Revolut has a conditional OCC Charter, 80m customers, and if you ask any European bank CEO is absolutely a threat to mainstream banks. They’re now coming to the US, and their US CEO will be in a live fireside chat at Fintech Nerdcon. This show is stacked. By far the best speaker lineup of the year, and we haven’t revealed everyone yet. People will talk about Nerdcon ‘26 for years to come, I promise you. Do not miss it; grab your tickets here.
Things to know 👀
The Information reported that Anthropic is hiring to build billing, fraud, and treasury tech in-house, and framed it as a threat to Stripe. Cue the entire timeline saying this is the PaymentsPocalypse and Stripe is cooked. I don't buy it.
🧠 The postings describe a merchant billing and risk team. Pricing primitives, reconciliation, promo abuse, chargeback thresholds, run "while leaning heavily on third-party providers such as Stripe."
🧠 Shopify, Uber, and DoorDash all started on Stripe and went multi-PSP. This is expected behavior. You go multi-PSP for better auth rates, lower cost, and more countries, and OpenAI already added Adyen alongside Stripe.
🧠 Remember the SaaSpocalypse? The market said Claude would eat Salesforce, then Salesforce jumped 22% on a partnership, because years of licenses and enterprise relationships are still a moat.
🧠 My read: this is expected behavior at ~$45B annualized revenue, and Anthropic ends up keeping Stripe, adding a second PSP and building its own billing layer on top.
The OCC granted Revolut Bank US preliminary conditional approval on Thursday, six months after it filed. Revolut still needs FDIC deposit insurance, Fed approval, and a final OCC sign-off, and it's targeting a 2027 launch. Today, its US customers sit on Lead Bank. With its own charter, Revolut gets FDIC-insured deposits, credit cards, loans, stablecoins, and direct access to Fedwire and ACH, in all 50 states, under one federal regulator.
🧠 The same regulator rejected Wise in July and Bunq in August. Wise fell on AML controls. Bunq fell for making European delinquency assumptions on a US card book, a board short on US card experience, and capital the OCC couldn't pin down.
🧠 The UK took four years. The US took six months. Overseas regulators lean on the home regulator. Once the PRA granted the full UK license in March, the US application went in the same week.
🧠 Revolut already runs as a licensed bank in 30 of its 40 markets. Lithuania gave it an EU banking license, passported across the EEA. The UK license landed in March. Mexico went live as a bank in January. Argentina came via the Banco Cetelem acquisition. Peru, Brazil and Colombia are in flight.
🧠 The thing crossing the Atlantic is a $6bn revenue, $2.3bn profit machine. 80m+ customers, 38% pre-tax margin, $67.5bn in customer balances, and 11 product lines each doing over $135m a year.
🧠 Revolut doesn't launch features. It launches business lines. FX, then trading, then crypto, then business banking. Each one became a nine-figure business. (Yes, AI wrote that bullet, but I stand by it in the edit)
🧠 That’s what makes them so different. If you judge them as a bank, their economics look weird. If you judge them as a consumer acquisition machine that cross-sells a ton of products banks never could, then they’re a real threat.
🧠 The US is a harder fight than Europe. Branches still matter. The card market is the most rewards-heavy on earth and incumbents will happily outspend a newcomer on acquisition.
🧠 The Revolut credit card is actually awesome. It comes with WeWork access, 1:1 points conversion in 30 airlines, unlimited lounge access (with partner lounges) plus one guest, and full medical and cancellation insurance(!). They did this for a lower fee than Sapphire, don’t require you to spend with partners, and use European interchange (0.3%). Imagine what they could do with US interchange (2%)
🧠 Every bank CEO I speak to in Europe tracks what Revolut ships, because it scares them. I've said it before: Revolut rarely announces without executing. It’s taking meaningful share in home markets from Europe’s largest banks. Do not bet against them doing that in the US. Slowly, then all at once.
🧠 The US is about to find out what Revolut is. And if you want to ask the person who has to build it, Cetin Duransoy, CEO of Revolut US, is on stage at Fintech Nerdcon.
MoneyLion founders Dee Choubey and Rick Correia came out of stealth with preliminary conditional approval to charter OpenReserve Bank, N.A. in Salt Lake City. The plan is institutional treasury, tokenized deposits, and a GENIUS-compliant stablecoin (rUSD) issued by a bank subsidiary. Backers include Coinbase Ventures and Wintermute.
🧠 The OCC wants $210M of paid-in capital and a 12% Tier 1 leverage ratio. OpenReserve has 12 months to raise it or the approval lapses. The seed round is $25M.
🧠 Coinbase, Circle, Paxos, Ripple and Bridge all applied for trust charters. Trusts custody assets and cannot take insured deposits or lend. OpenReserve applied for the full bank so it can do both.
🧠 Choubey told Bloomberg the bank is built to settle when Fedwire is closed. He also said it will accept other GENIUS stablecoins as collateral. So does collateral mean a stablecoin issuer could pledge some fraction of the daily settlement volume or would it be 100% prefunded?
🧠 The investors are exchanges and market makers. Silvergate SEN and Signature Signet served the same firms with 24/7 dollar settlement. Both closed in the same week of March 2023.
🧠 Erebor got its full-service charter in October for tech companies and their founders. OpenReserve is the first de novo national bank whose stated business is stablecoin issuance and tokenized deposits. Although I suspect these two banks to compete for that category in time.
🧠 MoneyLion went public by SPAC in 2021 at $2.4B, hit profitability in 2024, and sold to Gen Digital for ~$1B in April 2025. That team has filed 10-Ks and operated on a partner bank charter for a decade. The OCC's management test rewards exactly that. There aren’t many management teams with start-up and bank experience.
🧠 A stablecoin issued by an insured national bank can sit on a bank balance sheet, be lent against, and settle into tokenized deposits at the same institution. Circle cannot do any of those inside one entity. I think that combination is the whole reason for the full charter.
🧠 $25M raised against a $210M capital condition, due within 12 months. No de novo charter this year has come with a bigger gap between the money in and the money required.
Félix lets Latino immigrants in the US send money home by typing (or voice-noting) in WhatsApp; the only step outside the chat is a secure link for debit card details. Underneath, dollars convert to USDC and out to pesos, quetzales or reais. The round is $87m of equity led by a16z (QED, Castle Island, Switch, Contour and Endeavor Catalyst joined) plus a $113m credit facility from General Catalyst's Customer Value Fund. Félix says it has moved $8bn for 6 million people across 11 markets, and will now add lending and savings through partners, an AI "financial companion," and expansion into Brazil and Venezuela.
🧠 Most Félix users could not tell you it runs on stablecoins. Godoy told TechCrunch in 2024 it could be a donkey crossing the border for all the customer cares. They want pesos, instantly, at the best price.
🧠 From $1bn to $8bn lifetime volume. They had moved $1bn for all of 2024 at the Series B in April 2025. It’s still a fraction of Wise at $243.5 billion, but Wise has more segments and more markets. Felix’s growth is staggering, and if it compounds, they’re going to be special.
🧠 The debt is the interesting half of this raise. General Catalyst's Customer Value Fund lends against customer acquisition with predictable payback. Félix must be able to demonstrate that.
🧠 The debt is also key to growth. Even with stablecoins, you still deal with pre-funding and bank cut off times for the fiat off-ramps. The more capital you have the more volume you can deliver.
🧠 A 1% federal excise tax on cash-funded remittances took effect in January. Card-funded digital transfers are exempt. Policy just handed WhatsApp-native senders a price edge over the cash-agent incumbents.
🧠 When Stripe bought Bridge two years ago, stablecoins were peak hype as an asset category. That has cooled, and real-world volume has done the opposite. Félix is what stablecoins-as-plumbing looks like: a company solving a customer problem that happens to settle on-chain.
TabaPay already works with 20+ partner banks and is on track for $100B in payment volume this year. Now it's buying Transact Bank, an OCC-chartered, FDIC-insured bank in Denver, renaming it TabaBank and putting both companies under a new bank holding company. FTV Capital led the $155M round (primary plus secondary), and the deal is slated to close in Q4 pending regulatory approval. TabaPay made its name in push-to-card payouts for lenders and fintechs who had never run payouts at scale, and grew from there into a money movement platform across card, ACH, RTP, FedNow and wire.
🧠 Twenty partner banks and they still bought one. Relying on multiple sponsor banks adds complexity and limits redundancy, and the hard use cases (digital banking, debt repayment) are the ones partners struggle to cover. Owning the bank lets TabaPay say yes where partners say no.
🧠 The $155M capitalizes the bank. TabaPay expects the raise to qualify TabaBank as a card acquirer across all industries and all major networks. That takes a payouts specialist and makes it a sponsor and acquirer in its own right.
🧠 Acquiring is big business. The payout specialist is getting into pay-ins. The monster growth stories of fintech Stripe and Adyen have mastered this. Could Tabbapay compete?
🧠 Buy vs apply. Mercury filed for a de novo charter in December and got conditional OCC approval in April; TabaPay chose to buy a bank that's already running. Same conclusion by two routes: the plumbing is too important to rent.
🧠 $100B of volume makes take rate the interesting number. At 10bps that's a $100M revenue business, at 25bps it's $250M (my arithmetic, TabaPay hasn't disclosed). Margin is the unknown, but owning the bank is how you move it.
4 Companies 💸
1. Lane - Payments for agents, MCP servers, and merchants
Lane lets an AI agent complete a purchase without leaving the chat. Users can ask their agents to buy a shirt under $100 with limits like which merchant. Approvals happen in the chat, and then the agent is granted a virtual card. Lane also gives MCP developers auth and usage billing, and gives merchants a no-code way to sell into ChatGPT, Claude, and Perplexity.
🧠 This is a fairly complete payments offering, but for whom? The future where agents can buy things for you feels inevitable, but less than 9% of US consumers are willing to let their agents buy for them. That’s not the point though. The people that are using AI to build apps to solve minor daily frictions probably are, and those people have their own Hermes or Grok bot. My bigger worry here is: can you build a two-sided market here, and where’s the traction? Natural kept it simple; it's a wallet for agents to make B2B style payments over non-card networks. Who’s lane for first? And how do they land traction with that?
2. River Markets - A prime broker for prediction markets
River gives hedge funds and trading desks one API to trade across Kalshi, Polymarket, and Polymarket US. It routes each order to the best price, and consolidates positions and P&L in a single view. They claim smart routing alone averages 2 cents of price improvement.
🧠 Prediction markets badly need a prime broker, because the same contract trades at different prices on different venues. Today, capturing that opportunity means parking cash at different exchanges. Cash that’s not being useful. A prime broker gives a trading firm one place to keep its money while it trades everywhere. It holds the cash, lends against positions, and nets exposure so a win on Kalshi can offset a loss on Polymarket.. Every new market builds new financial institutions. River (or something like it) could be one for prediction markets.
3. Rivo - AI that sweeps idle cash into T-bills for consumers.
Rivo connects to your existing checking account and sweeps idle cash into 4-week U.S. Treasury bills, pulling it back before your bills are due. In its simplest mode, it checks once a day and moves anything more than $100 above the minimum balance you set.
🧠 This is a smart, simple feature that doesn’t feel like PFM. Their site says $5.9 trillion is sitting in US checking accounts, earning close to nothing. The startup is charging 5 basis points to do it, but IMO that’s worth it. The tools exist here; open banking exists. But if this catches on, banks will have more deposit-flight worries. I sort of hope this becomes a feature every Neobank has in 5 years.
4. Attestable - Cryptographic proof of what an AI actually did
Attestable produces a mathematical proof that an AI output came from the approved model, weights, and input, and nothing else. Checking a proof takes under a second and doesn't require the model weights or a rerun. If anything was tampered with, the proof breaks. It's built for frontier AI labs, critical infrastructure, and governments that need to vouch for each other.
🧠 Today, "the model did X" means "trust our logs” and that depends on whoever runs the datacenter. If you’re worried about cyber risk, or you have a brand-new model being tested by a government and some corporates, this is a great way to know who did what with AI. The catch is that proof-of-work tops out around 53 tokens a second on a 1B-parameter model. That’s slowww when scaled up to anything useful. Still it's a start, and if new models are cyber risks, we need solutions like this.
Good Reads 📚
Humans come with spiky parts, soft parts, and squishy underbellies. We’re vulnerable, hard to deal with, and lovely. Because this is hard to interact with, we develop masks called “personhood,” an implicit social contract that’s designed to help us get along well in public. Things like being sorry for being late become little social norms that, if you display, mean you get accepted as a person. We don’t expect these behaviors of children, but society is full of these little contracts. Personhood breaks down for poor hygiene, drug addicts, and in fanatics or ideologues who get too preachy, in soldiers who are faceless vessels of a government.
There is some force drawing you to an invisible median of behaviors. But there are subcategories of personhood that remain equally consistent. Consider putting a group of people with Asperger's together. Their world model that values logic and reason over emotion leads them to enjoy each other's company far more than they might in the rest of society.
🧠 As someone who’s still trying to learn society's norms, I enjoyed this. Society never came with a manual, but when you deconstruct it and explain it rationally, it all makes a lot more sense. Maybe you will too if you’re a smidge neurodivergent like me :)
Tweets of the week 🕊
That's all, folks. 👋
Remember, if you're enjoying this content, please do tell all your fintech friends to check it out and hit the subscribe button :)
Want more? I also run the Tokenized podcast and newsletter.
(1) All content and views expressed here are the authors' personal opinions and do not reflect the views of any of their employers or employees.
(2) All companies or assets mentioned by the author in which the author has a personal and/or financial interest are denoted with a *. None of the above constitutes investment advice, and you should seek independent advice before making any investment decisions.
(3) Any companies mentioned are top of mind and used for illustrative purposes only.
(4) A team of researchers has not rigorously fact-checked this. Please don't take it as gospel.
(5) Citations may be missing, and I’ve done my best to cite, but I will always aim to update and correct the live version where possible. If I cited you and got the referencing wrong, please reach out
