Financials

Financials — What the Numbers Say

1. Financials in One Page

Tripod is a NT$73B-revenue Taiwanese printed circuit board (PCB) maker that, after a two-year cyclical pause in 2022–2023, has just printed the best year of its public history. FY2025 revenue of NT$73.4B grew 11.5%, gross margin expanded 330 basis points to 25.9%, operating margin hit a fresh high of 17.6%, and net income reached NT$10.2B (EPS NT$19.45, up 22% year over year). Cash conversion is genuine — operating cash flow has covered net income in every recent year — but FY2025 free cash flow stepped down to NT$5.8B as capex more than doubled to NT$6.0B to build AI-server and data-center PCB capacity. The balance sheet is fortress-grade: NT$26.7B of cash and short-term financial assets against only NT$7.5B of total debt, leaving roughly NT$22B of net cash and 88x interest coverage. Return on equity has risen from 14% in FY2023 to 19.6% in FY2025, and the board has signalled confidence by lifting the cash dividend to NT$12.70 (65% payout). Against all of that, the market pays roughly 25x trailing earnings and ~14x EV/EBITDA — a clear premium to where Tripod has historically traded, but a noticeable discount to peers Zhen Ding, Unimicron, Compeq and Gold Circuit on a trailing-P/E basis. The single financial metric that matters next is the durability of the new gross margin — specifically whether 1H2026 prints at or above 25% rather than reverting toward the 18-19% pre-cycle band.

A note on terms. Operating margin is profit from selling PCBs (after factory and overhead costs) as a share of revenue. Free cash flow is operating cash minus capital spending — the cash the company can actually return to shareholders or stockpile. EV/EBITDA divides enterprise value (market cap plus debt, minus cash) by pre-tax cash earnings; it lets you compare companies with different capital structures. ROE is net income as a share of book equity — a measure of how much profit management squeezes from each dollar of retained shareholder capital.

2. Revenue, Margins, and Earnings Power

Investor question: Is Tripod's earnings power genuinely stepping up, or is FY2025 a cyclical spike that will mean-revert?

Revenue cycled with the global PCB market: a 2021–2022 plateau around NT$63–66B, a NT$58.9B dip in FY2023 as consumer PCs and handsets normalized, then a sharp recovery to NT$73.4B in FY2025 as AI server, networking, and 800G optical PCB orders ramped. Net income tells the more interesting story: it stayed flat at roughly NT$6B through 2021-2023, then stepped up 38% in FY2024 and a further 22% in FY2025 — far more than revenue grew, signalling real operating leverage.

The margin chart is the centre of the bull thesis. Gross margin held in a 17.9%–19.3% band for three years, then ratcheted to 23.2% in FY2024 and 25.9% in FY2025 — roughly 700 bps of structural lift in two years. Operating margin moved even more (10.3% → 17.6%) because operating expenses grew only 12% while gross profit grew 67% from FY2023 to FY2025. This is the signature of a favourable mix shift: rising share of high-layer-count (HLC), HDI, and AI-server PCBs, where Tripod is qualified and pricing is firmer than mainstream PCB.

The quarterly trajectory is the early warning we want to monitor. Revenue, gross margin, and operating margin all peaked in Q3 2025 and pulled back modestly in Q4 (revenue −2.1% QoQ, gross margin −130 bps, operating margin −140 bps). That is not yet a break — Q4 still ran 530 bps above Q4 2024 on gross margin — but it does say that the easy upside from mix shift may be slowing. The question for 1H 2026 is whether the margin re-accelerates with another AI-server cycle or settles into a 24%-25% range, which would still represent a permanent rebase from the 18%-19% pre-cycle band.

3. Cash Flow and Earnings Quality

Investor question: Are the reported earnings turning into spendable cash?

A pleasant story up to a point. Operating cash flow has covered net income in each of the last three years (1.4x–2.6x), confirming that accounting earnings are backed by real receipts. But the FCF line tells a more nuanced story: NT$12.6B → NT$8.8B → NT$5.8B. Cash earnings are not deteriorating; capital spending is rising, from NT$3.3B in FY2023 to NT$6.0B in FY2025 — a deliberate reinvestment to build HLC and AI-server PCB capacity that management has flagged in its BofA 2026 presentation.

Quick definitions. Operating cash flow is cash from running the PCB business — earnings adjusted for non-cash items and working-capital swings. Free cash flow subtracts capex (factory and equipment spending) from operating cash flow. FCF margin is FCF as a share of revenue and answers the question "of every NT$100 of sales, how many become free cash?"

Three observations. First, the FY2025 jump in capex (capex/revenue rose from 4.1% to 8.1%) is the proximate reason FCF nearly halved — without it, FCF would have stayed close to NT$11B. Second, financing outflows (NT$4.9B in FY2025) reflect dividend payments and modest debt repayment; there is no buyback program. Third, inventory rebuilt by NT$3.4B in FY2025 (from NT$8.6B to NT$12.0B) and other current assets grew NT$3.3B, both consistent with preparing for higher AI-server PCB volumes. The earnings quality is genuine, but the cash yield to shareholders compresses in FY2025 because the capital is being put back into the business.

4. Balance Sheet and Financial Resilience

Investor question: Does the balance sheet add capacity for upside, or risk of downside?

Cash exceeds debt by a wide margin in every year. Total debt sits at NT$7.5B; cash and short-term financial assets total NT$29.3B; net cash on the balance sheet is therefore roughly NT$21.8B — about 8.5% of the market capitalization. Tripod added its first NT$1B of long-term debt only in FY2025, and short-term debt was actually paid down from NT$16.4B in FY2021. This is a deliberately under-leveraged structure.

The current ratio (current assets divided by current liabilities — a quick gauge of whether near-term bills can be paid) recovered from a stressed 0.75 in FY2021 to a comfortable 1.80 in FY2025. The FY2021 figure was a function of classifying all debt as short-term, not actual stress; cash still covered debt 1.2x in FY2021. Debt-to-assets has flatlined around 43%-45%, but the bulk of "liabilities" is now trade payables and accruals rather than financial debt.

The takeaway: Tripod has the balance sheet to absorb a downturn, fund the AI-server capex cycle in cash, and still raise the dividend. Interest coverage is reported at 88x. The only mild caution is inventory: it grew 40% in FY2025 against 12% revenue growth — possibly a build-ahead for AI-server orders, but a metric to watch in early FY2026 prints. A working-capital reversal here is the most likely source of an FCF surprise.

5. Returns, Reinvestment, and Capital Allocation

Investor question: Is management compounding per-share value, or just growing the company?

ROE has rebuilt from a cycle trough of 14.1% in FY2023 to 19.6% in FY2025, and ROA from 7.9% to 11.0%. Tripod is now earning a meaningfully higher return on each retained NT$ than it did two years ago, despite a NT$10.5B build in equity over that period. With negligible debt and no buybacks, almost all of that return improvement comes from the operating-margin re-rate plus stable ~0.8x asset turnover.

Note: dividend cash outflows estimated as cash dividend per share × shares outstanding (525.6M shares). FY2025 includes the proposed NT$12.70 per share subject to AGM approval.

Three pieces of evidence that management is creating per-share value:

1. EPS has grown from NT$11.15 (FY2021) to NT$19.45 (FY2025) — a 75% rise over four years, with no dilution (share count is flat at 525.6M).
2. Book value per share has risen from NT$72.9 to NT$102.9 — a 41% accretion despite paying out 63%-65% of earnings as cash dividends.
3. Dividend per share has nearly doubled from NT$7.00 to NT$12.70, with the policy floored at 40% payout and a recent run-rate of 65%.

Capital allocation is conservative-but-deliberate: capex first, dividends second, debt repayment when convenient, no buybacks. Given a 25x trailing P/E, buyback restraint is the right call — the company would be repurchasing shares at richer multiples than the historical norm.

6. Segment and Unit Economics

Investor question: Which end-market is doing the work, and how concentrated is the dependence?

Tripod does not publish a segmented income statement in the public consolidated filings ingested in this run (the segment dataset returns an error). What management discloses in the BofA 2026 Asia Tech Conference presentation, and what investors track through industry reports, are revenue applications mix categories: communications (handset, networking), computing (PC/notebook, server, data center), automotive/industrial/medical, and consumer.

These shares are directional estimates from public presentations and industry research, not audited segment figures — treat as the right shape rather than precise weights. The investor implication is clear: the mix shift toward server, data center, AI, and 800G optical PCB is the engine behind the margin re-rate, and that is also the segment with the highest concentration risk (hyperscaler capex, AI training spend, and individual customer programs). FY2025 inventory build of NT$3.4B and the FY2025 capex doubling are both placed against this end-market.

A formal segment reconciliation would materially improve underwriting precision; until management publishes one, the cleanest read-across is through customer-program disclosures and peer commentary at Zhen Ding and Gold Circuit.

7. Valuation and Market Expectations

Investor question: What is the market paying for, and is the price supported by the underlying financial quality?

At NT$488 the equity is valued at roughly NT$256B (market cap), with an EV of about NT$237B once NT$22B of net cash is netted out. Against FY2025 net income of NT$10.2B and EBITDA of roughly NT$17.2B, the trailing multiples are 25.1x P/E and 13.8x EV/EBITDA. The trailing P/E sits well above Tripod's pre-2024 single-digit-to-low-teens range (the stock spent most of 2018-2023 at 8x-13x P/E) but compares with peer multiples that are materially richer (see Section 8).

The trailing-P/E expansion from ~11x (FY2023) to ~25x (FY2025) represents a roughly 2x re-rating layered on top of a doubled EPS — i.e., the share price has done much more than the earnings themselves. That is the central market-expectation: this is no longer a mid-cycle PCB processor, it is being repriced as an AI-infrastructure beneficiary with structurally higher margins.

The scenario table makes the underwriting tension explicit. The current NT$488 price already approximates the bull case, in which margins climb a further ~250 bps and the multiple holds at 22x. The base case implies ~28% downside, and the bear case (a return to pre-cycle margins and multiple) implies the stock could lose two-thirds of its value. This is therefore an expectations-laden setup: the market is paying for continuation, not normalization.

8. Peer Financial Comparison

Investor question: Where does Tripod sit on quality, growth, and price vs the top global PCB peers?

Notes: Market caps converted at the 2026-05-21 TWD/USD rate; TTM revenue is the latest available trailing-twelve-month figure per Yahoo Finance; EV/EBITDA computed from Yahoo enterprise value and reported EBITDA TTM. ROE not consistently disclosed for TW peers. Margins for Compeq and Gold Circuit not consistently disclosed in the source feeds and shown as null rather than estimated.

Two things stand out. First, on operating quality Tripod is best-in-class: its 17.6% operating margin and 19.6% ROE exceed Zhen Ding (5.2% net margin), Unimicron (cyclical trough), TTM (9.1% operating margin), and the Taiwan multilayer mid-caps. Second, on price Tripod is the cheapest of the group on trailing earnings (25.1x vs 47x-208x for peers). Part of the peer multiple inflation reflects depressed trailing EPS (Unimicron's 208x is a cyclical-trough P/E that compresses to ~32x on consensus forward), and part reflects investor expectations that Unimicron and Gold Circuit have larger upside in ABF substrate and pure-play AI server PCB respectively. But on any blended look — margin times multiple — Tripod looks like the highest-quality, lowest-priced entry to the same AI server PCB exposure.

9. What to Watch in the Financials

What the financials confirm: Tripod has earned its margin re-rate. Operating margin doubled in two years, ROE is at a structural high, the balance sheet has more cash than debt, and the dividend has nearly doubled. The new earnings power is genuine cash, not accounting credit, and management has chosen to redeploy it into the highest-returning end-market (AI server PCB) rather than buybacks at expensive multiples.

What the financials contradict: The FY2025 FCF step-down to NT$5.8B (only NT$11 per share) sits awkwardly against a NT$488 share price — current FCF yield is roughly 2.3% on the equity, so investors are paying for future free cash, not current. Inventory grew faster than revenue, hinting at a possible build-ahead that has not yet converted. And the valuation has done more lifting than the earnings: a 2x P/E re-rate on top of a 2x EPS double means the share price has compounded faster than the underlying business.

The first financial metric to watch is the Q1 and Q2 2026 gross margin print. Anything at or above 25% will validate the new earnings power and keep the multiple anchored. A retreat into the 22%-23% zone will reset expectations sharply, because the bull case requires not just current margins but further expansion to support today's multiple.