Isu Petasys 1Q26: Supply Still the Bottleneck as Multi-Layer PCB Demand Outpaces Capacity Expansion

Core Conclusion

Isu Petasys posted 1Q26 revenue W340bn and OP W67bn, in line with consensus (W329bn and W66bn), driven by sustained AI data center demand and initial phase‑1 capacity. Multi‑layer PCB supply remains structurally behind demand, prompting management to raise 1H27 capacity target to 13k sqm/month (from 10.5k). A deliberate product mix shift toward higher‑margin accelerator/switch products (vs. low‑margin server boards) and expected ASP hikes from 2H26 should support sequential margin expansion, offsetting near‑term raw material headwinds. At 34x 2026e P/E, the stock trades at a premium to industry peers (26x), justified by ~40% EPS growth and improving mix, but offers limited near‑term upside to the W114,000 target price.

What the Market May Be Mis‑Pricing

1. Persistence of supply tightness – Backlog surged 57% QoQ to W580bn, and long‑term contracts are increasing. The market may not fully price the length of the multi‑layer PCB shortage or Isu’s ability to capture incremental orders as it accelerates capacity.
2. Margin trajectory from mix shift – The focus on accelerator/switch products (vs. low‑margin server boards) should drive OPM above the current 19.7% level, especially as ASP hikes take effect from 2H26. Consensus may underestimate the combined impact of volume, price, and mix on 2026‑27 earnings.

Evidence Chain

1. 1Q26 results met consensus; top‑line growth driven by AI and new capacity.

• Revenue W340bn vs. MSe W338bn, Consensus W329bn; OP W67bn vs. MSe W69bn, Consensus W66bn.
• QoQ revenue +14%, OP +18%; YoY revenue +35%, OP +41%.
• OPM of 19.7% was 80bp below MSe (20.5%) due to raw material cost pressure, but still up 110bp YoY.
  Investment implication: Execution is on track; capacity is being absorbed immediately.

2. Multi‑layer PCB demand continues to outrun supply.

• Backlog W580bn, up 57% QoQ, signaling stronger long‑term visibility.
• Management raised 1H27 capacity target to 13k sqm/month from 10.5k, a 24% increase.
• Long‑term contracts are rising as customers lock in supply.
  Investment implication: The company is effectively pre‑selling capacity; revenue visibility is high.

3. Product mix is actively shifting to higher‑value applications.

• ~50% of sales exposure to key customer’s AI accelerators maintained.
• Management explicitly prioritises accelerator/switch products over low‑margin server products, even if it means ceding server share at the key customer.
  Investment implication: This strategic choice should lift blended ASP and OPM over time, despite potential volume dilution in the lower‑margin segment.

4. Profitability to improve sequentially from 2H26 onward.

• Raw material inflation pressured 1Q OPM, but ASP hikes and richer multi‑layer mix are expected from 2H26.
• Phase‑1 capacity contributions are already in revenue; phase‑2 ramp will add volume without proportionate fixed cost increase.
  Investment implication: The earnings trough in OPM is likely 1Q26; sequential expansion should resume from 2Q26 or 3Q26.

Key Risks

• Slower adoption of 800G or reduced AI accelerator investment could weaken end‑demand for Isu’s products, reversing backlog trends.
• Market share loss to regional peers capable of producing 40+ layer PCBs could compress pricing power, especially if server‑grade capacity floods the market.

Valuation / Trading Implications

At W128,000 (May 15 close), the stock trades at ~34x 2026e P/E, well above the peer average of 26x. This premium is supported by: (1) ongoing product mix improvement toward accelerator/switch, (2) exposure to NVIDIA and Google, and (3) ~40% EPS growth in 2026e. However, the current price exceeds the W114,000 price target (implying ~11% downside). Near‑term upside is limited unless management announces further capacity expansion or a major order win. Investors should monitor for either catalyst to re‑rate the stock. A pullback toward W110,000‑115,000 would offer a more attractive entry point.

Appendix Table: 1Q26 Earnings Summary vs. Estimates