DEEP RESEARCH · Protec (053610.KQ) · HBM / LAB

Protec's 4Q25 Earnings Surprise and the Structural Re-rating Begins

The Laser Assisted Bonder (LAB) growth engine layered on top of the dispenser cash cow — operating leverage driven by HBM and advanced packaging.

Date: 2026-01-30 · Personal research note · Source: Naver blog original

0. Bottom line first

Official fact: 2026-01-30 disclosure ("change in revenue/profit structure of 30%+"): FY2025 consolidated revenue KRW 230.19bn (+35.2% YoY), operating profit KRW 46.25bn (+244.0% YoY), net income KRW 35.30bn (+176.9% YoY). The company cited "strong demand for high-end semiconductors."

Interpretation: 36.6% of full-year revenue landed in 4Q with a 21.3% OPM, above the 9M cumulative 19.4%. This reflects operating-leverage on fixed costs plus a mix shift toward high-margin LAB and high-end dispensers (Mix Improvement).

1. Key investment points

2. Business model: precision control at scale

Founded in 1997, Protec has accumulated pneumatic-actuator and dispenser technology, evolving into an advanced-packaging equipment specialist. The two segments are the Systems Division (equipment) and the Components Division (pneumatic / inspection). The portfolio has shifted from SMT-centric to high-end semiconductor back-end equipment.

2.1 Dispensers — overwhelming market leadership

• Technical difficulty: control fluids of different viscosity at millisecond timing, dispense at nanometer accuracy. Globally a duopoly with US Nordson Asymtek.
• Expanding applications: semiconductor packaging (Underfill, Dam & Fill), LED phosphor coating, automotive camera-module bonding, etc.

2.2 LAB (Laser Assisted Bonder) — the game changer

Conventional Mass Reflow heats the entire substrate for 5–7 minutes, causing inevitable warpage. Protec's LAB heats only the chip area locally for 1–2 seconds via laser, minimizing thermal stress and dramatically improving throughput. Co-developed with US Amkor over multiple years — its mass-production reference data leaves followers (e.g., LaserSSEL) behind.

2.3 Role in the value chain

3. Decomposing the 4Q25 surprise

3.1 4Q standalone — back-calculated

3.2 Three triggers of the surprise

1. High-end semiconductor demand + LAB ramp: The company's own wording — "high-end semiconductors" — effectively means HBM/AI silicon. Customer mass-production lines likely absorbed LAB units in 4Q.
2. Customer capex restart: Global OSATs and US tech investments concentrated in 2H25. New orders tied to a customer with a prior LAB co-development history likely landed as 4Q revenue.
3. Subsidiary turnaround: Probe-card maker PMT and pneumatic-parts maker P&M rode the equipment cycle — boosting consolidated results.

4. Economic moat: technical lead and lock-in

4.1 Technical exclusivity of LAB

Mass Reflow's oven-style heating creates warpage, non-wets, and cracks the thinner the substrate. LAB heats only the chip area for 1–2 seconds → blocks thermal stress. The Protec / Amkor co-development reference is hard to imitate near-term.

4.2 Switching cost in dispensers

• Underfill and Dam & Fill require precise viscosity / dispense-volume / speed control — swapping the tool carries large yield-loss risk.
• Long-tail customers: Samsung Electronics, SK hynix, Amkor, StatsChipPAC plus many SME PCB/LED makers → strong lock-in.

4.3 Competitive risks

Hanmi Semiconductor dominates the TC bonder market and is accelerating hybrid-bonding development; players like LaserSSEL with area-laser tech are eyeing the market. Still, Protec advances its own area-laser tech (not spot-laser) and has verified mass-production references — so near-term moat erosion looks low.

5. End market: HBM and AI open a new horizon

5.1 HBM — LAB matters more as wafers thin

To stack HBM (8→12→16 layers), wafers must be ground extremely thin. As thickness drops, thermal warpage worsens — convection heating fails to secure yield. LAB, which heats only the bond area without warming the chip, is the most natural solution. Both Samsung Electronics and SK hynix are studying and expanding laser-based processes for next-gen HBM.

5.2 2.5D / 3D packaging & heterogeneous integration

AI chips place GPU, HBM, and logic on one substrate (CoWoS, etc.). CTE mismatches across interposer / substrate / die create warpage. Protec LAB delivers uniform laser irradiation even on large packages — a critical solution.

5.3 Growth mechanics: P and Q rise together

6. Financial health and valuation

6.1 Near debt-free capital structure

• Debt-to-equity: ~25% at end-3Q25, likely improved further on the 4Q profit surge.
• Cash: over KRW 100bn — Protec can fund R&D and capex without external financing.

6.2 Valuation — severely discounted vs. peers

• FY25 EPS ≈ KRW 3,209 (net income KRW 35.3bn / ~11M shares assumed).
• PER ≈ 10–12x on FY25 (varies with share price) vs. Hanmi Semiconductor at 50x+.
• Peers (PSK Holdings, EO Technics) trade at ~20x. If the market recognizes Protec's HBM value-chain entry, re-rating to 15–20x is justifiable.

6.3 Risks

• Hybrid bonding threat: bump-less Cu-Cu direct bonding post-HBM4 could reduce LAB demand. But cost and complexity push adoption out — TC bonding and LAB likely coexist for a while.
• End-market cycle: a temporary pause in AI-chip investment from 2026 could slow growth.

7. Conclusion: from "hidden HBM beneficiary" to "back-end leader proven by numbers"

The 2025 surprise proves Protec's body has changed — from "dispenser maker" to "advanced packaging solution company." The 244% profit growth is no coincidence: it is the match of capability and need.