Q1 2026 broke the FY2025 ceiling — but at NT$488 the price already pays for it.

• The margin print refutes the peak. Q1 2026 gross margin 26.5% landed above the FY2025 full-year 25.9% in the seasonally weakest quarter, on no working-capital benefit. April monthly revenue NT$8.2B (+28.9% YoY) annualises to NT$98B against FY2025 actual NT$73.4B.
• Re-rating outran earnings. Trailing P/E expanded from ~11× at the FY2023 trough to 25.1× at FY2025 — a 2× multiple lift layered on a 2× EPS double. FCF yield (2.3%) now sits below the dividend yield (2.6%); the equity is no longer cash-self-funding at this multiple.
• Q2 2026 is the decider. A Q2 print holding 25%+ gross margin with revenue above NT$22B is the durability test that supports the NT$610 sell-side mean target; a print below 24% compresses EPS and the multiple at the same time. Result lands around 6 August 2026.

660bp of gross-margin lift in two years came from mix, not volume.

Sales volume was flat (~69,800 thousand sq ft both years) — every basis point of margin came from moving square feet out of mainstream multilayer into 16–32-layer AI-server, HDI, and 800G optical boards. Operating expenses grew 12% while gross profit grew 67% from FY2023 to FY2025. Net cash of NT$22B (8.5% of market cap) funds the AI-server capex in cash; durability depends on the pull continuing through FY2027 and CCL pricing not fully reverting.

The market pays for AI-tier participation; the evidence says Gold Circuit is taking the tier.

• The moat ratio is collapsing. Tripod / Gold Circuit revenue ratio compressed from 1.96× in FY2023 to 1.69× in FY2024 to 1.22× in FY2025 — a 38% close in two years. On current trajectory the AI pure-play overtakes Tripod by revenue in FY2026 or FY2027.
• Gold Circuit is winning the share-race. Gold Circuit grew FY2025 revenue +54% versus Tripod's +11.5%, and earns a 220bp net-margin premium (16.0% vs 13.9%) in the same tier. Trade-press attribution of the Q1 surge cites Tripod as the spillover beneficiary when upper-tier capacity is full — overflow economics, not qualification economics.
• Mid-mix means mid-multiple. 37% of FY2025 revenue still sits in mainstream multilayer and handset, where Wus, Shennan, and Victory Giant each grew 30%+ in FY2024 and China's share of global PCB output rose from 56.0% to 57.7% in one year. A Tier-1 AI multiple on Tier-2 share is unstable.

FCF halved, the dividend now exceeds FCF, and intra-group loans grew 96% in 18 months.

• FCF conversion broke. Free cash flow stepped from NT$12.6B (FY23) to NT$8.8B (FY24) to NT$5.8B (FY25) as capex doubled to NT$5.96B. FCF/NI fell from 2.08× to 0.57×. The FY2025 cash dividend of NT$6.68B is the first year the payout exceeded free cash flow — the gap is funded from balance-sheet cash, not earnings.
• Inventory ran ahead of revenue. Inventory grew 40% on 11.5% revenue growth in FY2025 (NT$8.6B → NT$12.0B), the textbook precursor to a margin air-pocket if AI-server sell-through soft-lands. Days payable also stretched from ~172 to ~188 — working-capital lifelines support the headline cash flow.
• NT$19.45B intra-group lending. A January 2026 NT$9.53B intercompany loan took total intra-group lending from NT$9.92B (Mar 2024) to NT$19.45B — uncollateralised, to wholly-owned subsidiaries, funding Vietnam Phase-2 and Hubei Xiantao expansion the annual filing does not itemise.

Two founders age 75, and a charter that has to outlive them.

The charter. Tripod's discipline is partly encoded in the bylaws — zero stock-based compensation, cash-only director pay capped at 1% of pre-tax profit (actual usage 0.35% in FY2025), a 40% minimum dividend payout floor (actual run-rate 65% three years running), and no buybacks at any multiple. Dividend per share has risen every year since FY2018, including through the FY2023 downturn.

The founders. Chairman Wang Jing-Chun (75) has run the business since 1991. Co-founder Hu (75) and Director Hsu (75) sit beside him. CEO Huang Le-Jen was elevated in 2022 but owns just 0.02% of the company. The Nomination Committee was only constituted in November 2025; first deliverables not yet visible.

The risk. Every other long-term risk has a hedge — geographic for tariffs, mix-shift for CCL, balance-sheet capacity for capex air-pockets, qualification cycles for share-take. The discipline lever has no hedge other than the charter itself. A post-transition CEO who authorises buybacks at 25× P/E, or chases a debt-funded acquisition, rebuilds the thesis from scratch.