In March 2026, India’s neobank Fi quietly shut down the banking layer of its app. Customers who’d opened savings accounts through Fi’s sleek interface got an email saying the account itself was fine, safe, and fully funded, but they’d now need to manage it through FedMobile, the app belonging to Federal Bank, the actual licensed bank that had been holding those accounts the whole time. The Federal Bank called it a “business re-alignment.” Fi, which had served more than 3.5 million customers and processed over a billion transactions since 2021, said it was refocusing on deep tech and AI instead.
It’s a small story, but it says something big about how modern fintech actually works. Fi was never a bank. It was a beautifully designed app sitting on top of a bank’s infrastructure, and when that arrangement ended, the app was the part that disappeared, not the money or the underlying account. That’s true of nearly every neobank and investment app you use, and understanding why changes how you think about the fintech apps on your phone.
A neobank isn’t a bank, it’s an interface
A neobank is a digital-only financial brand with no branches, usually built around a slicker app and better UX than a traditional bank offers. What it almost never has is an actual banking license. In India, the Reserve Bank of India doesn’t recognize “neobank” as a licensed category at all. Companies like Jupiter and Fi built their savings accounts, debit cards, and UPI features in partnership with Federal Bank, while others lean on partners like Axis Bank or DCB Bank. The app handles design, rewards, and product experience. The regulated bank holds the deposits, the license, and the compliance obligation.
This isn’t unique to India. Chime, one of the largest US neobanks, doesn’t hold a bank charter either, it operates through partner banks. N26 and Revolut are different in that both do hold their own banking licenses in parts of Europe, which is part of why their regulatory footprint looks different from Chime’s or Jupiter’s. The pattern that matters is this: “neobank” describes a product layer, not necessarily a legal one, and the app you trust with your salary might be one commercial renegotiation away from changing hands, exactly like Fi’s customers just experienced.
Investment apps run on the same logic
Swap “bank” for “broker” and the same story plays out in investing. A slick SIP or stock trading app rarely holds its own clearing infrastructure or acts as its own depository participant from scratch. In India, every SEBI-registered intermediary, whether it’s a mutual fund app, a discount broker, or a wealth management platform, has to route KYC through a SEBI-registered KYC Registration Agency, a system introduced under the SEBI KRA Regulations of 2011. Complete your KYC once with any KRA, like CAMS KRA or NSE’s KRA arm, and every other SEBI-registered platform can pull that same verified record instead of putting you through the process again. It’s the investing world’s version of what Aadhaar-based eKYC does for bank accounts, one verified identity, reusable everywhere.
Globally, the shape is similar even if the plumbing differs. Robinhood, Wealthfront, and most retail investing apps sit on top of clearing firms and custodians that actually hold and settle the securities. The app is the experience layer. The infrastructure underneath is regulated, standardized, and mostly invisible to the end user, right up until something changes and you notice it was there all along.
Onboarding is the front door, and it’s an API problem now
Whether you’re opening a neobank account or a SIP, the first thing that happens is identity verification, and this is where the API layer earns its keep. A well-built onboarding stack pulls Aadhaar or PAN verification, checks the Central KYC Registry for an existing record, and for business or MSME accounts, verifies GST filings, all through a handful of API calls instead of a branch visit and a stack of photocopies. This is exactly the kind of infrastructure Decentro’s KYC APIs are built for, bundling Aadhaar, PAN, CKYC, and GST-based checks into a single integration so a neobank or investment app doesn’t have to build a separate relationship with each government registry on its own.
This matters more than it sounds like. Onboarding is usually the single biggest point of user drop-off in any fintech product. Every extra document, every manual review step, is a chance for a genuine customer to give up halfway through. The apps that feel instant on the front end are, almost without exception, running fast, automated verification on the back end.
The part nobody thinks about: recurring payments
Here’s the piece of neobank and investment app infrastructure that gets the least attention despite doing the most work. A SIP only works if the monthly debit actually happens, reliably, for years, without the investor having to manually approve a payment every single time. In India, that’s handled through UPI AutoPay, NPCI’s mandate system launched in 2020, which lets a user authorize a recurring debit once with their UPI PIN and then have subsequent payments processed automatically up to a set limit, ₹15,000 for general recurring payments, and up to ₹1 lakh for categories like mutual fund SIPs, insurance premiums, and loan EMIs, before re-authentication kicks in again. Older systems like NACH e-mandates still handle a lot of this too, especially for higher-value or bank-account-based recurring debits.
Neobanks lean on the same rails for a different reason: auto-save features, recurring transfers into savings “pots,” subscription billing for premium tiers, and standing instructions all depend on dependable, low-failure recurring payment infrastructure. Getting this wrong doesn’t just cost revenue, it means a SIP silently stops investing or a savings goal quietly stalls, and the user often doesn’t notice until months later. It’s exactly the kind of plumbing Decentro’s recurring payments APIs are built to handle, managing UPI AutoPay and NACH mandates so a fintech product can set up, track, and retry recurring debits without stitching together multiple bank and NPCI integrations itself.
What the Fi story actually teaches
Come back to Fi Money for a second. The reason its customers didn’t lose a rupee when the app shut down is precisely because the underlying infrastructure, the actual bank account, was never Fi’s to begin with. It belonged to the Federal Bank, a regulated entity, the whole time. The app could disappear overnight because the app was never where the real financial relationship lived.
That’s not a flaw in the neobank model, it’s the model working as designed. The lesson for anyone building, or just using, fintech products is that the interface is replaceable and the infrastructure underneath it isn’t, or at least shouldn’t be. Which is exactly why so much of the fintech industry’s actual innovation has moved from app design to the API layer: identity verification, ledger management, mandate handling, and compliance, the unglamorous stuff that determines whether your money is actually safe and your SIP actually fires on schedule, regardless of which app happens to be sitting on top of it this year.