Back to All Writing
    e272026-08-15

    Southeast Asia solved distribution: Now fintech has to scale on the balance sheet

    Originally published in e27 on 2026-08-15.

    Read Original Article on e27

    Core Argument

    A founder showed a growth chart, monthly actives climbing, merchant count doubling, transaction volume up across three markets, then asked which wallet licence to apply for next. Asked what his cost of funds would be, neither he nor his finance lead had an answer. That silence is the state of financial services in Southeast Asia: a decade spent rewarding reach, and the economics never followed. Fintech in the region has stopped being a distribution business and become a balance sheet business. The cause is public infrastructure: QRIS, PromptPay, DuitNow, PayNow, QR Ph, VietQR, KHQR, all built by central banks, made interoperable, then linked across borders. By December 2025 the region ran 29 cross-border payment linkages. Project Nexus, led by the BIS with India, Malaysia, the Philippines, Singapore and Thailand as founding participants, replaces bilateral stitches with a single hub. This inverts the Western pattern where rails are private and the toll is the business. When the state owns the rails, distribution stops being defensible. What carries margin is deposits, underwriting, and the spread between them. Singapore's three digital banks lost a combined S$358.75 million in 2024 while the three incumbents earned roughly S$25 billion between them. GXS sits inside Grab, MariBank inside Sea, neither had a reach problem, and reach did not save them. Indonesia shows the other half: eight of nine listed digital banks posted a 2025 profit, but only SeaBank's 2.3 percent return on assets looks like a conventional bank. The rest are thin or flattered by base effects. The 29 cross-border linkages processed 36.2 million transactions worth US$716.4 million, about twenty dollars a transaction, tourists buying lunch, not businesses settling invoices. If a regional plan assumes the QR corridor will carry business-to-business settlement, it has designed for a consumer rail and loaded it with trade. The conclusion: if payments-led, rebuild around funding cost, not downloads. If distribution-rich but outside financial services, do not build a bank, route, and partner with someone else's balance sheet.

    What I'd Revise Now

    Two figures in this column have moved since August, and one reframes the argument rather than just updating it. GXS did not narrow toward 2027 profitability as smoothly as I implied. Its FY2025 loss came in at S$208.1 million against S$214 million the year before, a real but marginal improvement, and coverage describes it as posting "the heaviest loss" among Singapore's digital banks even as its loan book tripled. The bank I should have named as evidence for the other side of the argument is Trust Bank, which posted its first profitable month in March 2026. Trust is majority-owned by Standard Chartered with NTUC Enterprise as a local partner, not a consumer platform with a large existing user base the way GXS and MariBank are. The digital bank that reached profitability first in Singapore was the one with a bank as its anchor shareholder, not the one with the app. Thailand has also moved past the column's framing. I described the virtual bank licences as due to launch "by mid-2026." Clicx Bank, backed by Krungthai, AIS and PTT's OR, went live in June. Early coverage already shows the same pattern this column predicts: a Forbes analysis from June concluded Thailand's entrants launch into "a model that barely pays," with the Bank of Thailand's own capital requirements, five billion baht rising toward ten billion, built around the expectation of years of losses. Both updates support the column's core claim more than they complicate it. Reach did not save GXS or MariBank. What is starting to work is a bank-anchored, deposit-first structure, which is the opposite of the distribution-first model most fintechs in the region were built on.

    Key Takeaways

    • Public payment rails removed the moat from distribution; margin now sits in deposits and underwriting, not reach
    • Singapore's digital banks lost S$358.75 million combined in 2024 despite sitting inside Grab and Sea, the two best-distributed platforms in the region
    • Only one of Indonesia's eight profitable digital banks, SeaBank, has a return on assets that looks like a real bank
    • Cross-border QR linkages average about twenty dollars a transaction, a consumer rail, not a trade settlement layer
    • The strategic choice is fund the book or route to someone else's balance sheet; building a licence for its own sake no longer works
    Visit Original Outlet (e27)

    Related Insights