Nik Storonsky says Revolut will be a bank with "effectively zero risk." He started his career trading derivatives at Lehman Brothers. Storonsky told the FT today that Revolut will cap lending at 10-20% of deposits. It will sell the loans it does make, whole or through securitizations, what some call an "asset-light model" Revolut lends about 6% of its deposit funding base. A typical bank lends out closer to 100%. At the end of 2025, Revolut held GBP 50.2bn of customer balances and a GBP 2.2bn loan book. It kept 90% of its assets in cash and treasuries. --- This is a fundamentally different business model from traditional banks where the goal was "net interest margin." Take deposits, lend them out, and earn the spread. Profit depends on the current interest rate, and every new loan needs more capital. Revolut uses the license to hold your salary and savings. It earns from cards, FX, wealth and subscriptions. Storonsky puts Return on Equity (ROE) at 40-50% after stripping out excess capital. That is about double the best banks, and higher than Nubank. And you can see this in how investors are reacting. They valued Revolut at $115bn in July, around 68x 2025 net profit of $1.7bn. Barclays was worth about $95bn the same week on $12.9bn of net profit. --- "Zero risk" here means credit risk. Three risks stay with Revolut: 1. Rate cuts shrink interest income, which was GBP 974m in 2025 (22% of revenue). 2. Loan sales need private credit and securitization buyers to keep buying. 3. Fraud, AML and outages become the main ways Revolut can lose money or a license. Northern Rock failed in 2007 when the securitization market shut, because it funded itself there. Revolut funds itself with deposits. A shut market would stop its lending growth and leave its funding intact. --- If you hold a license or want one, decide what you want it for. Fee income grows without tying up capital, and investors pay tech multiples for it. Private credit funds want consumer loans and have no customers to lend to. Revolut has 80M. It can originate the loans, sell them to the funds and keep the fees. What makes this so interesting to me is HOW different it is from Nubank. Nubank is THE subprime lender in Brazil. It made that wildly profitable. It is a bank that lends. Revolut is in Europe, where interest rates are far worse, credit spreads are tighter, and it builds an entirely different machine. h/t @maxkarpis for the story find

The ECB's Pontes is LIVE. Banks can now settle tokenized securities trades in central bank money. President of the ECB Lagarde also pushed again for a digital euro to compete with stablecoins. This I'm a little more dubious about. 1. Pontes: Pontes acts as a bridge connecting market distributed ledger technology (DLT) platforms directly into TARGET Services (the Eurosystem's wholesale gross settlement engine). This follows the 2024 trials, which included 64 institutions. Pontes launches with standard operating windows before gradually expanding toward 24/7 settlement, with full rollout expected by 2028. 2. The Digital Euro ECB President Christine Lagarde also pushed hard for the Digital Euro legislation, which is now in final talks between the European Parliament and member states. She wants it done by December. Her aim is geopolitical. A digital euro is a defensive move against US private stablecoins potentially gaining more of a foothold on the continent, one that is heavily reliant on Visa and Mastercard today. This has preoccupied European leadership for a while, but became especially pressing in June, when the US Government used export controls to cut foreign nationals off from Anthropic's Fable 5 model overnight. Access came back 18 days later. Sovereignty over payments is sensible, especially when you consider that launching a EUR stablecoin to compete is pretty hard. Under MiCA, at least 30% of your reserves must sit in commercial bank deposits (60% once you're a significant issuer), and there's no "Eurobond" equivalent to US Treasuries. The business model for a European stablecoin isn't there. But the problem is, a sovereignty policy objective isn't always what the market or citizens want. Banks don't want a cash-like Euro for consumers competing with their deposits. Consumers who use stablecoins today would probably like to continue to have something as simple, 24/7 and flexible. If what they launch is essentially a closed-loop form of digital cash that can't be moved 24/7 across borders, you've launched something that doesn't solve the problem you're worried about. Which makes Qivalis even MORE interesting to me. 37 banks are building a EUR stablecoin that would work like a stablecoin, but could potentially connect to Pontes so the banks behind it settle in central bank money. Put another way, 1 EUR = 1 EUR. That hasn't always held for 1 USDC and a dollar. Europe isn't sitting on its hands. Pontes is fascinating, and I think the Fed could take a lot of notes from it. But I still don't get the digital Euro 🤷‍♂️

Anyone got an in with the X support folks? We lost access to the email connected to @TokenizedPod after a recent phishing attack got one of our staff. We contacted X Support immediately but haven’t received a resolution yet. We would really appreciate any help getting this in front of the right team. Please amplify so we can reach @Michaelo

Jev can outperform the rules many companies use to prevent fraud. We found that without any pre-training, it accurately detected 93% of a fraud ring. In contrast, an LLM in a similar set up only achieved 62%.

Jev vs. Claude vs. Rules: We ran a real fraud ring through all three. The headline result is that Jev classified best at 93% recall, followed closely by the velocity rules our Data Analyst Agent proposed at 90%, with an LLM making the classification decision trailing at 62%. This shows Jev is great at reasoning and classification jobs compared to LLMs. The fact that Jev was so close to a Data Analyst agent (which could write multiple SQL queries to investigate the fraud ring) is the real finding. Classification in code was already near-optimal, and Jev matched it. Read more in the article:

If you see this chart, know this is my perspective: the revenue growth path for OpenAI and Anthropic has weakened due to competition. And that competition is from *other* US models, not China / open source.

The backlog can be committed But if the GPUs aren't plugged into power the revenue never shows up And the growth slows down Which is exactly what eventually broke the housing market...

Banks moved dollars from Singapore to New York on a Saturday, over a blockchain @sytaylor just published the clearest breakdown of what banks are building to answer stablecoins, and it is worth your time 👇 - On September 5 Citi and DBS sent dollars from Singapore to New York in minutes, using tokenized deposits on Swift's new ledger. Still a pilot - They are not upgrading the core. One chief architect described 60+ markets, a separate mainframe in each, the code's authors long retired, and no full map of what depends on what - So they route around it. A tokenized deposit is a bank balance represented on a blockchain, still a claim on the bank, with a second way to move - Nothing actually travels. Bank A signs an IOU to Bank B on a shared ledger, Bank B credits the supplier on Saturday, and the two banks net out what they owe each other on Monday - The ledger is shared because no bank wants to trust another bank's database. Keeping parties that do not trust each other in sync is the one thing blockchains are good at - Citi's Services made $21.3B in 2025, JPMorgan's Payments $19.3B. That is the revenue all of this is defending Bullish?

Tokenized deposits are real. The banks are coming for 24/7 money movement. “Just use a database” misses the point. How tokenized deposits actually work, why banks want them, and where stablecoins still have the edge 👇

Stablecoins woke the banks up. Now Swift has a blockchain and banks are moving dollars on Saturdays. “Just use a database” misses the point. How tokenized deposits actually work, why banks want them, and where stablecoins still have the edge 👇

🚨 TODAY: The SEC issued an order granting temporary, conditional exemptive relief to Tokenized Securities Venues from the definition of “exchange” in the Exchange Act to trade tokenized NMS stock using innovative permissioned automated market makers and liquidity pools.