Behind closed doors, people are asking: Can digital banks really be profitable?
I get asked this because I have been CEO of two banks and overseen two digital bank initiatives, as well as been deeply involved in FinTech and AI for Financial Services.
This has led me to insights and views that are different that traditional espoused.
People look at digital bank CEOs leaving—a theme particularly pronounced in Malaysia at the moment—and they notice that whilst Revolut commands a $75 billion valuation following its November 2025 secondary share sale, it took nearly a decade to achieve consistent profitability.
Revolut reported its fourth consecutive year of net profit in 2024, posting $1 billion in net income on revenues of $4 billion. Impressive numbers, but the journey was long and capital-intensive.
The broader picture is equally sobering.
According to TABInsights’ World’s Top 100 Digital Banks Ranking 2025, whilst 61% of the top 100 digital banks reported full-year profitability (up from 48% in 2024), the global break-even ratio across all digital banks remains below 25%.
The time to profitability has shortened for leading institutions—from five to six years historically to approximately two to three years for well-positioned second-generation digital banks—but this masks significant variance in outcomes.
Based on my experience and keeping an ear to the ground, add on a year or two to those estimates.
So are digital banks just for very long-term horizon investors seeking exposure to financial services?
Or can they genuinely deliver better solutions to existing customers and the underserved whilst generating economic value for shareholders?
The answer is both uncomplicated and complicated. Let me explain.
The Uncomplicated Answer
The uncomplicated answer is this: it does not make sense for conventional capital investors in an environment of significant profit-seeking to pursue digital banking, particularly is the value proposition is primarily for the underserved – which is typically required in obtaining a license.
The underserved community is not well understood, typically higher risk, and the nature of capital required cannot be one fixated on quarterly or yearly metrics.
That structural capital issue needs resolving, but we are not going to solve it here.
Despite this, yes, digital banks can deliver value to all parties if you understand banking and technology well enough.
With that dual understanding, you can absolutely deliver a business case on a typical three-to-five-year horizon that works for customers and creates a strategic long-term asset.
The Complicated Answer
To understand why this has seldom happened, we need to address three critical areas: Sales and Distribution, Proposition, and Talent.
If you address these effectively, even in a primarily profit-seeking system, digital banks can thrive and compete.
Before examining these factors, an important prerequisite: alignment between board, shareholders, and regulator.
If these parties are not aligned on expectations for success, even perfect execution on the three critical factors will be doomed.
That is a separate discussion for another article.
For present purposes, I assume—and it is a substantial assumption typically not borne out in reality—that this alignment exists.
Sales and Distribution: The Foundation
In my interactions with those involved in digital banking, there is often excitement about product and proposition
“Build it and they will come.”
Unfortunately, it seldom works that way. The reality is:
“Sell it and you’ll make money.”
Without a sales and distribution function, there is no business.
Digital banks face significant barriers: brand recognition, trust, top-of-mind awareness, network access, and meaningful customer engagement.
These barriers are particularly acute because financial services products are fundamentally a utility.
They do not define you the way a car or fashion brand might. Banks need to understand this reality.
The economics are stark.
Research indicates that customer acquisition costs (CAC) for traditional banks range from $150 to $350, whilst neobanks typically achieve $5 to $15 through purely digital acquisition channels.
However, this advantage evaporates if you lack an engaged customer base to begin with.
The successful cases in the short term have come from consortia where at least one member possesses a strong, engaged, and relevant existing customer base for distribution.
Without this, digital banks take an extremely long time to reach profitability because both acquisition and maintenance costs remain elevated.
Proposition: Three Paths to Viability
When it comes to proposition, there are only a few ways digital banks can deliver sustainable business activity and profitability.
It is not simply about great onboarding experiences or frictionless transactions. There has to be something more substantive. I see three genuine paths.
First, be credit-first. This means enabling people to access credit faster—not necessarily cheaper or better, but faster—and extending credit to a wider group.
If you can generate returns on financings, the rest of the business can follow. It matters less whether you have liabilities if your assets are growing at high margin. But this requires the capability to bank more difficult credits.
To deliver on this, you must address credit assessment differently—enabled by technology rather than traditional methods.
Around the world, institutions have deployed AI and machine learning to create scorecards capable of serving the underserved or subprime effectively.
Utilising these capabilities to book a high-margin but appropriately-priced risk book is one viable path to growth and profitability.
But you need to know what you are looking for in selecting the institutions that can help with this and unfortunately the typical names or those who shout loudest are not the ones.
So even if one is able to understand this issue, the potential solution, one eye is not enough, a bank needs someone with two eyes to help make the right choice.
Second, develop a best-in-class salary proposition.
You are essentially selling to individuals and HR departments a solution that allows employees to utilise their incoming salary in innovative ways: taking salary weekly, in different proportions, maximising financial flexibility.
This approach enables rapid customer growth and deposit accumulation. Whilst a salary proposition alone may not deliver profitability, it creates an engaged customer base for cross-selling.
The pathway to success on this is also incredibly important, you need folks who have been in the weeds for HR operations outside of financial services to be able to ensure value propositions hit both the hygiene factors as well as the differentiating factors.
Third, pursue an asset management-led approach.
To the extent digital banks can offer a suite of best-in-class asset management solutions, they attract customers interested in investments—along with their deposits.
This approach generates fee income rather than interest income, carrying significantly lower capital charges.
The complexity lies in requiring partnerships with asset managers, and the products must be genuinely competitive.
Digital banks in emerging markets can differentiate by partnering with asset managers employing AI to manage products and performance—an approach not yet widely adopted.
This approach is more important than just getting to profitability ever so let me spend a little more of your time on it.
Consider the reality facing young people today. They see stories of cryptocurrency millionaires, meme stock winners, and overnight fortunes.
What they do not see are the countless others who lost everything. Survivorship bias is powerful: you only hear about the winners.
Combine this with the entirely reasonable aspiration of young people to start building wealth early—which is genuinely a good thing—and you have a generation eager to invest but lacking the guidance to do so wisely.
Here is something most young people do not know.
In 2007, Warren Buffett made a $1 million wager that a simple S&P 500 index fund would outperform a portfolio of hedge funds selected by some of the brightest minds on Wall Street over a decade.
By 2017, Buffett had won decisively: the Vanguard S&P 500 Index Fund returned 125.8%, whilst the hedge funds averaged just 36.3%.
Buffett’s point was clear—massive fees charged by professional managers often leave clients worse off than if they had simply invested in a low-cost index fund.
As Buffett himself wrote in his 2017 shareholder letter:
“Performance comes, performance goes. Fees never falter.”
So should young people just index?
Here is where it becomes interesting for digital banks that should in theory be adopting and partnering with leading FinTechs, leveraging capabilities for customers.
Whilst traditional active management has consistently failed to beat passive indices, emerging research suggests that AI and machine learning can change this equation.
A Stanford Graduate School of Business study found that an AI analyst, using publicly available information, outperformed 93% of mutual fund managers over a 30-year simulation—by an average of 600%. A 2024 SEC report indicated that funds deploying AI-driven strategies outperformed peers by an average of 12%.
Research published in 2025 found that mutual funds in the top 10% of machine learning adoption generated annual risk-adjusted returns 2.4% to 3.0% higher than bottom-tier funds.
AI-led hedge funds have produced cumulative returns of 34% compared to 12% for the global hedge fund industry over comparable periods.
This creates a genuine opportunity.
Banks can play a role beyond providing typical banking accessibility to typical products.
They can start young people early on the journey of generating and preserving wealth—not through speculative gambling, but through intelligent, AI-enhanced investment strategies that have demonstrated the ability to deliver superior risk-adjusted returns, and outperforming indexing.
There is a way to educate the young, give them options, and provide access to wealth creation and preservation through investments that are safer and smarter than what they might otherwise pursue.
Young people educated on the benefits of indexing, helping them do that and also giving options for AI enhanced returns for any given risk profile.
For Islamic financial institutions, this is not merely an opportunity—it is an obligation.
The maqasid al-Shariah, the higher objectives of Islamic law, include the preservation and growth of wealth (hifz al-mal) as one of the five essential elements.
Islamic finance exists not simply to replicate conventional products in a Shariah-compliant wrapper, but to serve the community it represents in achieving genuine financial wellbeing.
Whatever primary approach an Islamic digital bank might take—whether credit-led, salary proposition-led, or otherwise—I believe there is incumbent upon Islamic financial institutions to also incorporate an asset management-led approach as a natural extension of their purpose.
This is not a commercial add-on. It is fundamental to supporting the principles of Islamic finance and the community it serves in growing and preserving wealth across generations.
Regardless if the proposition is credit-led, salary-led, asset-manager led or even some type of hybrid, they require robust sales and distribution arrangements to succeed.
Without embedded finance partnerships for asset growth, without consortium influence for salary propositions, without distribution networks for asset management—all of these approaches will struggle.
Proposition and distribution are inextricably linked.
Talent: The Most Complex and Most Controllable Factor
The final factor—and possibly the most complex yet most controllable—concerns people.
Digital banks face a fundamental challenge: where do you source a leadership team that understands both banking and technology at sufficient depth?
Even in traditional banks, successful bankers in business, risk, or finance find their way to CEO positions, then must learn areas outside their expertise.
I have known corporate and investment bankers who had to reinvent themselves as retail bankers because that is where the money originates.
The same dynamic applies to digital banks, but amplified: your choice is typically a banker or a technologist. For digital banking, you need someone who knows both deeply.
Going back two decades, people pursued technology routes or finance routes separately. The notion that you would need both skill sets in a deep way was seldom understood.
Consequently, people simply do not have both. If you are a shareholder, which do you choose? The banker will struggle to navigate essential technology decisions.
The technologist will struggle with banking requirements. Right at the top of the house, there is a talent gap that hampers execution—even with the right distribution arrangements and propositions.
This is compounded at the leadership team level.
Beyond the CEO, you need executives who understand both banking and technology in material ways.
When this gap exists throughout senior leadership, you get suboptimal decisions—not from bad intentions, but from knowledge and skill gaps.
What follows is heavy reliance on consultants who, whilst institutionally possessing success stories, are typically filled with people who know the business of consultancy rather than the business of building a business.
To mitigate risk, organisations engage the most well-known brands at premium prices, often without commensurate value delivery.
By the time people recognise the skills gap, they are already deep into cost and spend, with material impact on the capital base.
Pressing the reset button requires a new approach and more money.
The Agentic AI Opportunity
Here is where things become interesting.
Even with someone who understands both technology and banking, they cannot do everything. They rely on teams, typically large and expensive teams, and when those teams lack the same balance, execution remains problematic.
However, developments over the past twelve months in agentic AI could mitigate this risk very significantly.
According to BCG research, agents already account for 17% of total AI value in 2025 across industries, projected to reach 29% by 2028.
BCG’s Global Retail Banking Report identifies more than $370 billion in annual profit potential from AI by 2030. McKinsey’s analysis suggests that agentic AI could lower operational costs by 20% or more, equivalent to 9% to 15% of operating profits.
Early implementations indicate potential productivity gains exceeding 60%.
What does this mean for digital banks?
Essentially, agentic AI solutions can empower bankers to have agents undertake tasks and execute through systems, removing reliance on technology and operational teams.
If given the choice between bankers building a digital bank or technologists building one, you would choose bankers—and agentic AI makes this possible.
For regulated entities, the critical requirement is agents that do not hallucinate when performing tasks. Everyone has experienced hallucinations with large language models.
To control this, there needs to be a neuro-symbolic nature in the AI tool being deployed.
Through my advisory work with Syngentix in Singapore, I have seen this capability firsthand: their system empowers banking executives to build agents to perform work within a bank.
I have personally built agents using this platform.
If a CEO or senior management team can build agents, you can develop and build a bank without dependency on technology and operations teams.
Digital banks typically adopt vendors based on newer technology architectures, making it easier for AI agents to interact with these systems.
Today, if starting a digital bank, you would likely take one main vendor providing your core system, then add agentic AI capabilities.
That combination becomes your technology and potentially operations function.
Training people to use these systems—many of which offer conversational interfaces or no-code/low-code tools—becomes the focus rather than building large technical teams.
The capabilities now cover marketing, product, operations, compliance, and risk. It is possible with a team of thirty people to develop an entire working digital bank to regulatory standards.
This changes the economics fundamentally.
The Malaysian Context
What of digital banks today, particularly in Malaysia where I am based?
We have witnessed a steady stream of senior executives stepping away from their roles—CEOs, CTOs, familiar names who helped bring these digital banks to life.
At Ryt Bank, the CEO exited within the first year of operations. KAF Digital Bank saw its CEO change hands. GXBank has experienced leadership transitions as it shifts focus towards profitability and cost discipline, and at the time of writing it has been announced that the AEON Bank CEO also moving on too.
This is understandable.
Setting up a digital bank is hard work at a pace that cannot be endured indefinitely.
It is complex and involved based on how digital banks have typically been assembled. It usually costs more than anticipated.
Senior leadership feels pressure. Shareholders place pressure for better performance on investments. Executives who lack both banking and technology experience at sufficient depth—beyond high-level understanding—need to comprehend the implications of decisions from an execution and operational perspective.
Anyone falling into that category will not be having fun, and after reaching certain milestones, looking to move on becomes natural.
What typically happens next?
Replacements fall into predictable categories: another banker, another technologist, or another trusted person brought in hoping they can make the trouble disappear.
But fundamentally, what digital banks’ leadership requires—directed by boards and shareholders—is the capability to understand how to press the reset button when things are not working, to identify talent capable of navigating banking and technology, to leverage AI’s enabling potential (the correct types of AI), and to achieve certainty with sales and distribution.
What would I do if I were a shareholder of one of a digital bank?
The answer to that requires an article all on its own.
The Path Forward
Digital banks that have already been issued licences and have undertaken setup or market activities over the past three to five years will likely require some type of reset.
Otherwise, they will be throwing money at models that will not work, or making incremental changes insufficient to address compounding issues.
Part of that reset might be to sell. Part might be redefining a new operating model. The latter requires people with experience and capability to execute.
In this article, I have outlined various paths to successful digital banking.
However, a critical factor, regardless of which model you adopt, is execution capability. One of the challenges in emerging markets is talent, capability, and knowledge to navigate this landscape.
Existing partners and vendors can be leveraged effectively. There is no reason why an institution cannot succeed in delivering a better operating model.
But it becomes fundamentally a people issue: having the courage, conviction, and experience to deliver on operating models with higher probability of success.
My prediction: the majority of digital banks will either sell or scale down their ambitions.
From board level—both independent directors and shareholder representatives—there simply will not be sufficient knowledge spanning market understanding, network, influence, banking, and technology to execute successfully.
The ultimate losers are underserved customers.
Most digital banks were established to address gaps in service to the underserved and those not being served properly. Digital banks, despite being commercial in nature, can still support these segments.
But I doubt it will happen in any material way based on a business as usual approach.
There will be success stories, certainly—but success stories capable of scaling? That I highly doubt, unless there is fundamental change in operating model.
There is another option to truly support the underserved—one requiring considerable courage. But that is also an article for another time.
Featured image: Edited by Fintech News Malaysia based on an image by smth.design via Freepik.
