Long-Term Thesis
Long-Term Thesis — A 5-to-10-Year View
1. Long-Term Thesis in One Page
The long-term thesis is that the Taiwan battery-swap network — 665,000 paying riders, 2,700 stations, 1.3 million batteries, 800 million cumulative swaps, and 1.34% annual churn — eventually outgrows the loss-making scooter business attached to it, re-rates as a subscription utility, and compounds modest mid-single-digit subscriber growth and ARPU drift into a durable cash machine that earns its $933M cumulative capex base back over 7-10 years. This is not a long-duration compounder unless three things hold across the cycle: (a) the proprietary swap form factor remains the de facto Taiwan standard, (b) the energy segment delivers the 2026 non-IFRS profitability / 2027 FCF-positive guide as the first of multiple steps toward a 30%+ subscription gross-margin business, and (c) the equity stub survives the pre-committed NT$2.5B (~$80M) controlling-shareholder dilution path through December 2026 without further pricing resets. The bull case is a Taiwan urban-energy utility with a free option on Vietnam, Korea, and second-life batteries; the bear case is a slow-burn margin squeeze inside a contracting island market while bondholders and the controlling shareholder absorb most of the network's economics. Underwrite the energy utility; treat hardware and international as residuals; let the balance sheet decide the equity-versus-credit risk-reward over a five-year window.
Thesis Strength
Moat Durability
Reinvestment Runway
Evidence Confidence
The single most important long-term conclusion: the moat is real but geographically captive, and the equity may not capture all of the network economics. Position size to that asymmetry — not to consensus on quarterly margin.
2. The 5-to-10-Year Underwriting Map
The driver that matters most is the first one — subscription compounding in Taiwan. It is the single source of the recurring revenue that justifies the SOTP lens, the only line of the P&L that grew through a -23% hardware year, and the only metric that, if it breaks, takes the whole multi-year case with it. Every other driver — energy segment economics, capital allocation, PBGN flywheel — is downstream of whether 665K paying subscribers compound to ~1M with churn capped at 1.5%. Without that, no margin trajectory or international option can repair the thesis.
3. Compounding Path
The compounding question is whether subscription growth and ARPU drift, combined with falling capex on a built-out network, can convert today's $35.9M of OCF into a self-sustaining cash machine over five-to-ten years. The historical record is brutal — cumulative FCF of -$732M over seven years — but the recent trajectory is the first credible inflection, and the math from here works if subscribers compound and capex stays disciplined.
The path on the chart from 2027 forward is illustrative — it assumes ~3-4% subscriber CAGR off a contracting Taiwan PTW base, ARPU drift of 1-2% per year, no international subscription contribution, and no churn deterioration above 1.5%. Under those assumptions the subscription revenue line approaches ~$245M by 2030 and ~$260M by 2033 — roughly two-thirds higher than 2025. That is not aggressive: it assumes neither a Vietnam ramp nor a second wave of PBGN-driven share expansion. It does assume Taiwan ePTW share stays above 60% and that the swap form factor remains the de facto standard.
The compounding math is unusual because most of the asset base — the 2,700 GoStations, 1.3M batteries, 901-patent portfolio — has already been paid for through seven years of $933M cumulative capex. The remaining work is operating discipline plus modest subscriber growth, not another buildout phase. That is the case for owning a sub-book-value equity stub on a network that may not need another $500M of capex to monetize what is already there. The case against it is that a moat which still requires capital to maintain — and Gogoro's does, because batteries cycle out and stations need refresh — is only as durable as the financing access, and the balance sheet is the most fragile part of the multi-year case.
4. Durability and Moat Tests
The test most likely to be decided in the next 24 months — and most likely to settle whether this is a long-duration compounder or a value trap — is the second one: whether the proprietary swap form factor survives any regional standardization push. The 20-F risk factor states this plainly. A Chinese national standard or an India consortium standard could erase the switching cost overnight. The first test (churn) is the leading indicator of network gravity; the fifth (equity capture) is the leading indicator of whether minorities actually see the cash flow that the network generates. The third and fourth are the validation work over a 5-year window. All five have to hold for the long-term thesis to work; the absence of any one breaks it.
5. Management and Capital Allocation Over a Cycle
The capital-allocation track record splits cleanly into two regimes. The first — the Horace Luke / SPAC era through Q3 2024 — was a buildout strategy that spent $933M of cumulative capex against $2.5B of cumulative revenue, consistently promised international scale, missed revenue guidance three years running, and ended with a $32M China/India impairment, a Nasdaq min-bid notice, and a Taiwan subsidy investigation. The second — Henry Chiang / Bruce Aitken from Q4 2024 forward — has delivered against smaller, more controllable commitments: $24M of opex savings in 2025, capex cut in half to $65M, record adjusted EBITDA, and the long-promised crossover where subscription revenue exceeded hardware. The current team has roughly a year of credibility on the operating side. They have no track record on capital markets execution beyond consenting to the Gold Sino dilution path.
The hard question for a 5-to-10 year owner is whether incentives align over a full cycle. CEO Chiang has no meaningful equity stake; CFO Aitken's grants are deeply out of the money; SBC collapsed from $26.3M in 2023 to $2.3M in 2025 because no new awards were granted, not because the cost was earned. The forensic page rates non-IFRS metric hygiene a Red flag — adjusted EBITDA exceeded IFRS net loss by $140M in FY2025, the widest gap in company history, and the framework excludes battery-upgrade derecognition that management itself describes as recurring. This is presentation aggression, not fraud, but it matters over a multi-year window: every quarter that the IFRS/non-IFRS wedge persists is a quarter the equity story relies on management's framing rather than audited reality.
The single largest capital-allocation question is whether the NT$2.5B Yin undertaking — a director's personal commitment to procure ~$80M of equity by December 2026, signed as the price of the syndicated-loan covenant relaxation — closes at market terms or at the kind of negotiated discount the March 2026 Gold Sino tranche set. Each tranche issued under that undertaking is mechanically dilutive on the cap table, and Gold Sino is one share away from triggering Castrol's $25M put on change of control. The capital-allocation discipline visible in the 2025 capex line is real; the capital-markets discipline tested by the dilution path is the multi-year unknown. The single quarter where Q1 2026 OCF flipped positive on a built-out network is what the bull is buying; the contractually pre-committed dilution path that runs through December 2026 is what the bear is selling.
The honest read: the current team's operating instincts (cost discipline, capex throttling, focus on Taiwan) are correct for a 5-to-10-year thesis. Their capital-markets actions (consenting to a dilution path that walks a single shareholder toward majority control at negotiated discounts) are damaging to multi-year per-share value. The first leg is what management can control; the second is what the lenders have forced. The owner of this equity on a five-year horizon is making a bet on the first leg overwhelming the second — which only works if 2027-28 free cash flow is large enough to refinance the company on conventional terms before Gold Sino crosses 50%.
6. Failure Modes
The two failure modes most likely to define the next five years are the forced-dilution path through December 2026 and a regulatory standardization of the swap form factor. Both are partly exogenous, both are not yet priced into the equity, and either alone is sufficient to break the thesis. The other four failure modes are slower-moving and visible in quarterly disclosure with adequate lead time.
7. What To Watch Over Years, Not Just Quarters
The long-term thesis changes most if the energy segment crosses into non-IFRS profitability in 2026 and free-cash-flow positive in 2027 while the NT$2.5B Yin undertaking closes at market without further pricing resets — because that is the single combination of evidence that converts this from a "moat the equity may not capture" story into a "Taiwan subscription utility whose equity finally re-rates as one" story. Any other combination — operating success without equity discipline, or balance-sheet stabilization without operating proof — leaves the multi-year case ambiguous.