Tech Giants Rewriting Everything or Incumbent Banks Finally Rebooting Themselves?

This article is a synthesis of where some of the analytical minds across global banking industry currently stand - and why neither side has yet won the argument. If you are reading this as a practitioner, you are probably living one of these two theses. The discomfort is the point.

The Question Nobody Can Honestly Answer

Speak with any intellectual banker who plays on the int’l market, and you will likely find two types of people. The first speaks with quiet urgency about structural tech disruption - about distribution platforms swallowing customer relationships, fintechs continuously inventing business models addressing banks’ weaknesses, about AI architectures that make legacy banking look like a good old typewriter. The second gestures to their return on equity, their loan book quality, their deposit franchises built over decades, their client (especially corporate) trust and finally notes that the demise for traditional banking has been written many times before, hence no worries at all.

Both are right. Both are also dangerously incomplete.

This article aims to be the first in my series of four or five, built around some of the foundational questions that matter most for anyone with a stake in how banking might evolve over the next few decades. Of course, I’m not the one to tell you which side will win, I’m just as much climbing the practitioner’s learning curve as many curious others. But I’ll try to lay out some possible versions of each argument - with the numbers and some examples behind it - and leave the conclusion to the reader.

One warning before we begin: the most comfortable answer is probably the wrong one. And if you run a bank - or are trying to disrupt one – the title question is not academic, and neither am I. But maybe it could represent some sort of framework for every major decision your institution will make over the next couple of years.

"The threat from third-party agents could be material. Bank profit pools globally could decline by $170 billion, or 9 percent." - McKinsey Global Banking Annual Review, October 2025

The Fintech Wins Thesis: The Challengers Are Already Inside the Gates  

Let’s start with Nubank, which I honestly admire. In 2024, David Velez's Brazilian digital bank crossed 100 million customers - making it the largest financial institution in the Americas by customer count, surpassing banks that have operated for over a century. It achieved this in eleven years, with no branch network, no inherited infrastructure, and no legacy cost base. Its cost-to-serve is a fraction of the incumbents it competes against. And it turned genuinely profitable - not on an adjusted, excluding-this-excluding-that basis, but in the way that matters to long-term investors. It’s getting ready to conquer the US market now, behind a soon-to-be full banking license.

Nubank is not an emerging market curiosity. It is proof of concept for an entirely different theory of what a bank can be: built on a mobile-first architecture, growing through word of mouth rather than branch density, and acquiring customers at a cost that makes traditional retail banking economics look structurally broken. Warren Buffet understood this when he invested $500 million in 2021, not to bet on Brazilian digital banking specifically, but I guess to bet on the operating model.

Another Example:  Revolut, founded in London in 2015, posted £3.1 billion in revenue in 2024, growing 72% year-on-year, with £1.1 billion in pre-tax profit. At today’s 65 million customers, it did this with an IT cost structure and a product velocity that no incumbent bank of comparable customer scale can match. Nick Storonsky built it without a full UK banking license for most of its life, relying on an e-money license while extending into FX, stock trading, crypto, and travel insurance, but basically Revolut grew because the traditional banking failed to serve the travelling clients the way they should. When Revolut’s banking license finally came, it was an upgrade to a platform that already worked. And things can get very interesting from here on.

McKinsey (2025) describes the structural shift bluntly: cross-industry platformization has effectively eliminated size as a banking advantage. The challengers do not need to out-capital the incumbents. They need to out-distribute them. And distribution, in a smartphone-native world, belongs to whoever owns the daily relationship. Superior distribution will beat superior product, just ask Peter Thiel (book “Zero to One”).

Consider Kaspi, Lomtadze’s eternal legacy. The Kazakhstani super-app, founded as a payments platform, now a fully licensed bank with e-commerce, marketplace lending, and government services embedded, serves the majority of Kazakhstan's adult population through a single application. It has achieved this not by being a fintech that disrupts banking, but by being a platform that makes banking invisible. The user opens Kaspi to pay the utility bill, buy a flight ticket, or split a restaurant tab. The credit product, the savings account, the insurance policy - these are features of a daily-use platform, not the reason someone opens the app. This is actually the key distinction.

Klarna tells a similar story from a different angle. What began as a buy-now-pay-later product sitting between consumer and retailer has evolved into a financial services platform with 85 million active users across 45 countries. The short-term credit product created the distribution; the distribution is now being monetized across banking, insurance, and shopping. The incumbents that Klarna competes with in consumer credit were not disrupted by a bank, they were disrupted by a checkout button. Ok, it faces challenges now as the duration of BNPL products extends and starts to resemble a true consumer loan, but once you own the client, you can fix things along the way.

Accenture's 2026 outlook is unambiguous on where this all leads: by 2030, the largest retail banks may not be banks. Non-traditional retail/SME players will reshape the landscape, forcing incumbents to innovate or become invisible infrastructure: powerful, regulated, capital-intensive, but with less brand appeal from the customer's perspective.

The invisible infrastructure risk is perhaps the most psychologically difficult scenario for my senior colleague bankers to grasp. It does not mean losing customers per se. It means slowly being relegated to the role of licensed utility: providing the balance sheet, absorbing the compliance costs, taking the credit risk - while the customer experience, the brand, and most important of all, the data, sit with a platform that knows far more about the customer's life habits than any bank ever did. And we all work on all kinds of CRM systems integration.

"By 2030, the largest banks may not be banks." — Accenture Top 10 Banking Trends, January 2026

The data migration is namely already happening. In USA, fintech platforms now account for nearly 50% of US unsecured personal loan originations, up from zero a decade ago. A quarter of Americans are using BNPL services for groceries. Not electronics or luxury goods, but groceries. The credit intermediation function, the most fundamental commercial activity of a retail bank, is migrating, in real time, to non-bank actors. This is not a forecast. 

Have a look at Africa and the M-Pesa phenomenon - the mobile money platform originally launched by Safaricom in Kenya, now operating across seven African countries. It has demonstrated something even more radical: you do not need a bank at all to deliver the core functions of banking to people (retail) who have never had a bank account. M-Pesa processes more transactions in Kenya than the entire formal banking system combined. It has not disturbed the banking industry; it has actually replaced it for an entire client segment.

The Incumbents Prevail Thesis: Reports of Our Death Have Been Greatly Exaggerated 

In 2025, we were reading the reports which received considerably less attention than the disruption narratives: traditional banking staged a genuine comeback. Despite a decade of predictions that neobanks would commoditize the industry into oblivion, incumbent banks delivered historically highest profits. ROE across the major banks hit levels not seen since before the financial crisis. We swim in profits.

Andrea Orcel's UniCredit is perhaps the most striking European example. Orcel has driven a transformation that the disruption stories tend to ignore: aggressive digitization of an incumbent balance sheet, combined with cross-border M&A ambition that no neobank could contemplate. UniCredit's 2024 results showed net profit exceeding €9 billion. One can hardly see this as a performance of a dinosaur, it’s simply the performance of a large bank that decided to modernize on its own terms, at its own pace, with its own capital.

Another Example:  OTP Bank - the Hungarian incumbent that has systematically expanded across Central and Eastern Europe through acquisition only - has built a cross-border banking franchise that challengers cannot replicate anymore. It operates in 11 countries, serving 20+ million customers. Its strategy has been consolidation, not disruption - buying weaker incumbents, more or less integrating them onto a common platform, and extracting scale efficiencies that its fintech competitors cannot do without a balance sheet of equivalent size.

Accenture frames the incumbent advantage very simply: scale is emerging as the ultimate competitive factor. By 2030, the largest banks will leverage unmatched efficiencies and global or regional reach to outpace all competitors. Running a trillion-dollar loan book, absorbing regulatory capital requirements across a whole lot of jurisdictions, managing sovereign risk and systemic liquidity: these are capabilities that take decades and enormous capital to build.

Ana Botin's Santander provides a different illustration of the same point. The Spanish incumbent has built what is effectively a distributed digital banking operation: Openbank in Europe, a digital-first consumer bank in the US, and significant digital banking infrastructure across LatAm. And she is doing M&A quite successfully as well. All without abandoning the balance sheet or the regulatory relationships that underpin its competitive position. Santander did not need to become a neobank. It needed to build neobank capabilities while remaining a bank. 

Then there is DBS. Piyush Gupta - who ran the Singapore-headquartered bank from 2009 until his recent retirement - arguably executed the most complete incumbent digital transformation of any large bank anywhere in the world. DBS was named the world's best digital bank multiple times not because it disrupted itself, but because it genuinely rebuilt its technology architecture from the inside out: cloud-first, API-enabled, with an AI capability that runs across lending decisions, fraud detection, and customer service. Its cost-to-income ratio fell consistently over a decade while its ROE rose. As Gupta put it: the question was never whether to transform, but whether to do it at sufficient pace.

The regulatory moat is real and deepening. After all the recent stress events regulators across the world are tightening capital requirements, stress testing frameworks, and operational resilience methodologies. Every tightening of regulation that a challenger bank must now absorb as a growth-stage company, an incumbent has already absorbed and operationalized. Central bankers know that regulatory framework exists to ensure financial stability, and banks that have demonstrated they can operate through cycles carry a legitimacy premium that is not easily replicated.

Zafin's 2025 analysis articulates the modernization path that is working: progressive decoupling. Not the enormously expensive and deeply risky “shred and replace” of core banking systems, but a surgical extraction of specific capabilities into modern cloud while leaving the regulatory-grade processing engine intact. The banks modernizing carefully, layering new capabilities on proven foundations, are delivering results without betting on a single transformation program.

Lion Finance Group and its peer institution TBC, provide another compelling European evidence for the incumbent transformation thesis. Both Georgian bank groups operate in markets where the competitive dynamics of fintech are intense and the customer base is digitally sophisticated. Both have built super-app banking experiences around their incumbent charters, achieving cost-to-income ratios and digital adoption rates that most WE banks cannot easily match. They did this not by becoming a technology company, but by becoming a bank that thinks like one. Ok, and they employed some real smart people.

"Quality endures - companies with strong balance sheets and diversified revenue delivered strong returns." - Banks & Bankers, January 2026

The Third Path: What Evidence Actually Points To

Here is where the binary logic fails and where the most interesting strategic thinking is currently happening.

McKinsey's most provocative analysis proposes that the universal bank model is structurally over, and that the $20 trillion value creation opportunity in banking will flow to various types of specialized platforms. But the most compelling real-world evidence points to a different synthesis, what analysts call the combined model: a digitally native licensed bank, embedded inside an ecosystem, that merges regulatory legitimacy with platform-scale distribution.

Bank Jago in Indonesia is the example of that model. Jerry Ng's institution was a small regional bank before it embedded itself entirely inside the Gojek super-app. That’s a platform used by tens of millions of Indonesians for ride-hailing, food delivery, and digital payments. In three years, it went from obscurity to one of the fastest-growing digital banks in Southeast Asia. It did not need to acquire customers: Gojek's 38 million monthly active users were already there, already transacting, already generating the behavioral data that makes credit decisions possible. The bank provided the license and the balance sheet. The platform provided relationships and distribution.

Example: Toss (South Korea): the Korean super-app built by Lee Seung Gun — began as a peer-to-peer money transfer application in 2015, when Korean banking was famously cumbersome and over-regulated. By 2025 it had evolved into a full-spectrum financial platform: Toss Bank (a licensed digital bank), Toss Securities (a brokerage), Toss Insurance, and a payments layer used by millions of daily active users. Revenue exceeded $1 billion. The model proves that the combined approach - own the distribution, then earn the license - can work at scale even in highly regulated, bank-dominated markets.

Example: Mercado Pago (Latin America). The financial services arm of Mercado Libre - Latin America's dominant e-commerce platform - has become one of the region's largest financial institutions without ever positioning itself as a bank. It provides credit, insurance, investments, and payments to 50+ million active users across 18 countries, using transaction data from the marketplace to underwrite credit that traditional banks could not price accurately. The e-commerce platform created the credit bureau. The credit bureau made the bank possible. There is a pattern where Mercado Pago and Kaspi cross.

The pattern that emerges across these cases is consistent: the combined model, platform distribution at scale, plus licensed balance sheet ownership, appears to command the highest long-term valuations. Nubank trades at multiples that no traditional bank achieves while simultaneously reporting genuine profitability. Kaspi's market capitalization has exceeded that of several major European incumbent banks despite operating in Kazakhstan and being discounted for that. The capital markets are pricing the combination of distribution and balance sheet more generously than either alone.

But the questions that neither side answers clearly are mostly still tied to profitability, because on the challengers’ side it is harder than it looks for most of them, as the unit economics of customer acquisition look compelling for challengers until you model the full cost of risk management, regulatory compliance at scale, and the customer service infrastructure required to retain rather than just acquire. On the incumbents’ side the question is not whether they are profitable today, it is whether they are building the future capabilities that will make them profitable in a structurally different competitive environment. Because when adding the AI component, the game is altering quickly, as the analysis highlights a threat that is considerably more uncomfortable than cost-cutting alone: AI agents that act autonomously on behalf of customers, directly against bank revenue.

Imagine a GenZ grown-up whose phone-embedded AI agent continuously scans every savings product, every current account offering, every investment product on the market and automatically switches provider when a better option appears. Zero switching friction. Zero brand loyalty. Zero relationship inertia. The same AI capability that reduces a bank's operating costs by 30% to 50% could simultaneously compress its revenue margins by eliminating the information asymmetries and switching friction that have historically been the silent subsidy of retail banking profitability. 

Take my friend’s Coen Jonker's GoTyme Bank, operating in South Africa and the Philippines. It has arguably pushed further on this than most: underwriting credit in real-time using behavioral and transactional data, with AI-driven credit decisions that cost a fraction of traditional underwriting and reach customers that traditional credit bureaus cannot score. He also had some incredible people working alongside him to get there. When AI makes credit almost instantaneous, the distribution advantage of the bank branch network diminishes.

The Question Remains Open - Which Is Exactly Why It Matters

The synthesis from multiple credible analytical sources does not converge on a single winner. It converges on a single observation: the institutions most likely to own the next three decades of banking are those that combine data scale, AI-first architecture, and regulatory capital. The question is which type of bank gets there first and whether the combination is most naturally achieved by platform companies acquiring regulatory capability, or by incumbent banks acquiring genuine technological ambition (basically the right forward looking employees) and/or entire digitally native players.

What is clear is that the answer will not be the same in every geography. The combined model that works in Indonesia may not be the model that works in Serbia or Egypt. The AI advantage that a JPM can buy with $12 billion of IT Capex is not the same AI advantage available to a mid-size regional bank in Central Europe. Which means that the relevant question for any bank is not “who wins globally?” but “who wins in my market, with my customer base, against my specific competitive set?”

The comfortable answer, that banking will look roughly like banking today, with some digital features added, is almost certainly wrong. The range of outcomes is large. The window for making the strategic choices that determine which side of that range you end up on is narrowing.

"The era of the monolithic universal bank is structurally over. The question is what specialization model each institution chooses." - McKinsey, 2025

In the next article in this series I will try to take on the question that many bankers will find even more unsettling: what actually happens to the deposit-loan model -> the 800-year-old foundation of banking, in a world of platform finance, embedded credit, and near-zero switching cost? If the answer to this article feels unresolved, the answer to the next one could be even more eyebrow lifting.

Sources:

McKinsey Global Banking Annual Review (Oct 2025)

Accenture Top 10 Banking Trends (Jan 2026)

PwC Next in Banking & Capital Markets (2025)

HSBC Innovation Banking Fintech Horizons (2025)

Zafin Banking Modernization (2025)

Vega IT Fintech Trends 2026

McKinsey Agentic AI in Banking (Nov 2025)

Journal of Financial Services Research FinTech Score (Sept 2025)

BCG Global Wealth Report (2025)

CTO Magazine Neobank 3.0 (Jul 2025)

McKinsey $20T Banking Breakup (2022, updated 2025)