Banking’s Golden Age — or Its Final Chapter?

In 2024, the global banking sector posted $1.2 trillion in net profit — very high figure with respect to many other industries. This exceptional profitability is not a one-off; it has been building for several years, with banks outperforming virtually every other sector. So why is so much changing in an industry that looks, on paper, so healthy? And more importantly: is any of this sustainable? A closer look at the sector reveals a set of structural fault lines that are quietly widening beneath the surface.

During this period, it was not just profits that grew — revenues, funding sources, and balance sheets expanded as well. Yet beneath these record high numbers, structural challenges are accumulating. Let’s examine some of them.

• Customer relationships: in the traditional banking model, a customer relationship meant permanent ownership of the customer. People kept their savings at one bank, took out loans at that same bank, and stayed — creating a loyal, captive customer base.

But today, customer loyalty looks very different.

i) Customers no longer value loyalty in itself.

ii) They are well-informed enough to compare the cost, quality, and features of services across different banks.

iii) They can move their funds from one institution to another easily, quickly, and usually at no cost.

iv) Thanks to a growing range of products, they can even transfer their debts and loans from one institution to another.

v) Open banking has forced banks to offer increasingly generous concessions to attract customers from competitors — and in terms of both pricing and product features, those concessions keep shifting further against the banks’ interests.

• Fintechs: day by day, the sector’s pie is being divided more favorably for fintechs and less so for traditional banks — both in terms of customers and transaction volume. Banks do hold a competitive advantage built on years of experience, brand recognition, reach, and above all, trust. But while large, established banks struggle to keep up with rapidly evolving technology, fintechs — with their agile structures — can respond to customer needs and expectations far more quickly.

• Digital banks (Neobanks): digital banks now account for 17% of total global banking and fintech revenues, and have reached 37% of combined market capitalization. Players like Revolut, Monzo, Nu, Starling, and N26 are growing their customer bases and transaction volumes at rates that outpace traditional branch-based banks. Several of these institutions — all of which entered our lives within roughly the last decade — have now achieved net profitability.

• AI and agents: there is something critically important happening here. AI is not showing up merely as an efficiency tool. Beyond guiding customers, AI agents are now tracking which institution to leave, when to leave it, and which one to move to — at the product and transaction level — and then taking those actions automatically. This is precisely what will erode the “customer inertia” that banks have profited from for years. In many situations, many customers simply could not be bothered to act on small improvements in returns or marginal cost savings — so they stayed put. And banks benefited enormously from that inertia. The decisions and transactions that once required too much effort from the customer are now being made and executed automatically by agents.

AI agents search for the best options across markets, products, and banks; they make intelligent decisions based on the person’s current circumstances and needs; they adjust deposits and loans by term and rate to secure better yields, lower costs, and reduced fees; and they manage liquidity on the customer’s behalf.

• Technology investment: looking at technology investment as simply a budget line is misleading. The more meaningful question is: how much of that budget is going toward genuinely improving and renewing existing technology, versus merely maintaining it? This is one of the key constraints weighing down legacy banks. Outdated software and aging hardware generate high maintenance and upgrade costs, and there is a ceiling on how far you can take them — you can only develop, optimize, and accelerate them so much. Newer banks (and here we mean not only digital banks, but also recently established banks in emerging markets) have a significant advantage in this regard. Building on modern technology and a software architecture designed from scratch carries a compounding edge.

• Data: We have come to realize that banks no longer just store our money — they store our data. And newer banks have a decisive advantage here too. Conventional banks may hold larger volumes of data, but “smart data” matters more than sheer quantity. Data that sits in siloed warehouses, unlinked, uncleaned, and uninterpreted, does more harm than good. To deliver “intelligent banking” to customers in an AI-driven world, the data architecture itself must be intelligent. New-generation banks are therefore building not just modern computing infrastructure, but data architectures designed from the ground up to be AI-compatible.

“Customer-centric product development and service delivery” has taken a leap forward in this new era. What banks must now offer is not just customer-centric service for segments — but customer-centric service for each individual customer: hyper-personalization. For example, a bank’s research team producing three or four daily market commentaries segmented by risk appetite — low, medium, high — will no longer be enough in the near future. The infrastructure must be capable of generating a tailored report and commentary for each individual customer seperately.

Banks’ profits have been high in recent years, but the structural questions outlined above suggest this is not a sustainable trajectory. And we cannot say that the alarm bells are still some way off — vulnerability is growing with every passing day. The pace of change in customer behavior is high, and technology adoption levels are already elevated. We should not take comfort from how slowly internet banking was adopted in the early 2000s. The demographic reality then was simply that adapting to new technologies was harder and slower. Smartphones and apps were genuinely new for that generation, and the learning curve was steep. Today’s population is already fluent in the technology it uses.

Another risk on the near-term horizon for banks is integration. Legacy banks may not disappear — but they may lose direct access to customers. Instead, they will find themselves providing infrastructure to the new-generation banks that do reach customers. In other words, conventional banks — weighed down by bureaucratic structures and heavy operational burdens — will end up serving fintechs and digital banks through APIs, lending out their networks and distribution channels. Customers will receive services as clients of new-generation banks, without even realizing they are using the infrastructure of legacy institutions. The old-model banks will become the invisible backbone supporting smart platforms — the plumbing of a network economy they no longer control. But do not confuse this “invisible infrastructure” with “invisible banking.” That is a topic for another