Revolut CEO likens bank lending to selling options: collect a decade's rent, lose it in one year
Lending profits are wiped out by economic cycles, so Revolut keeps lending under 10% of assets, securitizes the rest, and prefers to be a subscription tech company rather than a traditional net-interest-margin bank.
The video won't play here. Listen to the audio instead:
The argument · timestamps estimated from transcript position
Revolut's first seven years had zero growth budget
Storonsky says he never deliberately pursued rapid expansion, only constantly refined the product and solved problems end-to-end. Daily new customer growth climbed from single digits to hundreds to thousands, driven entirely by word of mouth and PR. The company set a formal growth budget for the first time in 2022, began measuring customer payback periods and lifetime value, and launched referral marketing products—the prior seven years had been pure organic growth.
— Nik StoronskyNew product validation turns on unit economics, not creativity
Storonsky says 90% of new ideas will make money if loaded onto the existing customer base, since team costs are far below customer lifetime value, so most ideas don't need strict validation. But real exceptions exist: Revolut launched Revolut Lite, an unlicensed cross-border payment product in 100+ countries, betting on low-cost organic growth. It didn't grow, forcing a shift to paid marketing, and the unit economics didn't work out, so it was killed.
— Nik StoronskyLending is like selling options: collect a decade's rent, lose it in one year
Storonsky believes every loan requires a reserve provision, and reserve size depends heavily on macroeconomic variables like unemployment. If unemployment rises from 3% to 6%, book profits could evaporate or flip negative overnight. Subscription businesses only face customer churn in downturns, not negative profit. This is why tech businesses typically trade at 40-50x PE while bank businesses trade at 10x PE. Revolut caps lending at under 10% of the balance sheet and securitizes everything else.
— Nik StoronskyProducts race by launch cohort; bottom performers get defunded
The company has accumulated historical data on roughly 100 new products and tracks each product's gross margin curve by month-of-launch cohort. Top 25th percentile products get team expansions, growing from one team to two, three, four. Bottom 25th percentile products lose funding and staff get reassigned. Storonsky says he's almost never seen a slow-starter turn around later—the first 12 months' trajectory nearly always holds, and that's the main basis for the company's investment decisions.
— Nik StoronskyAn LLM customer manager lives inside the cross-sell engine
Machine learning models score users based on transaction events and where they shop. For example, if a user lands in Ireland and buys an Airalo data plan, they get pushed Revolut's eSIM product. On top of that sits an LLM layer that acts like a real customer manager, weighing all candidate recommendations while balancing the ‘negative value of interruption,’ and decides when, to whom, and what product to recommend—the entire flow runs automatically, so Revolut essentially has no traditional distribution problem.
— Nik StoronskyCapture under-18 users and they naturally become lifelong customers
Revolut builds products specifically for minors. Once children become ‘Revolut natives,’ switching to an adult account at 18 faces far less friction than pulling adult customers away from traditional banks. Across nearly every European market, Revolut penetration among under-35 is now over 50%, but those 50, 60, 70 and older hold most of the liquid assets and are nearly impossible to migrate from traditional banking.
— Nik StoronskyAI could halve team sizes, but org structure isn't ready
Storonsky believes one engineer can now handle front-end, back-end, and data science simultaneously, so in theory a two-person team could match an eight-person team's output. But Revolut still organizes permissions and workflows by specialty—iOS, Android, back-end, data science—so it's hard for one person to get full system access and use AI to ship all code. His strategy is to keep running the old framework for now while experimenting with new approaches on the side, then switch once validated.
— Nik StoronskyEntering America means head-to-head competition with Amex and Chase
The US has historically been stingy with new bank licenses, and most players using bank partnership models have failed—fintech teams and bank compliance teams are essentially two companies forced together, with mismatched cultures and systems. Revolut chose to get its own license and has received conditional OCC approval, targeting remaining approvals next year. Entering the US requires a credit-card-driven product offering benefits at least as good as Amex or Chase, plus Revolut-exclusive lines like wealth management and crypto trading.
— Nik StoronskyIn their own words · checked verbatim
I mean, to be honest here, I was never optimizing to get scale quickly. I was just working on the product and then solving problems end-to-end.
Nik Storonsky0:00
I just had too many failures in the past when my judgment was wrong, so now I'm fully optimistic.
Nik Storonsky7:41
It's similar to selling options, right? You kind of collect premium for 10 years, and then year number 11, you lose it all, right?
Nik Storonsky17:24
Usually trend is your friend. The thing is growing very quickly in the first 12 months. It will continue to grow very quickly.
Nik Storonsky22:29
Technically, one person can accomplish much more now.
Nik Storonsky34:12
It's now, in terms of returns, top 3%, top 4% of funds in the world. The model really works.
Nik Storonsky45:56
Figures
| Revolut Business daily new merchant count | ~1,500/day | 2:43 |
| Target lending as % of balance sheet | Under 10%, rest securitized off-balance-sheet | 13:58 |
| Tech business vs. banking business PE comparison | Tech ~40-50x, banking ~10x | 15:00 |
| Payroll accounts as % of monthly active users | 10%-20% | 28:51 |
| Revolut penetration among under-35 in Europe | Over 50% | 30:07 |
| AI venture fund performance ranking | Top 3%-4% of global funds | 45:56 |
| AI fund investment count in past 2-3 years | 40-50 | 45:56 |
| Target IPO date | 2028 | 47:22 |
Glossary
- K-factor
- Measure of how many new users each existing user brings on average; above 1 indicates self-sustaining exponential growth.
- NPAP (New Product Approval Process)
- Revolut's internal governance and compliance approval process for new products before launch.
- OCC
- U.S. regulator responsible for issuing federal bank licenses.
How to listen
Fintech entrepreneurs, banking executives, and investors focused on new-bank business models and product organization methods.
After 48:41, discussion of quarterly performance reviews and lawyer KPIs is generic and lower in information density.