The Money Rails: Why Your Bank Stopped Being a Vault

For most of our lives, the mental image of a bank was incredibly simple. It was a dignified, slightly boring building with a heavy vault. The business model was just as heavy and stationary: we gave them our money to keep safe (deposits), they lent that money to businesses or homebuyers (loans), and they pocketed the difference in interest rates.

We called this “prudence.” If the bank made more money, it meant they were lending more. If they made less, the economy would slow down. It was a world explained by a single, simple metric: the spread.

That story still comforts us. It feels safe. But if you are trying to understand modern banking with that mental model, you are looking at a map of a world that no longer exists.

The shift: From storage units to highways

The biggest change in banking isn’t happening in the vaults; it’s happening in the pipes. Banks have realized that simply holding money and lending it out is a slow way to get rich. The new game isn’t about storing value; it is about moving it.

To see this, you have to ignore the loan book for a moment and look at the “activity engine.”

In India, the smoking gun is the humble credit card. Between 2017 and 2021, the number of credit cards in circulation didn’t just grow; it doubled from 29 million to over 60 million. But the number of plastic cards isn’t the point, it’s the obsession with using them. Even during the pandemic, when the world stood still, transaction values skyrocketed.

This wasn’t just people buying groceries. This was the banks quietly building a massive toll‑road system.

The old model: Make money once, over 20 years, on a home loan.
The new model: Make a fee every time a customer buys coffee, pays a bill, or shops online.

Every swipe is a fee. Every transaction is a piece of data. Every interaction is a chance to sell insurance or a mutual fund. Banks stopped acting like vaults and started acting like rails.

The giants of Indian banking: HDFC and Axis Bank

This transformation is clever because the best banks hide it well. Look at HDFC Bank. On the surface, it looks like the classic, prudent bank your grandfather loved. The accounting looks traditional, the asset quality is stable, and the margins are healthy. But if you look under the hood, HDFC isn’t just sitting back and collecting interest. They are aggressive gatekeepers.

A huge chunk of their revenue now comes from fees, commissions, and third‑party sales. They aren’t just lending you money; they are charging you for access to the financial system. They have turned the customer relationship itself into a product.

Axis Bank is even more transparent about this. When they bought Citi’s consumer business in India, many scratched their heads. Why buy a credit card portfolio? The logic was purely about velocity. They didn’t just want loan accounts; they wanted high‑frequency users.

The data proves the strategy: By late 2025, Axis Bank reported that “Granular Fees,” those tiny, recurring charges from transactions, constituted over 92% of their total fee income. They understood that a customer who uses their card three times a day is worth more than a customer who just parks money in a savings account.

The allocation pivot

This philosophy has fundamentally changed where banks send their money. In the vault era, capital was allocated to heavy industry: steel plants, power grids, and infrastructure. It was slow, secure, and nation‑building.

Today, the credit allocation matrix has pivoted toward consumption.

Industry vs. individuals: In early 2025, credit growth to the industrial sector slowed to roughly 5–7%, while personal loans continued to surge at 12–14%.
The retail takeover: Retail loans now make up nearly 33% of total bank credit in India, up from just 24% in 2020.

Why? Because a corporate loan is a low‑yield marriage, but a personal loan is a high‑yield fling. Banks have moved from funding production (factories) to funding consumption (iPhones and vacations).

Shifting from lazy money

Why did banks have to change? Because the customer changed.

There was a golden era for banks when deposits were lazy: you put your salary in the bank and left it there. You were a passive source of cheap funding for the bank.

Those days are dead. Today, money is fidgety. With a single tap on a phone, you can move your savings to a high‑yield liquid fund, a stock broker, or a digital wallet. Fintech apps have trained us to chase every 0.5% of extra return.

Because banks can no longer count on your money sitting still for free, they can’t rely solely on lending spreads to survive. If they did, they would go broke the moment interest rates tightened. They need that fee income, the recurring, non‑lending cash flow, to keep the lights on when the lending cycle gets tough.

The proof in the pipes

If you want evidence of this pressure, just look at the credit‑deposit gap. Throughout 2025, credit (loans given out) grew at ~12%, while deposits (money coming in) lagged at ~9%.

We are borrowing faster than we are saving. To bridge this gap, banks are forced to rely on high‑velocity earnings (fees) rather than just stationary earnings (interest).

Hidden risks in the system

This brings us to the paradox. Modern banks look safer than ever. They have capital reserves and fewer bad loans (NPAs). Yet, the stock market remains nervous. Why?

Because while the balance sheets look cleaner, the risk hasn’t disappeared. It has just become off‑balance‑sheet.

1. The co‑lending black box

Banks and fintech apps are doing this cute little arrangement where the bank provides the money and the app provides the borrower. To the customer, it feels like “the app gave me a loan,” but the risk ultimately sits on regulated lenders. That structure creates a co‑lending black box: credit origination is happening at app speed, while transparency and risk visibility lag behind.

The real problem is credit stacking. A borrower can take five small ₹10,000 loans from five different apps, and each app might underwrite them as if they are only taking one small loan. The borrower looks safe in each app’s narrow view, but collectively they’re over‑leveraged. And because the same few large banks or NBFCs often fund multiple apps, the bank can end up with concentrated exposure to the same stressed borrower clusters while thinking it’s diversified across “thousands of users.”

This isn’t just a storyteller’s paranoia. RBI’s digital lending rules explicitly try to remove hidden intermediaries from fund flows: disbursals should go into the borrower’s bank account and repayments should go directly to the regulated entity, limiting how much a lending service provider can “sit in the middle.” That’s basically the RBI acknowledging the opacity risk in app‑led lending.

RBI also tightened the co‑lending framework in August 2025 (effective January 1, 2026) to force stronger accountability and reduce moral hazard. The rules require each co‑lending partner to retain at least 10% of each loan, mandate reporting within 15 days, limit default‑loss guarantees, and crucially push borrower‑level classification, which helps prevent the “it’s diversified across partners/apps” illusion when the risk is actually concentrated in the same borrowers.

2. The liquidity mismatch

Bank runs used to be slow. Panic meant hearing a rumor, walking to a branch, and standing in a queue. That friction mattered. Today, panic moves at phone speed. A single viral message can trigger millions of digital transfers in minutes. That change in speed is what makes the modern liquidity mismatch genuinely dangerous.

The core issue is a classic maturity mismatch, now amplified by digital banking. Banks are increasingly funded by money that behaves like it has a maturity of seconds, while liquidity regulation is built for much longer stress horizons. Under Basel III, the Liquidity Coverage Ratio (LCR) requires banks to hold enough high‑quality liquid assets (HQLA) to survive a 30‑day stress scenario. The problem is that digital panic does not wait 30 days. It can hit in 30 minutes.

On the liability side, banks rely heavily on transactional and retail deposits such as UPI‑linked balances and savings accounts. These are low‑cost funds, but they are also instantly withdrawable. With UPI and mobile banking, deposits can be moved out with almost zero friction. Regulators have acknowledged this shift. In April 2025, the RBI amended LCR norms to impose an additional 2.5% run‑off factor on digitally enabled retail and small‑business deposits, explicitly recognizing that digitally accessible money can run faster than traditional assumptions.

On the asset side, this near‑instant funding is increasingly deployed into longer‑tenor, illiquid loans. In India, personal loans now account for roughly 32–33% of total bank credit, reflecting the growing focus on unsecured retail lending. Once these loans are issued, cash is locked in for years. In a stress event, banks cannot suddenly convert these assets back into liquidity.

The danger is a digital bank run. A rumor spreads, confidence breaks, and withdrawals become self‑fulfilling. The global example is Silicon Valley Bank, where regulators reported $42 billion in deposits withdrawn in a single day, with requests for another $100 billion the next day. That episode showed how social media coordination and frictionless transfers can overwhelm liquidity buffers almost instantly.

The system now resembles a high‑speed train with weak brakes. Financial technology has massively increased the speed of money, but the mechanisms to slow mass outflows have not kept pace. As a result, digital finance turns liquidity risk from a slow‑burning problem into a flash event, making the mismatch between “three‑second money” and “three‑year loans” a central risk for modern banking stability.

The new reality

Banking hasn’t become boring again; it has just become quieter. The drama has moved away from the loan department and into the back‑end technology that processes our lives.

The banks winning today aren’t the ones with the biggest vaults. They are the ones that have successfully woven themselves into the fabric of your daily existence.

The question for the next decade isn’t “Can the bank lend more money?” It is: “Can the bank keep you inside its ecosystem, swiping, clicking, and transacting, while the world of finance gets faster and more competitive?”

If you are still judging a bank solely by its loan book, you are watching the steam engine while everyone else is boarding the high‑speed train.