On 11 March 2026, Fi Money wound down the banking services on its platform, ending its consumer neobanking run a little over four years after launch. The company — founded by former Google Pay executives and backed by Sequoia (now Peak XV) and Temasek — had raised $137 million and reached a peak valuation of $522 million. Some 3.5 million users were redirected to Federal Bank’s own FedMobile app. Fi survives as a B2B AI enterprise software company; the consumer bank it set out to build does not.

The Diagnosis

Three failure modes compounded, but one was fatal.

First, the unit economics never turned. In FY23, Fi reported revenue of ₹38 crore against losses of ₹301 crore — roughly ₹8 lost for every ₹1 earned, four years and $100M+ of deployed capital into the journey. Indian neobanking’s structural problem is well documented: without a banking licence, the fintech layer earns thin distribution economics on deposits and payments while bearing venture-scale customer-acquisition costs.

Second, the regulatory perimeter never opened. India issues no digital banking licences, so every neobank is contractually a front-end to a licensed partner. Fi’s entire product ran on Federal Bank’s infrastructure — accounts, deposits, cards, the lot.

Third — and decisively — the dependency was single-threaded. When Federal Bank ended the partnership in March 2026, Fi’s consumer product ceased to exist with no alternative banking partner ready. The February 2026 pivot announcement and 40% workforce reduction preceded the shutdown by weeks; the partner’s exit converted a strategic retreat into a hard stop. A company can survive bad unit economics for as long as capital tolerates it. It cannot survive the withdrawal of the licence it rents.

Why It Matters

The lesson generalises well beyond neobanking. Any startup whose core product depends on a single upstream counterparty — a bank partner, a platform API, an exclusive data provider, a sole cloud or model vendor — carries a concentrated existential risk that rarely appears in its pitch metrics. Fi’s dashboards showed users, engagement, and deposit growth. None of those numbers measured the only variable that ended the company: one counterparty’s willingness to continue.

For founders, the diligence question to ask yourself is brutal but simple: if my most important partner sends a termination notice tomorrow, do I have a product in ninety days? If the answer is no, dependency diversification is not an infrastructure chore — it is the survival roadmap.

For investors, partner-concentration risk deserves the same scrutiny as customer concentration. A company earning 100% of its existence through one contract should be assessed at a structural discount, whatever the growth curve says.

The Charaka View

Our postmortem database — 387 documented startup failures as of 3 July 2026 — shows dependency failure is chronically under-weighted next to the fashionable killers (no market need, founder conflict, cash-out). Deaths like Fi’s get recorded as “business model failure,” which obscures the mechanism: the model was known to be hard from day one; the timing of death was set entirely by a third party. In our own analytical work we treat single-counterparty dependence as a named risk category precisely because it fails silently — every metric looks fine until the day none of them matter.

This analysis draws on TechCrunch, the Value for Startups Fi Money investor report, and TechBuzz. Human editorial oversight applied.

This analysis is informational and does not constitute investment advice, a research report, or a recommendation to buy, sell, or hold any security.

Charaka Notes by Manthan Intelligence. Subscribe