Register D | Date: 2026-04-26 | Mode: New Idea / Full Write-Up > Industry: Semiconductor Equipment — Advanced Packaging Laser + Plasma Tools > Prior primer: SiP/OSAT/Advanced Packaging covered 2026-04-20 (6451 deep-dive → primer)
Register D | Date: 2026-04-26 | Mode: New Idea / Full Write-Up Industry: Semiconductor Equipment — Advanced Packaging Laser + Plasma Tools Prior primer: SiP/OSAT/Advanced Packaging covered 2026-04-20 (6451 deep-dive → primer)
Advanced packaging semiconductor equipment is covered in the
sip-osat-rf-primer (written 2026-04-20). The specific laser
+ plasma equipment niche overlaps with optical components (6777
deep-dive) and the fiber laser / directed energy work (LASR deep-dive).
No separate primer needed — industry context embedded in this
write-up.
Bull thesis (2-3 sentences): E&R is the only identified company to have validated TGV (Through-Glass Via) laser processing at production accuracy for a major North American IDM — almost certainly Intel — after a five-year co-development program. With Intel’s EMIB ramp accelerating in 2H26 and EMIB-T in 2026-27, E&R is entering a production equipment market that did not exist at scale 18 months ago. The AIS multi-module integration service has already secured “significant” North American orders into 2027 and Q4 2025 showed the first operating profit in five quarters — the fundamental inflection may be underway.
| Item | Value |
|---|---|
| Current price | TWD 156.50 (Apr 24, 2026) |
| Market cap | TWD 16.44B (~USD 503M) |
| Enterprise value | TWD ~16.72B |
| EV/Sales (FY2025) | ~9.2× |
| Target price | No explicit target; 1 analyst PT TWD 98 (set pre-run; likely stale) |
| Conviction | Medium — thesis is real but execution risk is high and timing uncertain |
Full name: E&R Engineering Corporation / 鈦昇科技股份有限公司 Ticker / exchange: 8027 / Taipei Exchange (TWO) Sector / industry (GICS): Information Technology — Semiconductors & Semiconductor Equipment HQ: Yanchao District, Kaohsiung 824, Taiwan Founded: 1994 | Listed: Taipei Exchange (TWO) Website: https://en.enr.com.tw/ Latest investor presentation: Not found. IR page: https://en.enr.com.tw/
What E&R does in plain language: It makes machines that prepare, clean, drill, and mark semiconductor packages. Before a chip is bonded to a substrate, E&R’s plasma tool removes oxide contamination from the bonding surface. After glass is laser-drilled for a through-glass via, E&R’s laser did the drilling. After a fan-out panel is diced into individual packages, E&R’s laser did the scribing. The machines are sold to OSATs and IDMs; they are capital equipment — one-time revenue per unit with an aftermarket (spares, service, upgrades).
Key business lines (estimated revenue split — E&R does not disclose segments):
| Line | Products | Est. Revenue % |
|---|---|---|
| Laser Solutions | Marking (wafer ID, package backside), micromachining (scribing, grooving, cutting, TGV drilling), laser debonding | ~55-65% |
| Plasma Solutions | Pre-bonding cleaning, oxide removal, pre-underfill/mold treatment, post-drill de-smear, descum after debond | ~25-35% |
| Other | FPC processing, tape-and-reel, UniDRON lab equipment | ~5-10% |
Business model: Capital equipment — design, manufacture, sell. Revenue is one-time per unit with an aftermarket tail (field service, consumables, engineering upgrades). Machines are manufactured, assembled, and tested entirely at Yanchao HQ in Kaohsiung with key components sourced externally (Coherent for femtosecond lasers, HIWIN for precision motion stages). Order-to-revenue lead time is typically 12-24 months for complex capital equipment; revenue is therefore a lagged indicator of order-book health.
Geographic mix: Not disclosed. Service network spans Taiwan, mainland China (Nantong plant), Southeast Asia, and the U.S. (Portland/Hillsboro OR, Phoenix AZ). The North American revenue share has been growing — the Hillsboro site opened early 2026, and AIS North American orders are described as “significant.”
| Facility | Location | Purpose | Status |
|---|---|---|---|
| Yanchao HQ | Kaohsiung, Taiwan | R&D, manufacturing, assembly, testing — the core facility | Operating |
| Qiaotou Science Park plant | Kaohsiung, Taiwan | Capacity expansion: AI/HPC, glass substrate, automotive electronics, SiC | Under construction; operating license ~2026 |
| Nantong Plant | Jiangsu, China | Regional manufacturing for China market | Opened ~July 2024 |
| Hillsboro, OR (site 2) | Oregon, USA | Second North American site; adjacent to Intel campus | Opened early 2026 |
| Portland area (site 1) | Oregon, USA | Existing service hub | Operating |
| Phoenix, AZ | Arizona, USA | Field service | Operating |
Asset-heavy for a Taiwanese equipment company: the Qiaotou plant expansion has driven TWD 495M capex in FY2024 and TWD 328M in FY2025, turning FCF sharply negative. Once operational (~2026), this plant should be the capacity base for TGV and glass substrate order fulfillment.
E-Core System Alliance (August 28, 2024): E&R-led consortium for glass substrate manufacturing. Non-equity alliance of complementary suppliers:
| Partner | What They Do | Public? |
|---|---|---|
| Manz AG (M5Z.DE) | Glass handling and laser processing equipment | Yes — XETRA |
| Coherent (COHR) | Femtosecond laser light sources (also a component supplier to E&R’s own machines) | Yes — NYSE |
| HIWIN (1515.TW) | Precision linear motion stages and ball screws | Yes — TWSE |
| Scientech | Wet chemical etching systems | Private (Taiwan) |
| ShyaWei Optronics | Automated optical inspection (AOI) | Private (Taiwan) |
| Lincotec, STK Corp., Skytech, Group Up | Sputtering and ABF lamination equipment | Private |
| Mirle Group (2374.TW) | System integration and automation | Yes — TWSE |
| Keyence Taiwan (6861.T subsidiary) | Sensors and measurement | Yes (via parent) |
The alliance’s purpose is to offer a validated, turnkey glass substrate process flow to IDM customers — so a customer evaluating glass core can validate the full manufacturing sequence from one coordinated group rather than five separate vendors. This reduces customer qualification risk and positions E&R as the integrating node.
North American IDM TGV co-development (5-year program): E&R disclosed it worked for five years with a “North American IDM” to develop and validate laser modification TGV technology. The process passed validation in 2024: up to 8,000 vias/second on fixed-pattern layouts, 600-1,000/sec on random layouts, +/-5 µm accuracy. The customer is unnamed. Given that: (1) Intel is the only North American IDM with an active glass substrate + EMIB program; (2) E&R opened a second site specifically in Hillsboro, OR (Intel’s campus city); and (3) Intel publicly displayed a glass + EMIB module at NEPCON Japan in late 2025, the Intel identification is strongly implied but not confirmed. Intel is not named in any E&R public communication.
Advanced semiconductor packages are densifying: more chiplets, more dies, more heterogeneous integration. This creates two equipment needs that E&R addresses:
Surface preparation: Bond interfaces must be atomically clean. Any ionic contamination, organic residue, or native oxide on a copper pad before underfill application causes delamination, electromigration failures, or reduced reliability. Mechanical cleaning cannot reach the geometries involved. Plasma treatment — bombarding the surface with ionized gas at controlled energy levels — removes contamination without physical contact, down to sub-nanometer depths.
Precision material removal: At advanced packaging geometries, mechanical dicing blades create micro-cracks, debris, and heat damage. Laser scribing and drilling avoids these problems by ablating material with photons rather than physical abrasion. For glass specifically — brittle, hard to cut mechanically without fracture — femtosecond laser modification is the only viable approach at production accuracy.
Plasma treatment works by driving a gas (argon, nitrogen, O₂, forming gas) into a plasma state via RF power. Ions and radicals in the plasma bombard the substrate surface, breaking chemical bonds in organic contaminants (C-H, C-O, C-N bonds) and reducing native oxides (CuO → Cu). The energy is controlled precisely enough to treat the bonding surface without damaging the underlying chip. The key metrics are: ion energy (eV), flux density (ions/cm²/sec), uniformity (±% across substrate), and process time (seconds to minutes).
Femtosecond laser micromachining works differently from conventional nanosecond lasers. A femtosecond pulse (10⁻¹⁵ seconds) delivers energy before heat can diffuse into the surrounding material (thermal diffusion time in glass is ~picoseconds to nanoseconds). This enables cold ablation — material is removed without a heat-affected zone, no cracks, no recast layer. For TGV, the laser modifies the glass structure along a via path (stealth modification or direct ablation), then wet etching opens the via cleanly. The key metrics are: pulse duration (fs), repetition rate (MHz), wavelength (typically 1030nm or 515nm or 343nm), beam quality (M²), throughput (vias/second), and accuracy (µm).
TGV process flow (E&R’s validated process):
Glass panel/wafer
↓
Laser modification pass
(femtosecond pulses modify glass structure along via path)
↓
Wet etch (HF-based chemistry via Scientech equipment in E-Core Alliance)
(etch selectively removes modified glass, opening vias)
↓
Metallization (sputtering via alliance partners)
(deposit seed layer of Ti/Cu in via walls and top/bottom surfaces)
↓
Electroplating (copper fill of vias)
↓
CMP / planarization
↓
Build-up RDL layers (RDL applied to top and bottom surfaces)
↓
Final glass core substrate ready for EMIB integrationE&R owns the laser modification step. The breakthrough was achieving 8,000 vias/second on fixed layouts — prior art was in the hundreds. This is a throughput inflection that moves TGV from a laboratory curiosity to a production-viable process.
| Metric | Why It Matters | E&R Status |
|---|---|---|
| Via throughput (vias/sec) | Determines cost per substrate; needed to compete with conventional substrate on cost per unit | 8,000/sec (fixed), 600-1,000/sec (random) — validated |
| Via accuracy (µm) | Determines interconnect density achievable; must match RDL pitch tolerances | +/-5 µm — validated |
| Panel size support (mm) | Larger panels = more dies per laser pass = lower cost per die | 700×700mm FOPLP capability confirmed |
| Warpage tolerance (mm) | Glass is brittle; warped panels are common; tool must handle without breakage | 16mm warpage capability confirmed |
| SeWaRe (micro-cracking) incidence | Glass fractures during processing are yield killers | Intel demonstrated zero SeWaRe at NEPCON Japan 2025 |
What it does: Laser processing systems for semiconductor package manufacturing. Three sub-families:
Laser Marking: Dot-matrix or barcode marking on package surfaces for traceability. Applied to FCBGA top surfaces, wafer backside (wafer ID), and flip-chip packages. ASP range: TWD 1-5M per unit (estimated). High unit volumes to OSATs. Mature, established market — growth tracks OSAT capex.
Laser Micromachining (Scribing/Grooving/Dicing): Precision material removal for singulation, cavity formation, and de-layering. Applications: FCBGA panel singulation, ABF substrate scribing, epoxy mold compound grooving, SiC device dicing. Uses pulsed Nd:YAG or femtosecond lasers depending on material. ASP range: TWD 5-20M per unit. Mid-range, growing with FCBGA volumes.
TGV Drilling / Laser Modification: The strategic product. Uses femtosecond laser (Coherent Monaco or equivalent) to modify glass for TGV formation. This is the new, high-value application. ASP not disclosed; likely TWD 20-50M+ per system given complexity. Currently in early production ramp.
Customers: OSATs (ASE, Amkor, JCET, Powertech) and IDMs (implied Intel for TGV; potentially others for marking/scribing).
Competitive alternatives: DISCO (6146.T) for dicing applications; 3D-Micromac (private, Germany) for laser micromachining and TGV; Delphi Laser (private, China) for TGV systems; EKSPLA (private, Lithuania) for femtosecond laser systems (not OEM).
Why customers choose E&R: Price (20-40% below DISCO on equivalent applications); co-development history for TGV (only validated supplier in market at production accuracy); Taiwan-based with U.S. service capacity (CHIPS Act / supply chain resilience angle).
What it does: Vacuum plasma processing chambers for semiconductor packaging. Four application areas:
ASP range: TWD 3-12M per unit depending on chamber size and configuration. Attach closely to FCBGA/FCCSP production volumes; plasma tools are standard process steps at most OSATs.
Customers: Same OSAT base. Plasma tools are higher-volume, lower-ASP than laser systems. This segment provides revenue stability; laser provides growth optionality.
Competitive alternatives: Diener Electronic, PVA TePla, March Instruments, SAMCO. Plasma is a more commoditized market than precision laser; pricing competition is real.
FPC processing equipment: Flexible printed circuit (FPC) laser cutting and drilling systems. Serves the mobile electronics supply chain. This is a legacy product area with modest growth.
UniDRON: E&R’s brand for lab/small-production units. Likely serves process development customers and university/research labs. Low ASP, low volume.
Tape-and-reel: Tape-and-reel systems for packaging discrete components. Commoditized; not a growth driver.
Raw Materials Components/Subsystems Equipment/Tools Manufacturers End Products
(glass, gases, → (lasers, stages, optics) → (E&R machines ★) → (OSATs, IDMs) → (AI processors,
chemicals) Coherent (COHR), E&R Engineering ASE, Amkor, FPGAs, CPUs,
HIWIN (1515.TW), (8027.TWO) Intel Foundry, GPUs)
Keyence (6861.T) SamsungE&R sits at the equipment layer. It sources components from technology companies and sells to manufacturers. This is an intermediate position — lower pricing power than component/IP companies (Coherent, HIWIN) but lower cyclicality than pure-play chip manufacturers.
Revenue pool at this layer: The advanced packaging equipment market is ~$8-11B globally. E&R’s TWD 1.8B revenue (~$55M) represents less than 1% of the market — substantial room to grow without displacing any major player.
Barriers to entry at E&R’s layer: Customer qualification cycles (18-24 months), process recipe IP, co-development relationships, field service network. The TGV development specifically creates a moat that took 5 years and is not replicable quickly.
Pricing power: Limited in plasma (commoditized); moderate in laser marking/scribing (Taiwan cost advantage vs. Japan); potentially strong in TGV (sole-source or near-sole-source for validated glass substrate laser processing).
| Supplier | Ticker | Layer | Bypass-ability | Supplier MC vs E&R | Note |
|---|---|---|---|---|---|
| Coherent | COHR | Laser light sources | Partial — alternative femtosecond laser vendors exist (EKSPLA, Amplitude, IPG) but switching requires re-qualification | COHR ~$5.5B vs E&R ~$503M (~11×) | E&R success story on Coherent’s website suggests deep relationship |
| HIWIN | 1515.TW | Precision motion | Partial — NSK, THK are alternatives; HIWIN preferred for Taiwan-native supply | HIWIN ~$2.5B vs E&R ~11×+ | Taiwan-based; logistics advantage |
| Keyence | 6861.T | Sensing/metrology | Partial — Cognex, SICK alternatives | Keyence ~$30B+ (50×+) | Not a concentration risk; replaceable |
Bottleneck verdict: Coherent is the closest upstream bottleneck — E&R’s key application (femtosecond TGV) requires best-in-class femtosecond laser sources, and Coherent’s Monaco is cited explicitly in E&R press materials. If femtosecond laser supply tightens (possible given multiple semiconductor companies accelerating glass substrate programs), Coherent benefits disproportionately. COHR is already on the radar as a much larger and more diversified company, however, so this is not an under-priced secondary long.
| # | Customer | Ticker | Est. Rev. Share | Relationship Type | Details |
|---|---|---|---|---|---|
| 1 | Unnamed North American IDM (implied Intel) | INTC | Unknown — likely significant for TGV/AIS | Co-development + production equipment supply | 5-year TGV program; Hillsboro site proximity; AIS orders through 2027 |
| 2 | Major OSATs (ASE, Amkor, JCET, Powertech) | ASX / AMKR | Collectively meaningful; each likely <20% | Capital equipment supply — laser marking, plasma, scribing | Standard tools in FCBGA/FCCSP process flow |
| 3 | Additional IDMs for advanced evaluation | Various | Small | Development/evaluation equipment | Entry point for potential production wins |
Customer 1 — Implied Intel: Intel (INTC) is the only North American IDM with: (a) an active EMIB glass substrate program; (b) TGV drilling as a production requirement; (c) campus in Hillsboro, OR where E&R just opened a site. Intel’s EMIB-T is expected to ramp 2H26; external customer EMIB commitments are “prepaying for production” and expected to exceed $1B annually. If E&R supplies the laser modification tool for Intel’s glass substrate line, this would be a material revenue inflection. Intel’s financial health is under pressure (operating losses in foundry segment) but its EMIB/advanced packaging business is specifically the area of strategic priority and investment.
Customer 2 — ASE / Amkor: Both are public companies (ASX, AMKR) undergoing significant FCBGA capacity expansion for AI/HPC. E&R’s plasma tools are standard process steps in FCBGA lines; laser marking tools are ubiquitous. These relationships are steady-state, not transformational — they track OSAT capex cycles.
Concentration Risk: Cannot be assessed without segment disclosure. Qualitative judgment: if Intel represents >30% of revenue on TGV/AIS, that is a high concentration risk in an Intel-specific technology. If the customer is merely one OSAT among several for standard tools, the risk is moderate. The market appears to be pricing as if Intel is a dominant customer; the company’s financial disclosures do not confirm or deny this.
Switching costs: For TGV tools: High. Switching requires qualifying a new machine supplier from scratch on a process that took E&R 5 years to validate. DISCO or 3D-Micromac would need 2-3 years minimum to achieve the same qualification status. For plasma/marking tools: Low to moderate. Standard tools with established alternatives.
The problem E&R solves: Semiconductor packages are becoming more complex faster than the tools to build them. The transition to glass core substrates (required for EMIB 3.5D and beyond) demands TGV drilling at production accuracy — a capability that did not exist at commercial scale before E&R’s 2024 validation. Without TGV-capable tools, Intel’s glass substrate program stalls. E&R is one of the critical path items for a $100B+ industry transition.
End-use applications:
| Application | E&R Tools | End-Market Driver |
|---|---|---|
| FCBGA packaging (CPUs, GPUs, FPGAs) | Plasma cleaning, laser marking | AI/HPC server buildout; every data center GPU is FCBGA |
| FOPLP (Fan-Out Panel Level) | Laser marking/cutting/descum, plasma cleaning | AI ASICs, RF front-ends, edge AI |
| Glass substrate (EMIB 3.5D+) | TGV laser modification, plasma pre-treatment | Intel EMIB-T; future Samsung/other adoption |
| SiC power devices | Laser scribing, plasma treatment | EV drivetrains; renewable energy inverters |
| CPO (Co-Packaged Optics) | Plasma surface prep for photonic die bonding | AI datacenter switch ASICs with integrated optics |
TAM (advanced packaging equipment):
| Segment | 2025 Market | Growth Rate | E&R Addressable |
|---|---|---|---|
| IC advanced packaging equipment (total) | ~$10.7B | 13.6% CAGR to 2034 | Partial |
| Laser dicing/scribing (sub-set) | ~$0.9B | 6-10% CAGR | Core |
| Plasma dicing/treatment (sub-set) | ~$0.1B | 6.5% CAGR | Core |
| Glass substrate TGV equipment (new market) | Near-zero today | Pre-production; expected mass production 2027-28 | Potentially E&R’s most valuable segment |
E&R’s served portion of existing markets: ~$1-2B SAM (laser marking/scribing + plasma treatment at back-end). Current market share: ~3-5% of SAM (est. ~$55M revenue). Glass substrate TGV equipment: effectively first-mover in a market that is forming now.
Secular tailwinds: 1. Heterogeneous integration (chiplets) — every node in the packaging hierarchy becomes more complex, expanding tool intensity per wafer 2. EMIB + glass substrate ramp — Intel’s program, with Samsung and LG Innotek following (Samsung 2028 target), creates a new TAM for TGV equipment 3. AI infrastructure buildout — FCBGA demand for Nvidia H/B-series, AMD MI-series, custom ASICs at hyperscalers 4. SiC volume growth — EV adoption requires laser dicing of SiC wafers at scale 5. CPO inflection — datacenter optics integration requires new surface prep steps
Demand inflection: Advanced packaging equipment spending is entering a multi-year upcycle driven by AI infrastructure. The AI chip buildout requires unprecedented packaging sophistication — HBM stacks, FCBGA at >100mm die sizes, EMIB bridges. OSAT capex at ASE and Amkor is recovering from the 2023 trough. Intel’s EMIB program is adding an entirely new demand layer as it moves into external foundry services. Demand is accelerating, not stabilizing.
Supply constraint: The TGV laser processing tool supply is effectively just E&R and one or two smaller companies (3D-Micromac, Delphi Laser) — none of which have completed a multi-year production validation with an IDM. This is a genuine early-stage bottleneck. For conventional plasma/marking tools, supply is adequate with several qualified vendors.
Inventory cycle: Taiwan OSAT capex was in a trough through 2023-early 2024. Q4 2025 at E&R showed operating profit for the first time in five quarters (TWD 66M vs. losses in Q1-Q3 2025) — this is consistent with a turn in the order book.
What has fundamentally changed in the last 12 months: 1. TGV validation completed (2024): E&R crossed from R&D partner to production equipment supplier. This is an irreversible step — it takes years to replicate. The market is now pricing in the option, but the contract hasn’t fully landed in revenue yet. 2. Intel EMIB outsourcing to Amkor (Dec 2025): Intel has expanded EMIB packaging beyond its own fabs to Amkor (Songdo K5, Portugal, Arizona). This actually expands E&R’s addressable market — any EMIB line needs substrate preparation tools. 3. EMIB-T ramp timeline firmed up: External customers are “prepaying for production” of EMIB capacity. Apple eyeing Intel 18A for 2027 M-chips. MediaTek already using EMIB-T for large-die configurations. The EMIB market is no longer an Intel-internal program — it is a foundry service. 4. AIS (Automation Integration Service) product line launched: This is a new revenue model — design-to-implementation multi-module contracts vs. one tool at a time. The AIS backlog into 2027 is the most concrete recent signal of order momentum.
What consensus is missing: The sell-side has one analyst on E&R (PT TWD 98), clearly set before the recent developments. The market consensus for a loss-making Taiwan OTC micro-cap is minimal. The significant North American AIS orders, the Intel EMIB-T timeline firming up, and the Q4 2025 profitability turn are not reflected in any meaningful analyst coverage.
Near-term (0-12 months): - Q1 2026 earnings (May 11, 2026): First quarter with Hillsboro site fully operational; confirm whether AIS revenue is being recognized; test whether Q4 2025 profitability was a one-off or a trend - Monthly MOPS Taiwan revenue data: E&R files monthly revenue with the exchange; tracking this gives real-time order conversion visibility - Intel EMIB-T production ramp (2H26): If Intel begins EMIB-T production at scale, E&R’s TGV tools should be in the production flow
Medium-term (1-3 years): - Qiaotou plant operating license (~2026): Adds manufacturing capacity specifically for TGV/glass substrate tools; enables E&R to handle larger order volumes - Glass substrate mass production (2027-2028): LG Innotek targets 2027-28; Samsung targets 2028; Intel’s pilot line already running. If any major program reaches production scale, E&R’s TGV tool is the critical path item - EMIB-T external customer ramp: Apple (2027 M-chip), MediaTek (now), Google/Amazon (exploring) — all potential packaging customers for Intel’s foundry, all requiring glass substrate-capable tools at scale
Leading indicators to watch: - E&R monthly MOPS revenue (TWSE filing, monthly) — the single best real-time indicator - Intel quarterly EMIB packaging revenue disclosures - OSAT capacity utilization rates (ASE, Amkor quarterly calls) - Any E&R press releases about production equipment deliveries (vs. development) - Qiaotou operating license announcement
E&R is at a rare inflection: after five years of co-development burn, it has validated a technology (production-accurate TGV) at the exact moment the customer (Intel) is ramping the product (EMIB-T) that requires it. The AIS integration service model extends the revenue relationship beyond one-off equipment sales. Q4 2025 showed the first operating profit in five quarters, suggesting the order conversion is beginning. The stock has already priced in optionality (9× EV/Sales), but the specific catalysts — May 11 earnings, EMIB-T ramp in 2H26, Qiaotou operating license — give the thesis concrete near-term checkpoints rather than open-ended technology optionality.
| Name | Title | Tenure | Background |
|---|---|---|---|
| Eric Chang | Group President | Long-tenured (founder era) | Attended Qiaotou groundbreaking May 2024; identified as President/IR; described as “Group President” suggesting holding company structure; clearly the strategic decision-maker |
| Kevin Chang | Overseas Sales & Service Director | Current | Manages international customer relationships and field service network; based in Kaohsiung |
| Lin Allen | Marketing Director | Current | Taichung-based; product marketing |
| Chao Vic | COO | Current | Kaohsiung-based; manufacturing operations |
Data quality caveat: E&R’s management team is lightly disclosed in English-language sources. The MOPS 年報 (annual report) would contain full executive compensation disclosures, board structure, and insider transaction history. This section reflects what is publicly available in English.
Founder-led dynamics: Eric Chang appears to be founder-era management. This is standard for Taiwan OTC industrial companies. Founder-led companies in semiconductor equipment often have: deep customer relationships embedded in the founder’s network (positive for revenue stickiness; negative for key-person risk), conservative financial management with preference for organic growth (positive: no empire-building M&A; negative: may under-invest in sales/marketing), and long product development cycles (positive: TGV’s 5-year development reflects patience; negative: investors may get impatient).
Data gap: No English-language source quantifies insider holdings. Given Eric Chang’s apparent founder status, his ownership is likely significant (10-30% is typical for Taiwan founder-led OTC companies). This should be verified from MOPS filings.
Recent insider activity: Not available from English sources. MOPS Form 4 equivalents would need to be reviewed.
Shell & Cross-Holdings Red Flag Scan:
No red flags identified in available English sources. E&R is a single-product industrial company in semiconductor equipment — the business is relatively transparent. The E-Core Alliance is a non-equity consortium, not a shell structure. No related-party transaction patterns identified. This section would need to be revisited with MOPS annual report data.
| Item | Assessment |
|---|---|
| M&A | No M&A history identified; organic-only growth |
| Buybacks | None identified; current losses make buybacks inappropriate |
| Capex | TWD 495M (FY2024) + TWD 328M (FY2025) = ~TWD 823M over 2 years for Qiaotou plant + Nantong plant. Represents ~50% of total assets invested in ~2 years. This is aggressive for a company in losses |
| Dilution | No evidence of significant equity issuance; share count stable at ~105M |
| Dividends | None in FY2024-25 (losses); were paid in profitable years (FY2021-22) |
Capital allocation grade: C+. The Qiaotou capex bet is the right strategic move if the glass substrate market materializes, but the timing — investing heavily during an operating loss cycle — reduces financial flexibility. No M&A history is a positive (no expensive empire-building). Stable share count is a positive.
Not available from English sources. Would require MOPS 年報 CD&A equivalent.
Board composition not available from English sources. Taiwan TWO-listed companies have mandatory independent director requirements and audit committee structures, but the specific names and backgrounds are not in available public English sources. MOPS 年報 should be pulled.
Governance flags: - Company structure appears standard for Taiwan founder-led industrial; no dual-class shares identified - “Group President” title suggests Eric Chang may have oversight of related entities (E&R USA LLC at minimum) - E-Core Alliance is a strategic consortium, not a financial entity — governance concern level is low
| Dimension | Rating | Key Finding |
|---|---|---|
| Skin in the Game | Yellow | Likely significant founder holding, but unquantified — need MOPS filing |
| Holdings Concentration | Yellow | Data gap — MOPS required |
| Shell / Cross-Holdings | Green | No patterns identified in available sources |
| Capital Allocation | Yellow | Aggressive capex in loss years; strategic rationale is sound but financial risk is real |
| Compensation Alignment | Unknown | Data gap |
| Governance Quality | Yellow | Standard Taiwan OTC structure; founder-dominated; limited disclosure in English |
| Litigation / Enforcement | Green | No litigation identified |
| Overall Management Grade | C+ / Yellow | Founder-led with strategic vision; governance disclosure is opaque from available sources; no red flags but also limited verification |
| Company | Ticker | Segment Overlap | Revenue | Market Share (dicing/scribing) | Moat Type |
|---|---|---|---|---|---|
| DISCO Corporation | 6146.T | Laser dicing, grinding, thinning | ~$2.8B (FY2025e) | 20%+ in dicing (dominant) | Brand + installed base + precision manufacturing |
| Accretech (Tokyo Seimitsu) | 7729.T | Plasma dicing, wafer probing | ~$500M semi division | Meaningful in plasma dicing | Precision engineering + Japan quality brand |
| 3D-Micromac | Private (Germany) | Laser micromachining, TGV | ~€50M est. | Small; specialist | Technology focus + European customer base |
| Delphi Laser | Private (China) | TGV laser systems (newer entrant) | <$30M est. | Emerging | China market; price competition |
| Coherent (selected products) | COHR | Some OEM laser system competition | ~$4.7B total (lasers + other) | Small in systems | Vertical integration into laser subsystems |
E&R’s competitive position: - Versus DISCO: Price advantage (20-40% cheaper) + Taiwan supply chain + AIS integration model. Weakness: scale, brand, global service depth. - Versus 3D-Micromac: E&R has a validated production customer (implied Intel); 3D-Micromac has European customers but no confirmed IDM production validation of same scale. E&R has the AIS integration model. - Versus Delphi Laser: China-market player; E&R is not competing in China for the strategic TGV business.
Competitive moat (E&R-specific): 1. TGV co-development IP: 5 years of joint development with an IDM customer created process recipes, machine parameters, and reliability data that cannot be rapidly replicated. This is the strongest moat element. 2. Taiwan cost-structure: Against DISCO and Accretech, E&R prices 20-40% below for equivalent applications. For customers prioritizing cost (all of them), this matters. 3. AIS integration model: Selling a complete process module (laser + plasma + handling + integration) rather than a single tool increases switching costs and revenue per customer engagement. 4. Geographic positioning: Hillsboro site adjacent to Intel campus; Taiwan HQ close to ASE/Amkor. This proximity reduces time-to-response for field issues, which matters for production tools.
Moat weaknesses: 1. No protected IP disclosed (no patents mentioned in public materials) 2. Small scale limits ability to invest in customer support at DISCO’s level 3. Single dominant customer concentration (implied) creates vulnerability
5-year lock-up test: Would I happily own this business for 5 years without trading? Uncomfortable but possibly yes — the TGV technology is validated, the market is forming, and the company has no debt crisis. But “glass substrate ramp in 18-24 months” has been said since 2022. If the glass substrate timeline slips again, the business earns losses for another 2 years while burning remaining cash. This is a development-stage bet in an industrial company’s clothing.
Unique economic engine: E&R’s economic engine is: co-development relationships that become production equipment sales, with AIS integration expanding per-customer revenue. The source of uniqueness is the 5-year TGV validation — a durable relationship asset that took long to build. Durability: medium — it holds as long as the customer (Intel) is active in glass substrates. If Intel exits glass substrates (as it nearly did in 2022-23), the engine loses its main fuel.
Blank-check disruptor test: A well-funded competitor (DISCO, Applied Materials, Lam Research) could allocate $200M and 3 years to replicate E&R’s TGV qualification. Applied Materials (AMAT) already has deep plasma chemistry expertise and could enter the TGV drilling market with a femtosecond laser division. This is a real threat over a 3-5 year horizon. The question is whether E&R can generate enough revenue from TGV in the next 2-3 years to build an installed base that makes switching costs prohibitive.
Quality verdict: Vulnerable-to-Durable. The business is not high-quality by standard metrics (currently losing money, limited pricing power in core markets). The TGV moat is real but narrow and time-limited. Quality improves significantly if TGV production scale is achieved before a better-capitalized competitor qualifies.
Industry structure: Advanced packaging equipment is moderately consolidated at the top (DISCO, AMAT, Lam dominate by segment) but fragmented in the niche tool categories where E&R operates. The laser-for-packaging niche has 5-10 meaningful global players; the TGV-specific niche has 2-3.
Barriers to entry: Customer qualification cycles (18-24+ months at a major OSAT or IDM), process recipe IP, field service network, and capital equipment manufacturing know-how. High for new entrants; moderate for existing semicap players wanting to expand.
Cyclicality: Semiconductor equipment is highly cyclical, tracking OSAT and IDM capex budgets. E&R’s revenue collapse in FY2023 (-52% YoY) was a direct consequence of the post-COVID semicap correction. The company had peak revenue of TWD 3.2B in FY2022 and has not recovered to that level. However, the current recovery is against a trough baseline, and the new TGV/AIS revenue stream is counter-cyclical to traditional tools in the sense that it is tied to a structural transition (glass substrate adoption) rather than a capacity cycle.
Current cycle position: Recovering from trough. OSAT capex is re-accelerating for AI/HPC; Intel’s EMIB packaging investment is counter-cyclical to its foundry losses. E&R’s Q4 2025 operating profit (TWD 66M) after five quarters of losses is the clearest signal of cycle turn.
Leading indicators for E&R specifically: - OSAT capacity utilization (published quarterly by ASE, Amkor) - Intel EMIB packaging revenue (Intel quarterly reports) - E&R monthly MOPS revenue filings (available on Taiwan exchange website)
Organic revenue growth: Revenue in FY2025 was TWD 1.81B (+10% vs. FY2024’s TWD 1.64B). Q4 2025 alone was TWD 721M — nearly equal to H1 FY2025 combined. Growth is accelerating in Q4, suggesting AIS orders are beginning to convert to revenue. FY2023 (-52%) and FY2024 (+6%) were the bottom; the Q4 2025 surge is the first real evidence of inflection.
Margins: Gross margins are stable at 33-40%; this range is typical for Taiwan industrial equipment. Operating margin has been negative since FY2023. The Q4 2025 operating profit of TWD 66M (9.2% operating margin) is the first positive operating quarter in over a year. If Q4’s run-rate sustains, the annual operating income picture changes substantially.
Capital intensity: Capex-heavy during the plant expansion period (TWD 495M FY2024, TWD 328M FY2025 = TWD 823M over 2 years on a TWD ~4B asset base). Once Qiaotou is operational (operating license ~2026), capex should normalize to maintenance levels (TWD 150-200M range). FCF should turn positive by FY2026-27 if revenue continues recovering.
Capital deployment: No M&A, no buybacks, no dividends (suspended during loss years). Capital is going entirely into Qiaotou plant + Nantong plant + North American service expansion. This is the right strategic deployment for the opportunity, but it requires the TGV/AIS revenue to materialize.
| Q2 2024 | Q3 2024 | Q4 2024 | Q1 2025 | Q2 2025 | Q3 2025 | Q4 2025 | |
|---|---|---|---|---|---|---|---|
| Revenue (M TWD) | 503.9 | 422.5 | 430.7 | 314.9 | 374.5 | 398.5 | 721.4 |
| Revenue QoQ Δ | — | -81.4 | +8.2 | -115.8 | +59.6 | +24.0 | +322.9 |
| Revenue YoY % | — | — | — | +9.4% | -25.7% | -5.7% | +67.4% |
| Op. Income (M TWD) | +22.3 | -39.5 | -29.4 | -75.5 | -44.4 | -23.7 | +66.2 |
Second-derivative assessment: - Revenue growth rate is strongly positive in Q4 2025 (+67.4% YoY). This is the inflection quarter — the second derivative went from negative (Q2 2025: -25.7%) to strongly positive in one step. - Operating income improved across every quarter from Q1 2025 (-75.5M) through Q4 2025 (+66.2M) — a clear, consistent trend. - Q4 2025 alone represents ~40% of FY2025’s total revenue. This suggests an order bunching effect (AIS orders recognized in one quarter) or a structural step-change in order rate. - The exit rate from Q4 2025 annualized would imply FY2026 revenue of ~TWD 2.9B — a 60% step up from FY2025. That is a significant implied expectation embedded in the Q4 data. Whether it sustains depends on whether AIS orders continue to convert.
| Metric | Value |
|---|---|
| Market cap | TWD 16.44B (~USD 503M) |
| Enterprise value | TWD ~16.72B |
| P/E (TTM) | N/A (net loss FY2025: -97.8M) |
| EV/EBITDA | N/A (operating losses; EBITDA not disclosed; est. EBITDA near break-even including D&A) |
| EV/Sales (FY2025) | 9.2× |
| EV/Sales (Q4 annualized) | ~5.8× |
| FCF yield | Negative |
| Dividend yield | None |
| 52-week range | TWD 70.60 – 163.00 |
| Forward P/E (1 analyst) | 233.6× (2026E EPS ~0.67 TWD) |
At 9× EV/Sales on a TWD 1.8B revenue business losing money, E&R is priced as a growth option, not an industrial company. The only way to make 9× EV/Sales defensible is if FY2026 or FY2027 revenue accelerates materially (toward TWD 3B+) as AIS and TGV production orders convert. At Q4 2025’s annualized run-rate (~TWD 2.9B), EV/Sales drops to ~5.8× — still elevated for an industrial, but less extreme.
For comparison: DISCO trades at ~5-6× EV/Sales with positive operating margins and market leadership. E&R at 9× EV/Sales with negative operating margins is pricing in a significant growth step-up. This is justified only if the glass substrate / EMIB thesis materializes on the 2026-27 timeline.
| Metric | FY2022 | FY2023 | FY2024 | FY2025 | FY+1E (FY2026) |
|---|---|---|---|---|---|
| Revenue (M TWD) | 3,223 | 1,549 | 1,645 | 1,809 | N/A (1 analyst; likely stale) |
| Revenue growth | +26.6% | -51.9% | +6.2% | +10.0% | est. +30-60% if AIS sustains |
| Gross profit (M TWD) | 1,124 | 614 | 598 | 637 | — |
| Gross margin | 34.9% | 39.7% | 36.3% | 35.2% | — |
| EBIT (M TWD) | 404 | 17 | -119 | -77 | — |
| EBIT margin | 12.5% | 1.1% | -7.2% | -4.3% | — |
| Net income (M TWD) | 391 | 31 | -51 | -98 | — |
| EPS (TWD) | 3.94 | 0.32 | -0.51 | -0.93 | ~0.67 (1 analyst) |
| Metric | FY2022 | FY2023 | FY2024 | FY2025 |
|---|---|---|---|---|
| Op. cash flow (M TWD) | -234 | 246 | 355 | 137 |
| Capex (M TWD) | -184 | -158 | -495 | -328 |
| FCF (M TWD) | -418 | 89 | -140 | -191 |
| FCF margin | -13.0% | 5.7% | -8.5% | -10.6% |
| Cash (M TWD) | 1,065 | 846 | 730 | 496 |
| Net cash/(debt) (M TWD) | 566 | 218 | 248 | -279 |
Cash runway: TWD 496M cash, TWD 949M total debt, net debt -TWD 279M. If capex normalizes in FY2026 (to ~TWD 150-200M) and operating cash flow tracks Q4 2025’s trajectory (positive), FCF should turn positive. If the revenue inflection does not sustain, cash of TWD 496M buys 2-3 years at current burn rates.
| Q3 2024 | Q4 2024 | Q1 2025 | Q2 2025 | Q3 2025 | Q4 2025 | 4Q Avg | |
|---|---|---|---|---|---|---|---|
| Revenue YoY Δ (M TWD) | — | — | +27.1 | -129.4 | -24.0 | +290.7 | +41.1 |
| GP YoY Δ (M TWD) | — | — | +6.3 | -62.6 | +3.4 | +92.7 | +10.0 |
| Incr. Gross Margin | — | — | 23.2% | 48.4% | n/m | 31.9% | — |
| EBIT YoY Δ (M TWD) | — | — | -3.2 | -66.7 | +15.8 | +95.6 | — |
| Incr. EBIT Margin | — | — | n/m | 51.5% | n/m | 32.9% | — |
What the incrementals tell us: - Q4 2025 incremental gross margin of 31.9% (on +TWD 291M incremental revenue) is consistent with the historical 33-35% reported gross margin — new revenue is being added at normal margins. There is no margin sacrifice to win orders. - Q4 2025 incremental EBIT margin of 32.9% is exceptional — this means nearly all gross profit drop-through is reaching operating income, implying operating leverage is very high on incremental AIS/TGV revenue. - This pattern is consistent with what happens when a fixed-cost equipment maker wins large-order contracts that run through an existing factory: the incremental revenue accrues almost entirely to operating profit because SG&A and R&D are fixed. - Q2 2025 was distorted by a tax or one-time charge (net loss of -118.9M on operating loss of only -44.4M — the gap is unusually large).
Peer multiples (2025-2026):
| Company | Ticker | EV/Sales | EV/EBITDA | Notes |
|---|---|---|---|---|
| DISCO | 6146.T | ~5-6× | ~15-18× | Dominant; positive margins; Japan OEM |
| Accretech | 7729.T | ~2-3× | ~10-12× | Smaller scale; positive margins |
| BESI | BESI | ~8-10× | ~20-25× | Advanced packaging pure-play; growth premium |
| Applied Materials | AMAT | ~4-5× | ~14-16× | Diversified semicap leader |
| E&R Engineering | 8027.TWO | 9.2× | N/A (losses) | TGV/glass substrate optionality |
Implied expectations at current price: At TWD 156.50 and EV ~TWD 16.7B, the market is pricing ~9× EV/Sales on FY2025 revenue. For E&R to justify this multiple: - If you apply a 5× EV/Sales target (matching DISCO), E&R needs TWD ~3.3B in revenue (~+85% from FY2025) to hold the current price - If FY2026 revenue hits TWD 3B (consistent with Q4 2025 annualized run-rate), EV/Sales drops to ~5.6× — closer to justified on a structural-growth view
Bear case valuation: If AIS orders do not sustain and TGV production is delayed, FY2026 revenue stays at ~TWD 2B. At 4× EV/Sales (trough multiple for a loss-making equipment company), EV = TWD 8B, equity = TWD ~7.7B, price = TWD ~73. That is -53% from current.
Base case valuation: FY2026 revenue of TWD 2.5-3B with positive operating income (~5% margin). At 6× EV/Sales, EV = TWD 15-18B, equity value ~TWD 14.7-17.7B, price TWD ~140-168. Roughly in line with current.
Bull case valuation: FY2026 revenue of TWD 3-3.5B with EMIB-T production tool recognition, operating margin recovering to 8-10%. At 7-8× EV/Sales (justified by TGV sole-source premium), EV = TWD 21-28B, price = TWD ~195-260. Meaningful upside.
| Tailwind | Mechanism | Magnitude | Durability |
|---|---|---|---|
| AI/HPC FCBGA demand | Nvidia/AMD/ASIC GPUs all require FCBGA packaging; E&R’s plasma and marking tools are standard process steps | Significant; tracking data center capex | 3-5yr+ secular |
| Glass substrate / EMIB-T ramp | TGV drilling demand scales with every glass core substrate line installed; E&R is the validated supplier | Potentially transformational if market scales as projected | 5-10yr if glass substrate succeeds |
| CHIPS Act / reshoring | Intel’s domestic EMIB capacity expansion requires locally-supportable equipment; E&R’s Hillsboro presence is directly aligned | Moderate; selective benefit | 3-5yr policy cycle |
| SiC device growth | EV adoption requires SiC wafer dicing; E&R has SiC laser capability | Moderate; follows EV penetration curve | 5-10yr |
| CPO (Co-Packaged Optics) | Datacenter switches with photonic integration require new surface prep steps | Emerging; uncertain timing | 2-5yr |
| Catalyst | Expected Timing | Impact |
|---|---|---|
| Q1 2026 earnings | May 11, 2026 | Critical: confirm or invalidate Q4 2025 inflection |
| Monthly MOPS revenue (monthly) | Every month on Taiwan exchange | Leading indicator of order rate |
| Intel EMIB-T production ramp | 2H 2026 | If Intel confirms EMIB-T volume, E&R’s TGV tools should be in the flow |
| Qiaotou operating license | ~2026 | Removes capacity constraint; reduces need for further capex |
| Catalyst | Timeline | Potential Impact |
|---|---|---|
| Glass substrate mass production (LG Innotek, Samsung, Intel) | 2027-2028 | New TAM opening; E&R is first-mover with validated tool |
| AIS backlog conversion | FY2026-27 | Already secured; question is recognition rate |
| Additional EMIB outsourcing (Amkor expansion) | 2026-27 | More EMIB production lines = more surface prep tool demand |
AIS North American Orders (unnamed customer, implied Intel-linked): - What it is: Design-to-implementation integrated multi-module process systems (laser + plasma + handling, integrated into a turnkey production module) - Counterparty: Unnamed; described as “North American” in E&R’s ISIG 2026 announcement - Value: “Significant orders” — no specific value disclosed - Status: Already secured; described as providing visibility into 2027 - Revenue impact: Q4 2025 revenue surge (TWD 721M) may represent initial AIS order recognition; further deliveries expected through 2027
TGV Production Equipment (unnamed North American IDM, implied Intel): - What it is: Production-scale laser modification tools for TGV formation in glass core substrates - Counterparty: Unnamed IDM; process validated 2024 - Status: Validation complete; production order status unknown — this is the key unknown - Revenue impact: If production orders placed, each TGV tool likely TWD 20-50M+ ASP, and multiple tools needed per production line
| Risk | Likelihood | Existing Mitigants | Mgmt De-risk Plan | Can It Be Closed? |
|---|---|---|---|---|
| Intel strategy shift / glass substrate delay | Medium | TGV validation done; Intel EMIB-T timeline firmed up; Intel is now outsourcing EMIB, widening the addressable customer base | E-Core Alliance broadens beyond Intel dependency; AIS serves multiple customers | Partially — Intel dependency is structural; reduces if multi-customer production wins are achieved |
| Q4 2025 revenue was one-off, not sustained | Medium | Monthly MOPS data will clarify by May 2026; AIS order backlog through 2027 | Hillsboro site expansion supports order fulfillment | Yes — closes at Q1 2026 earnings (May 11) |
| Cash burn / capital needs | Medium-Low | TWD 496M cash; Qiaotou capex should taper post-license; FCF improving Q4 | Capex normalization post-Qiaotou; AIS revenue conversion | Yes — closes if FCF turns positive, which Q4 trajectory implies for H2 2026 |
| Competitor qualification (DISCO, 3D-Micromac, AMAT) | Low-Medium in near term; Medium over 3-5yr | 5-year co-development moat is time-consuming to replicate; AIS installed base creates switching costs | Scale-up quickly using Qiaotou capacity; lock in multi-year service agreements | Cannot be closed — permanent competitive risk; can be managed via speed and service relationships |
| OSAT capex correction (semicap cycle) | Low-Medium (currently in recovery) | E&R already in trough; AIS/TGV are counter-cyclical to the OSAT capex cycle | Diversify across OSATs and IDMs; AIS creates multi-year backlog | No — structural cyclicality; can only be diversified away at revenue mix level |
| Key-person risk (Eric Chang) | Medium | Founder has deep customer relationships; management team has operational depth (Kevin Chang, Chao Vic) | Not disclosed | Cannot be eliminated — management succession plans not public |
| Intel foundry execution failure | Low-Medium | Intel 18A is in production (Panther Lake shipping); EMIB-M already mass production; EMIB-T timeline firmed up | E&R serves the packaging layer, not the wafer process — partially insulated | No — Intel’s 18A yield trajectory affects EMIB adoption volume |
What would make me wrong: - Q1 2026 earnings (May 11) shows Q4 2025 was a one-time order recognition, not a trend - Intel announces a pause or restructuring of its glass substrate program (not impossible — Intel nearly exited glass substrates in 2022-23 before recommitting) - EMIB-T ramp is pushed to 2028+ as external customer adoption is slower than expected - 3D-Micromac or another competitor completes a production qualification with a major IDM before E&R can lock in multi-year supply relationships
Downside price: If the thesis breaks, EV/Sales falls to 4× on TWD 2B revenue = EV ~TWD 8B, equity ~TWD 7.7B, price ~TWD 73 (-53%). If the thesis collapses (Intel exits glass substrate, AIS orders dry up), trough revenue could return toward FY2023 levels (TWD 1.5B), implying price of TWD 40-50.
Data limitation: Taiwan OTC (TWO-listed) companies do not file 13F/13D/13G equivalents that are available in English. MOPS filings contain shareholder registers, but accessing these requires reading Chinese-language filings. The ownership section reflects what is available from English sources.
| Item | Data |
|---|---|
| Total shares | 105.06M |
| Implied founder (Eric Chang) holding | Unknown — likely significant; estimate 10-30% based on Taiwan founder-led OTC norms |
| Institutional holders | Not quantifiable from available English sources |
| Short interest | Not disclosed (Taiwan OTC) |
Analyst sentiment:
| Item | Data |
|---|---|
| Number of analysts | 1 |
| Consensus | Neutral |
| Price target | TWD 98 (single analyst) |
| Implied return from current price | -37% |
This single-analyst PT at TWD 98 was almost certainly set before the Q4 2025 revenue inflection and the ISIG 2026 AIS announcement. At current TWD 156.50, the stock is 60% above the lone analyst’s target. This is not a red flag in itself — it reflects the thin coverage universe for Taiwan OTC micro-caps. The information edge for investors willing to do primary research is significant.
Conviction level: Medium. The thesis is coherent and the technical foundation is real, but: - The Q4 2025 inflection needs to be confirmed at Q1 2026 earnings - The Intel dependency cannot be quantified from public sources - The valuation (9× EV/Sales) requires a major revenue step-up to be justified
Entry strategy: Scale in — do not take a full position before Q1 2026 earnings (May 11, 2026). A starter position to establish a cost basis before the catalyst, with a meaningful add if Q1 2026 confirms the Q4 2025 trend.
Stop-loss / re-evaluation triggers: - Q1 2026 revenue below TWD 400M (below Q3 2025 level — would suggest Q4 was a one-off) - Any public announcement by Intel of a pause in glass substrate programs - Cash falling below TWD 300M without corresponding revenue inflection - Any major MOPS governance disclosure (undisclosed related-party transactions, executive departures)
What would cause adding: - Q1 2026 revenue > TWD 500M with positive operating income (confirming Q4 trend) - Intel public naming of E&R or confirmation of production TGV orders - Additional large AIS order announcement - Glass substrate mass production timeline confirmation by Samsung or LG Innotek
Data as of April 26, 2026. Financial figures in TWD unless noted. FY = calendar year.