.3B); profile cites ~€4.2B / ~€800M net debt — DISCREPANCY, see financials
EV/EBITDA (TTM) ~11x (adj. EBITDA ~€375–410M TTM)
EV/EBITDA (FY2025/26E) ~12x (on ~€390M)
EV/EBITDA (FY2026/27E) ~7–9x if guidance delivered
Net Debt/EBITDA 2.2x (Sep 2025) → 2.0x (Q3 FY2025/26)
P/E (TTM) NM / negative (near-zero to negative net income)
Dividend yield 0% (suspended during investment cycle)
Beta 1.98 (high)
52-week range €12.92–€93.90 (checklist); deep-dive/canonical cite ~€16.54 trough → ~€87+; profile cites a stale Yahoo €10.42–€35.65 window — DISCREPANCY

Key stats: FY2024/25 revenue €1,590M (+2.6% YoY); FY2024/25 adj. EBITDA €408M (25.7% margin); FY2025/26E revenue ~€1,700M; FY2026/27 guided revenue €2.1–2.4B at 24–28% EBITDA margin. Competitive: #5 global ABF substrate maker at ~8% share; only European scaled player.

Business

AT&S makes the physical platform that connects chips to circuit boards — printed circuit boards (PCBs) and IC substrates, the interposer-like platforms that flip-chip BGA packages sit on, bridging a leading-edge die (nanometer pitch, 700W+ thermal) to the millimeter-scale motherboard. Its highest-value product is ABF (Ajinomoto Build-up Film) FC-BGA (Flip-Chip Ball Grid Array) substrates for AI accelerators and server CPUs. Fewer than five companies on earth can make the highest-end ABF substrates at scale; AT&S is the only one headquartered in Europe.

Segments

Segment Rev share Description
Electronics Solutions ~62% (~€990M) HDI PCBs for mobile, automotive, industrial, medical; mature, cash-generative (~20% EBITDA margin); funds substrate investment
Microelectronics ~38% (~€600M, growing fast) ABF FC-BGA IC substrates for HPC/AI; the thesis segment; Kulim + Leoben fabs

A third "Others" segment exists but is immaterial. Microelectronics was ~€350M in FY2023/24, guided to be the primary driver of the €2.1–2.4B FY2026/27 target — implying substrate revenue likely reaches €1.0–1.2B by then.

Customers (no precise per-customer disclosure; estimates)

  • AMD (NASDAQ: AMD) — anchor IC substrate customer at Kulim; HVM-certified May 2025; EPYC + Instinct MI-series. Est. revenue share ~15–20% (deep-dive/checklist) vs ~20–25% (profile) — DISCREPANCY. Kulim is purpose-built and AMD-co-designed; re-qualifying another vendor would take AMD 18–24 months. AT&S's most strategically important and financially healthy customer.
  • Apple (NASDAQ: AAPL) — PCB supply for mobile devices (iPhone HDI main/sub-boards); in Apple Top 200 suppliers. Est. ~25–30% of total revenue (deep-dive) / ~15–20% (profile) — DISCREPANCY. Presumed the "volatile order behavior of a key customer" that disrupted FY2024/25 Q1 guidance.
  • Intel (NASDAQ: INTC) — IC substrate pilot (3nm double-sided ABF) + PCB; "Certified HVM Site"; awarded AT&S Preferred Quality Supplier 2025 (one of 26 globally). Est. ~5–10%.
  • Three undisclosed U.S. AI chip companies — "further key players of the global AI chip industry"; won in FY2023/24; ramping/qualifying at Kulim through FY2025/26; co-financing Leoben capacity.
  • Minor: Sony Ericsson / Fairphone (mobile PCB).

Concentration: Apple = PCB concentration, AMD = substrate concentration; combined top 2 ≈ 40–50% of revenue (profile: top-3 ≈ 50–60%). This is the primary business risk.

Moat / competitive position

Narrow but real. (1) Geographic uniqueness — sole scaled European IC substrate producer; EU Chips Act beneficiary; Western supply-chain-security argument. (2) Qualification lock-in — 18–36 months to qualify a new substrate supplier; switching costs high once production-qualified. (3) Technology — mSAP (modified Semi-Additive Process) heritage from the PCB side (AT&S was an early commercializer); ABF process IP over 20+ years; ahead on glass core substrates via Intel collaboration. (4) Oligopoly pricing — 5 players control ~74–75% of capacity. Moat is NOT wide: Unimicron, Ibiden, Shinko are larger and better-capitalized; AT&S competes on geography + program-specific qualification, not absolute scale. Margins are volume-leveraged (yield and utilization are the dominant P&L drivers), not inherently wide — quality verdict: durable but asset-heavy, capital-intensive, conditional on continued capex to keep pace with the technology roadmap.

Production footprint

Site Location Output / status
Leoben HTB1/HTB2 Austria (HQ) ML/HDI PCBs, embedded solutions, IC substrate cores — operating
Leoben HTB3 Austria Europe's first IC substrate fab — opened Jun 2025; ramping to 20,000 FC-BGA panels/mo by 2027 (EU Chips Act funded); ~700 jobs targeted
Fehring Austria Standard/specialty PCBs — operating
Shanghai China IC substrates + mobile PCBs; group R&D hub — operating
Chongqing China High-volume PCBs — operating
Nanjangud Karnataka, India Industrial/auto PCBs — operating
Kulim Campus Kulim Hi-Tech Park, Kedah, Malaysia Large-scale ABF FC-BGA; AMD anchor; ~255,000 sqm; ~500 high-tech machines; RM 5B (~€1B) invested — HVM since May 2025, ramping

Divested: Ansan, South Korea — sold Sep 2024 for ~€405M to SO.MA.CI.S. (book ~€38M EBITDA); ~€317M after-tax gain; key de-leveraging event. Workforce distribution proxy: China 62.9%, Austria 16.4%, Malaysia 11.4%, India 8.7%.

Value chain & key suppliers

AT&S sits at the IC substrate layer (high barriers, high switching costs, 5-player oligopoly pricing), feeding OSAT/packaging (TSMC CoWoS, Amkor/ASE, Intel Foundry) → AI server / HPC / consumer CPU. Upstream bottlenecks: Ajinomoto (2802.T, ~$9B cap) — monopoly supplier of ABF film (the dielectric the product is named for); controls film allocation globally; the single most structurally advantaged upstream node. Nitto Boseki / Nittobo (3110.T, ~$1B cap) — dominant supplier of low-CTE T-glass fiber cloth for large-format AI substrates; T-glass shortage extending into 2026, flagged by AT&S management. Panasonic (6752.T) — laminates/core, partial alternatives. Laser/drilling tools (LPKF, ESI) — 28-week lead times constrain capacity adds.

Financials

Revenue history (FY ends March 31)

FY Revenue (€M) YoY
FY2019/20 ~1,170
FY2020/21 ~1,490 +27%
FY2021/22 ~1,790 +20%
FY2022/23 ~1,791 (profile) / ~1,790 ~0% (cycle peak)
FY2023/24 ~1,549–1,550 -13.5% to -14% (Ansan reclassified + market weakness)
FY2024/25 €1,590 +2.6%
FY2025/26E ~€1,700 +7% reported / +20% organic
FY2026/27E (guided) €2,100–2,400 +24–41% (vs €1.59B FY24/25 implies 32–51% — nearly all organic, Kulim-driven)

Margins / income statement (profile detail; reported FY2024/25 distorted by Korea gain)

Metric FY2022/23 FY2023/24 FY2024/25 FY2025/26E
EBIT €146M €31M €277M* +ve (guided)
EBIT margin ~8.2% 2.0% 17.5%* ~2–5% adj.
Net income €137M -€37M €90M* Negative (loss guided)
EPS €3.03 -€1.39 €1.86* ~-€1.93 consensus
Adj. EBITDA ~€417M €384M €408M ~€390–435M
Adj. EBITDA margin ~23.3% 24.8% 25.7% ~23%

*FY2024/25 EBIT/net income boosted by ~€317M after-tax gain from the Korea plant sale; adjusted EBIT was €97M (6.1% margin). Gross profit / gross margin are not disclosed by AT&S.

Current YTD (Q1–Q3 FY2025/26, 9 months to Dec 2025)

Period Revenue EBITDA EBITDA margin EBIT
Q1–Q3 FY2024/25 €1,196.7M €231.7M 19.4% €(1.4)M
Q1–Q3 FY2025/26 €1,314.0M €296.8M 22.6% €34.0M
YoY change +9.8% +28.1% +320 bps +€35.4M

Net loss €39.3M (improved from €95.3M loss prior year). Q3 standalone (Oct–Dec 2025): revenue €467.7M / €468M (+17.9% YoY), EBITDA €122.1M at 26.1% margin, net profit €24.2M (first profitable quarter in a while; highest quarter on record).

Second-derivative / quarterly cadence

Q1 FY25/26 Q2 FY25/26 Q3 FY25/26
Revenue (€M) €399 €447 €468
QoQ growth +12% +5%
YoY growth +14% ~+8% +18%
EBITDA margin ~20% ~21% ~26% est.

Second derivative is positive and accelerating — exactly what you want as Kulim ramps. (Note: text variously describes Q3 as "+18% QoQ" and "+18% YoY"; the table shows +5% QoQ / +18% YoY — the +18% figure is the YoY number; QoQ leg-up phrasing in the fragments is loose.)

Incremental margin

Q1 vs Q1 Q2 vs Q2 Q3 vs Q3
Δ Revenue +€50M ~+€48M ~+€69M
Δ EBITDA ~+€20M ~+€25M ~+€20M
Incremental EBITDA margin ~40% ~52% ~29%

Average incremental EBITDA margin ~40%, well above the reported 22.6% — confirms new Kulim revenue is above-average-quality, operating leverage is real.

Balance sheet & cash flow

Metric FY2024/25 Q3 FY2025/26 YTD
Net debt ~€1.4–1.6B (canonical/deep-dive); profile implies ~€1.0B at Mar 2025 from 2.5x×€408M — DISCREPANCY ~€1.3B
Net debt / EBITDA 2.5x (Mar 2025) → 2.2x (Sep 2025) 2.0x (improving; was 6.1x at Dec 2024 cycle peak)
Operating cash flow -€75M FY24/25 (profile) / -€91M H1 prior yr; -€29.4M prior YTD €331.8M (Q1–Q3) — vs -€29.4M prior year
Net capex €415M FY24/25 (€327.5M prior-year YTD) €108.4M (down 66.9% YoY)
FCF Negative Turning positive — inflection confirmed
Cash on hand €485M (H1 entry) / €676–793M €793M / €886M incl. credit lines
Equity ratio 23.2% (profile) → 19.2% (Sep 2025) — thin; covenant-breach risk flagged est. improving
ROCE / ROIC (1.6%) prior year 2.2% (turning positive)

The cash-flow inflection is the most important recent data point: operating cash flow went from roughly -€91M (H1 prior year) to +€209M (H1 FY2025/26), and the capex collapse is generating real free cash. Capex path: €855M+ expansion peak → €415M (FY24/25) → €250M (FY25/26 guidance) → revised €200M (Q3). ROCE of 2.2% reflects the huge invested-capital base vs still-low earnings; peers (Ibiden, Unimicron) run 10–15% ROIC mid-cycle, implying significant ROIC expansion embedded in the guidance scenario.

Financing structure

€350M perpetual hybrid bonds (5% coupon, 2021; classified as equity under IFRS, subordinated) — market instruments, not related-party. IFC (World Bank) $250M debt facility secured Mar 2025 for Kulim. EU Chips Act €500M grant for Leoben HTB3 (awarded Nov 2025). No dividend, no buybacks — pure deleveraging phase. Cost reduction: €120M in FY2024/25, further €130–150M targeted FY2025/26 — supports margin recovery without top-line growth. IFRS-reported, Big-4 audited; no evidence of accounting irregularities.

Dilution

Share count 38.85M, unchanged throughout the entire expansion — AT&S financed €1B+ Kulim with debt (hybrids, IFC, bank) + asset sales (Korea €405M) + grants (EU €500M), NOT equity. The aborted €450–500M capital increase (planned via Öbag, Austria's state holding) in 2023/24 was a governance fight (Androsch blocked it; Dörflinger supported it). No material convertibles or warrants. Genuinely shareholder-friendly; dilution risk LOW near-term, re-emerging only if the ramp materially underperforms and emergency liquidity is needed.

Industry landscape

ABF/IC substrate is a consolidated oligopoly — five players control ~74–75% of capacity, with high barriers (>€500M greenfield capex, 5+ years to master ABF yield, 18–36 month qualification per product, Ajinomoto ABF film allocation as a gating item). Approximate ABF market share: Unimicron (3037.TW, ~26.6%), Ibiden (4062.T, ~14.6%), Nan Ya PCB (8046.TW, ~13.5%), Shinko Electric (6967.T, ~12.8%), AT&S (~8.0%). The cycle turned sharply from the 2023–2024 glut: AI accelerator demand is institutionally budgeted by hyperscalers on multi-year commitments, while supply additions are mechanically constrained (28-week laser-tool lead times, Ajinomoto-rationed ABF film, tight T-glass). ASP evidence confirms the turn — ABF ASPs rose from ~$65 (2024) to ~$82 (2025), +26%, with large-format AI substrates (>80mm) at $100–200+. Projected supply shortfall: 10% in H2 2026 → 21% in 2027 → 42% in 2028. TAM (sources vary): advanced IC substrate ~$11.4B (2026) → $15.7B by 2031 (6.6% CAGR); ABF subset ~$5.3B (2025) → $9.5B by 2033 (10.6%) per deep-dive, or ~$5.3B → $10.2B by 2030 (9.9%) per profile — DISCREPANCY kept. See sector page: advanced-packaging

Management

Leadership (full refresh executed 2023–2026)

  • Dr.-Ing. Michael Mertin — President & CEO (May 1, 2025 – Apr 30, 2028, 3-yr term). Physics doctorate (RWTH Aachen; Fraunhofer Institute for Laser Technology postdoc); ~10 yrs Carl Zeiss; CEO of JENOPTIK 2007–mid-2017 (10-yr tenure: grew revenue >35% from ~€500M to ~€670M by 2015, earnings from near-breakeven to ~€50M net, refocused to a clean three-division model). 2017–2025 independent PE/M&A advisory. Relevant precision-manufacturing/capex-cycle background; clean JENOPTIK exit. Yellow flag: 8-year gap from last operating role before the most critical ramp in AT&S history. Hired to execute someone else's blueprint, not originate strategy.
  • Gerrit Steen — CFO (Feb 1, 2026 – 2029). MBA WHU Otto Beisheim. ~20 yrs Fresenius (Global CFO Fresenius Kabi; CEO Fresenius Kabi Asia); Heraeus (Asia CEO); Group CFO DAMAC Group (Dubai, >$5B revenue real estate/tech) until Dec 2025. Strong IFRS/multi-currency/Asia and balance-sheet-stress background — relevant to the deleveraging phase. Flag: DAMAC is private, harder to validate externally. Only ~2 months in role at time of research — cannot yet assess.
  • Dr. Peter Griehsnig — CTO (Apr 1, 2023 – 2028). Applied Physics doctorate (TU Graz). AT&S lifer since 2001 (22+ yrs); Shanghai-based since 2002; designed, built, and commissioned both the Shanghai and Chongqing plants; led mSAP commercialization; COO Mobile Devices & Substrates from 2012. The single most important executive — institutional memory, knows exactly where the yield problems are and how to fix them. Kulim is a scaled evolution of his China-plant playbook. Key-person risk: effectively irreplaceable near-term, no disclosed succession plan.
  • Andy Mattes — Supervisory Board Chair (Jul 3, 2025 – 2030). Business degree LMU Munich. ~20 yrs Siemens; 7 yrs HP senior roles in California; CEO Diebold Nixdorf 2013–Dec 2017 (led the Diebold/Wincor Nixdorf merger to the world's largest ATM provider — but stepped down Dec 2017 after a downward earnings guidance revision and execution problems; stock had underperformed); CEO Coherent 2020–2022 (operationally solid, short tenure); McKinsey Senior Advisor; Cohu board. Independent under Austrian CGC C-Rules 53 & 54. Yellow flag: the Diebold departure pattern-matches AT&S's exact situation (ambitious manufacturing transformation, complex execution) — watch board resolve if Kulim disappoints.

Prior CEO Andreas Gerstenmayer resigned Sep 30, 2024 (forced out in the Androsch boardroom conflict). Petra Preining / Simone Faath CFO lineage retired (Aug 31, 2025). Both Kulim and Leoben were committed under Gerstenmayer; Mertin inherited and must execute them.

Ownership

Holder Type Stake
Dörflinger Privatstiftung Founding-family foundation (Willi Dörflinger; MBO 1994) ~18% (~€612M; ~7.0M shares)
Androsch Privatstiftung + AIC Late Hannes Androsch's foundation + Androsch Intl. Consulting ~17.55% (~€597M; ~6.8M shares)
Free float Institutional + retail ~61–64%
Norges Bank Sovereign wealth (passive) ~1.66%
Vanguard (combined) Index ~1.3%

Founding bloc ≈ 35.6% combined — de facto blocking power, no anti-takeover structures needed; no dual-class shares. Institutional ownership was only ~2.49% through June 2025 (under-owned in distress), likely to rise post-re-rating. Their wealth is overwhelmingly AT&S-concentrated — genuine, irreversible skin in the game; they cannot exit without moving the market.

Capital allocation & alignment

Primarily greenfield capex (Kulim ~€1B; Leoben HTB3 ~€300–400M est., EU co-funded; Chongqing ~€300M) over acquisitions. The Korea divestiture (€405M, Sep 2024) was excellent, well-timed capital recycling. Zero share dilution over a decade of massive investment (grade: excellent on equity discipline). Cycle timing was imperfect — the €800M+ capex commitment through FY2022–2024 created leverage stress — but the directional AI-substrate bet has proven correct. Overall capital allocation grade: B / Yellow-Green. Compensation: Austrian CGC remuneration report; STI tied to EBITDA + revenue + strategic milestones, LTI multi-year on TSR/operational metrics; SBC minimal (<0.5% of shares/yr, consistent with the flat share count). Specific LTI hurdle structures for Mertin/Steen not yet retrievable — verify against the FY2024/25 annual report (due May 2026). If hurdles sit at/near the FY2026/27 guide, management is well-aligned; if below, an easy layup. No unusual perks disclosed.

Insider transactions

  • Dörflinger foundation bought €13.4M (326,342 shares @ €41.10, Feb 8, 2022) plus ~€1.08M more June 2022 at €53–55 — now worth 2–4x and NOT sold through the +420% re-rating (strongest insider signal).
  • Dörflinger also bought 11,147 shares @ €26.72 on Nov 21, 2025 (profile) — buying through the trough.
  • Gertrude Tumpel-Gugerell (independent supervisory board member, former ECB Executive Board) bought twice (Dec 12, 2025 and Apr 14, 2026) on the open market — genuine board-level alignment signal.
  • New management (Mertin, Steen, Griehsnig) have NO disclosed open-market purchases — a yellow-flag gap to track for the next 12 months (Mertin/Steen may be in blackout / still building positions).
  • Günter Pint SELL Feb 19, 2026 (apparent departing executive, no pattern of concern).

Governance flags

Supervisory board (8–9 members, majority independent) is materially improved and now semiconductor-literate: Hermann Eul (former Intel CTO of Germany/Austria) and Lars Reger (NXP CTO) bring active industry oversight; Regina Prehofer (finance/audit), Tumpel-Gugerell (macro). Georg Riedl — Androsch foundation representative, independent under C-53 but NOT C-54 — is the remaining Androsch legacy. Hannes Androsch passed Dec 2024, partially resolving the historical Androsch–Dörflinger friction (two large private foundations co-controlling a public company with conflicting strategic views — the structural governance weakness that forced out Gerstenmayer). The one unresolved RPT item is AIC (Androsch International Consulting): possible historical advisory fees to AT&S during Androsch's dual board-chair/shareholder role — no confirmed material payment; risk reducing post-death; warrants forensic annual-report review. Clean IFRS subsidiary structure (China, Malaysia, India, Austria operating subs + standard Austria/Netherlands holding entities); no off-balance-sheet SPVs, no circular ownership. Not SEC-registered (no ADR), no FMA enforcement actions, no material personal litigation on any current executive.

Mgmt-DD verdict (2026-04-26): Overall Grade B / Yellow (deep-dive phrased it B / Yellow-Green)

Dimension Rating
Skin in the Game Yellow (founding bloc not selling = strong; Mertin/Steen no open-market buys)
Holdings Concentration Green
Shell / Cross-Holdings Yellow-Green (AIC historical RPT reducing post-Androsch)
Capital Allocation Yellow-Green (zero dilution; correct bet; cycle timing imperfect)
Comp Alignment Yellow (principles correct; hurdles unverified)
Credibility / Follow-Through Yellow (~60–65% historical rate; improving under Mertin)
Governance Quality Yellow-Green (board now semi-literate; Riedl legacy remains)
Litigation Green (no FMA, no personal litigation)

Upgrade to B+ if the first full year under Mertin (FY2025/26, due May 2026) delivers at/above the €1.7B / ~23% EBITDA guide. Red flags: none identified.

Guidance credibility history

Gerstenmayer era over-promised on medium-term targets (the FY2022/23 "€3B by FY2025/26, 27–32% EBITDA margin" target was pushed right by a year to FY2026/27 in Mar 2023) and cut guidance twice: FY2023/24 (revenue cut to ~€1.55B, actual €1,549M; EBITDA margin guided ~25%+ → actual ~17.7% adj. — MISS) and FY2024/25 (€1.7–1.8B orig. cut to €1.5–1.6B Oct 2024, actual €1,590M near top of revised range; full-year guidance abandoned for H1-only). Under Mertin the reset to H1-only guidance is deliberate conservatism: H1 FY2025/26 guided ~€820M → actual €846M (+3.2% beat — first clean beat). The FY2026/27 €2.1–2.4B guide explicitly assumes no further USD depreciation; USD has fallen ~10% since Dec 2024 (€1.07 → €1.17), moving guidance to the lower end — disclosed transparently, not buried.

Catalysts & risks

Catalysts (bull)

Near-term (0–12 months): FY2025/26 full-year results (~May 2026) confirming ~€1.7B revenue and ~23% EBITDA margin — first major de-risking event; any Kulim utilization / panel-volume / yield disclosure; new co-financing customer names at Leoben; Q4 FY2025/26 print (Jan–Mar 2026) could be a record quarter given Q3 momentum.

Medium-term (1–3 years): FY2026/27 revenue ramp to €2.1–2.4B — the single most important de-risking event; EMIB-T capability addition at Kulim (management flagged the need, timing undisclosed); dividend reinstatement once net debt/EBITDA < 1.5x (signals balance-sheet normalization).

Long-term (3–10 years): glass core substrate commercialization — AT&S and Intel jointly advancing TGV (Through-Glass Via) technology; first commercial revenues possible post-2027; AT&S is on the right side of this disruption rather than threatened by it.

Secular tailwinds: AI accelerator build-out (each AI chip needs 1–4 high-end substrates at $80–200+ ASP; hyperscalers spending $50–100B+/yr); chiplet/advanced-packaging complexity (Intel Foveros/EMIB, AMD 3D V-Cache, NVIDIA multi-die = more substrate content per design); ABF supply scarcity (10% → 42% shortfall 2026–2028 with ASP uplift); EU/Western supply-chain mandates (AT&S the sole European option); cost-reduction program restoring margins.

Risks (bear)

  1. AMD concentration (High, structural): AMD ~15–20% of revenue and the Kulim anchor; a >20% AMD order cut (NVIDIA AI share gains, inventory build, AMD's own cycle) would collapse Kulim utilization and margins. Mitigant: 3 undisclosed U.S. AI customers + Intel diversifying, but Leoben customers take 12–24 months to offset a Kulim AMD cut. Partially closeable.
  2. EUR/USD FX (High, structural): ~50%+ of revenue USD-denominated (U.S. substrate customers); cost base EUR/MYR/CNY/INR. EUR +~10% vs USD since the guide was set (€1.07 → €1.17). Each 5% EUR appreciation ≈ -€40–50M revenue / -€20–30M EBITDA. Guidance moved to lower end on this; not closeable, only short-term hedgeable.
  3. Kulim yield fails to reach 90% (Medium): started ~70%; peers run 90%+; each point of yield 70–90% has meaningful EBITDA impact. If yield stalls at 80%, EBITDA margin stays ~20–22% vs 24–28% and the bull case dissolves. Mitigant: CTO Griehsnig's China-plant yield-ramp track record — but Kulim is the largest, most complex site AT&S has ever run. Closes only when margin reports confirm ≥90% by end-CY2026.
  4. Leverage / thin equity cushion (Medium): equity ratio ~19.2%; net debt/EBITDA 2.0–2.2x; management flagged covenant-breach risk if business cases deviate. Less margin for error than a debt-free peer. Closes as EBITDA grows; expected by FY2026/27.
  5. New management execution (Medium): new CEO + CFO during the most critical ramp; Griehsnig provides continuity but he is CTO, not CEO. Closes over 12–18 months of delivery.
  6. Governance (Medium): Androsch–Dörflinger bloc historically confrontational; cannot be fully closed (structural, depends on two foundations' alignment) — but improving post-Androsch.
  7. T-glass / ABF film scarcity (Medium): structural bottleneck (Nittobo T-glass, Ajinomoto film), not AT&S-specific; partial mitigation via dual-source development and design changes.
  8. EMIB-T capability gap at Kulim (Medium): investment underway; closes post-FY2026/27.
  9. Glass core substrate disruption (Low, 5–10 yr): AT&S is Intel's glass-core partner — early-mover, on the right side. Other emerging threats: in-house substrate fabs at NVIDIA/AMD (theoretical, capex-prohibitive); Chinese domestic substitution (geopolitically constrained from AMD/Intel/NVIDIA today); TSMC CoWoS (a different stack level, not a substrate displacement).

Bear case

FY2026/27 revenue stuck at €1.7–1.8B (guidance miss from AMD softness + yield problems + FX), EBITDA margin 20–21% (~€360M EBITDA), net debt still ~€1.1–1.3B. At 8x bear-case EBITDA → EV ~€2.9B → shares re-rate to ~€35–46 (the checklist's precise bear is ~€46, -47% from €87; deep-dive cites €35–45). Thesis invalidators: AMD materially reduces Kulim orders (>20% volume cut); Kulim yield fails to reach 85% by end-CY2026; an additional large Kulim Phase-2 capex commitment before leverage normalizes (<1.5x); EUR appreciates to €1.25+/USD with a mid-year FY2026/27 guidance cut.

Valuation / DCF

Multiples (Apr 2026)

Metric Value
Market cap ~€3.4B
EV (est.) ~€4.7B (deep-dive); ~€4.2B (profile)
EV/EBITDA (FY2025/26E, ~€390M) ~12x
EV/EBITDA (FY2026/27E, ~€500–650M) ~7–9x
EV/Revenue (FY2025/26E) ~2.8x
EV/Revenue (FY2026/27E) ~2.0–2.2x
P/E (FY2025/26E) NM (near-zero net income)
Dividend yield 0%

AT&S trades at a modest premium to immediate peers on FY2025/26 earnings but at a discount on FY2026/27 guidance — implying the market is only partially pricing the ramp. Peer EV/EBITDA (NTM): Unimicron ~8–10x, Ibiden ~10–14x, Shinko ~8–11x. The wide analyst price-target dispersion (€15–105) reflects genuine execution uncertainty.

Scenario analysis (share price = (EV − net debt) / 38.85M shares; net debt declining toward ~€0.9–1.0B by FY2026/27 in base case)

Scenario FY2026/27 Revenue EBITDA margin EBITDA EV/EBITDA Implied EV Implied share price
Bear €1,800M 20% €360M 8x €2.9B ~€40 (checklist precise: ~€46)
Base €2,200M 25% €550M 10x €5.5B ~€105
Bull €2,400M 28% €672M 12x €8.1B ~€175

At ~€87, the market is pricing roughly mid-base-case delivery, with no discount for execution risk. Deep-dive base target €90–110 (FY2026/27 re-rating on EBITDA delivery); expected return 0–25% from current levels, asymmetric to the bull on a beat. Bull exit zone €110–175. Would you still buy 10–15% higher? No — at €95–100 the FY2026/27 base case is fully priced; at €100+ you need the bull case. Margin of safety at ~€87 is thin to moderate. Consensus: Buy (9 analysts: 0 strong sell, 2 sell, 1 hold, 4 buy, 2 strong buy); average target ~€50–70 — significantly below the price, the key signal that estimates/targets need upward revision if FY2026/27 guidance is delivered. No formal multi-stage DCF was built (asset-heavy, capex-cycle, near-zero current net income make EV/EBITDA the operative frame).

Decision log

  • 2026-04-26 — /profile: full company profile written (profile skill).
  • 2026-04-26 — /deep-dive: full investment write-up. Conviction Medium-High; base target €90–110; suggested sizing 3–5% of portfolio (not a core position at current multiples); scale in over 2–3 tranches weighted to FY2025/26 full-year confirmation. Quality verdict: durable but asset-heavy.
  • 2026-04-26 — /mgmt-dd: overall management grade B / Yellow (deep-dive phrasing B / Yellow-Green). Founding bloc 35.6% concentrated and not selling = strongest alignment; CTO Griehsnig institutional anchor; board now semiconductor-literate (Eul/Reger). No red flags. Upgrade to B+ contingent on FY2025/26 delivery at/above €1.7B / ~23% EBITDA.
  • 2026-04-26 — /checklist (Pre-Buy): VERDICT = WATCH → SCALE BUY (STAGED). All 5 FundamentEdge gates pass (Gate 2 second-derivative strongly positive; Gate 5 estimate revisions upward). FOMO flagged (+420% in 11 months, near 52-week high €93.90, beta 1.98). Valuation borderline (~12x FY2025/26E EBITDA; ~8x FY2026/27E if guidance delivered). Pre-Buy scorecard: thesis clear (Yes), business understood (Yes), FundamentEdge (Yes, all 5), incentives partially aligned (Yellow), financials Yellow (leverage 2.0x, equity ratio ~19%), valuation borderline (Yellow), behavioral FOMO flagged, technicals Wait/Partial, sizing OK if staged, exit plan Yes. Technical verdict: wait for better entry (€75–80 prior support or post-results) — near 52-week high, high beta.
    • Entry plan: Starter 30–35% at ≤€85 now → Add 1 (35–40%) on FY2025/26 full-year confirmation May 2026 (≥€1.7B revenue, ≥22% EBITDA margin) → Add 2 (25–30%) on H1 FY2026/27 ramp evidence or guidance raise. Holding period 18–36 months. Position size 3–5% of portfolio max.
    • Stop-think: €60 (bear-case pricing without a fundamental reason). Thesis-break exits: AMD order cut >20% (exit immediately); Kulim yield stall <85% by Dec 2026; net debt/EBITDA not declining toward 1.5x by FY2026/27; FY2026/27 guidance cut. Bear case: ~€46 (8x EBITDA on €360M). Bull case / bull exit: €110–175 (trim/exit 50–75% on FY2026/27 revenue ≥€2.2B, EBITDA margin ≥26%). Max loss tolerance ~40% drawdown (€87 → €52) before forced reassessment.
  • 2026-04-26 — Briefing published: AT&S (ATS) · vault.
  • 2026-05-30 — Consolidation (W3): five research fragments merged into this thesis-first canonical page; sector reassigned to semicap-equipment per reorg taxonomy (business is more precisely IC-substrate/advanced-packaging — noted for sector-page placement). No new pricing or thesis change; all figures carried forward as of the Apr 2026 research snapshot.

Sources

Folded-in fragments (all dated 2026-04-26 unless noted):

  • ats.md — prior canonical page (created 2026-04-26, updated 2026-05-15)
  • ats-deep-dive.md — /deep-dive full write-up
  • ats-mgmt-dd.md — /mgmt-dd management due diligence
  • ats-buy-checklist.md — /checklist pre-buy checklist
  • ats-vi-profile.md — /profile company profile

External sources cited across fragments:


Consolidation queue (merged 2026-05-30)

These five research fragments were folded into this canonical page and remain live pending Pink's archive confirmation:

  • [ ] ats-deep-dive.md
  • [ ] ats-mgmt-dd.md
  • [ ] ats-buy-checklist.md
  • [ ] ats-vi-profile.md
  • [ ] ats.md

Source updates (auto-maintained)

Intake (May 12, 26) - cu-wiring-resin-primer

AT&S is identified as a scaled ABF substrate fabricator consuming the concentrated chemistry inputs (MEC CZ microetchant, Ajinomoto ABF film, Mitsui MicroThin carrier foil) that gate yield and ramp speed at Kulim; the primer frames substrate fabricators as owning customer relationships and qualification slots while chemistry suppliers capture the spec inside each slot.

Relevant to your thesis: Directly reinforces the Kulim yield/utilization execution risk — AT&S's ramp trajectory depends on single- or two-vendor consumables (ABF film, MicroThin foil, CZ chemistry) where any supply disruption or qualification delay would slow the yield climb from ~70% to 90%+ that underlies the FY2026/27 guidance.

Source: intakefile://cu-wiring-resin-primer.md