š Welcome back to another edition of Fintech Radar!
If youāre new, here is a breakdown of what you can expect from each issue.
If you missed our recent editions, you can catch up here. Some previous issues you might want to check out if youāre new include āA Deep Dive Into The Cash Appās Growth Machineā, āThe Future Of Payment Initiationā, and āCurrent: Doing It Differentlyā.
Your ad could be HERE!
Fintech Radar is a must-read for fintech founders, operators, and investors. If thatās your target audience, placing an ad right HERE is a cost-effective way to reach them.
If this sounds like a good fit for your brand, head to our sponsorship page for more details and to secure your advertising slot. Prices start at $100 per issue!
If you have any questions, reply to this email and ask away!
š Airtel Money files for a London IPO, The Guardian
š The Rundown: Airtel Money, the mobile money arm of Airtel Africa, announced last week that it intends to list on the London Stock Exchange. The offer is made up entirely of shares sold by existing shareholders, with a free float of at least 10%, and the FT reported the company is targeting a valuation of $8-9 billion. The IFC has signed on as a cornerstone investor for up to Ā£67.2 million, with pricing expected in mid-October.
š„” Takeaway: Airtel Money is such a great story. Airtel Africa exists because Indiaās Bharti Airtel, Sunil Mittalās telco, bought Zainās African operations in 2010 for around $10.7 billion. Airtel Money grew up inside that business as a way to send and store money over basic mobile phones, much like the well-known M-Pesa did in Kenya. Airtel Africa listed in London in 2019, and in 2021 Airtel Money raised about $500 million from TPGās Rise Fund, Mastercard and the Qatar Investment Authority at a valuation of roughly $2.65 billion.
It now has 53 million monthly active users across 13 African markets, processed $213 billion in payments over the last twelve months, and made $1.35 billion in revenue last financial year at an EBITDA margin of around 50%. It carries no debt and spends about 3% of revenue on capex. CEO Ian Ferrao said that is exactly why nobody is raising fresh money in the float: āThe business is debt-free, capital-light and highly cash generative.ā Wow.
Interestingly, Airtel Money has converted 41% of Airtel Africaās 128.9 million telco subscribers, which still leaves more than 75 million people who already have an Airtel SIM and donāt yet use the wallet. App users also bring in $9.50 a year against $1.80 for someone on a feature phone dialling USSD codes, so every customer who moves onto a smartphone is worth roughly five times as much. Iād expect a fair chunk of that shift to happen with or without a push from Airtel.
Pair that with a dividend policy of at least 80% of net profit and it is being pitched a lot like a toll road. Thatās probably the right pitch for London, where investors have spent the last few years watching the more exciting tech names choose the NYSE. The obvious comparison, M-Pesa, still sits inside Safaricom rather than trading on its own, so this would give global investors a rare pure-play way into African mobile money.
The listing was originally pencilled in for the first half of the year and got pushed back, and Airtel looked at a Middle East listing before settling on London. Nils Pratley at the Guardian reckons itās ānot yetā the end of the Cityās listing drought, and Iām inclined to agree that one deal doth not a trend make. It would be a very good one to have on the board, though.
š Amazon blocks Metaās Muse agent from shopping its site, Payments Dive
š The Rundown: Amazon started blocking Metaās Muse AI agent from its retail site last week, after Meta declined a request to keep Muse off Amazon. Shoppers who send Muse to Amazon now get a pop-up saying ācontinued access by an unauthorized AI agent violates Amazonās Conditions of Useā. In the same week, Shopify and PayPal both announced deals letting Muse check out across their merchants.
š„” Takeaway: Surprise, surprise, Amazon doesnāt want bots shopping on its site. It makes the case that agents shopping for customers should operate openly and let the other side decide whether to take part, the way delivery apps work with restaurants and travel sites work with airlines. Its more specific complaint is that Muse doesnāt identify itself and appears to capture and store customer credentials, which Meta disputes, saying credentials sit in secure storage that the model canāt see. The identity point is a fair one, if you squint. To be fair, an agent that logs in as you without saying itās an agent looks a lot like the account takeover activity that fraud teams spend all day trying to stop.
But letās be serious, this is also about who owns the customer. Amazonās advertising business depends on people browsing its pages and scrolling past sponsored listings, and an agent that goes straight to the item it was asked to buy skips all of that. Not to mention all the upsell techniques/tricks Amazon has been honing for years. On the more direct competition front, Amazon also runs its own agent, Buy for Me, which purchases from other retailersā sites, and sellers whoād rather not be included have to email Amazon to opt out (lol, hypocritical much?).
Itās interesting to note that the payments companies have decided to go in the opposite direction. Muse launched three weeks ago with Stripeās Link built in, Shopify brought Shop Pay checkout to Muse across all Shopify stores the day after Amazonās block, and PayPal followed the day after that. Those are rails the merchant has already opted into, where the agent pays with a credential issued for it rather than logging in as the customer. Amazon is big enough to keep agents out for now (and is still fighting Perplexity in court over the same issue), but most merchants arenāt. I suspect the agents that end up doing the most shopping will be the ones that come in through a checkout built for them.
š¤ Six global banks publish principles for agentic commerce, Bank of America
š The Rundown: Bank of America, Capital One, Commonwealth Bank, ING, NatWest and New Zealandās ASB published a joint paper last week called āBuilding Trust in Agentic Commerceā. It sets out five voluntary principles for AI agents that buy things on a customerās behalf: transparency, safety, privacy and data, choice, and interoperability. A second paper on implementation is to follow, and other banks have been invited to sign on.
š„” Takeaway: At first blush this is only a paper, itās non-binding, so nobody has to change anything tomorrow. Some of the details are still interesting though. More specifcally, every party in a transaction should know when an AI agent is involved and who it is acting for. Customers should be able to see and revoke whatever theyāve delegated to an agent. And liability should sit wherever the risk or the error was introduced.
Liability is going to be the hardest one to settle for agentic payments. Card disputes and chargebacks were designed around a human cardholder authorising a purchase. When an agent buys the wrong flight or the wrong size, the question of who eats the loss (the bank, the merchant, the agent provider or the customer) is still largely unanswered. Visa and Mastercard have both spent the year publishing their own agent frameworks, and this reads to me like the issuers making sure they get a say before the networks and the model providers write the rules for them.
The banks own the customer relationship, the dispute process and the fraud losses. That gives them more leverage here than their slow start on agentic products might suggest. It is also good to see CBA in the group, which should mean Australian customers arenāt an afterthought when this gets built out.
š· GoCardless processes an agentic bank payment in the UK, GoCardless
š The Rundown: GoCardless announced last week that it has processed what it calls the UKās first agentic account-to-account payment, a recurring donation to the anti-poverty charity Trussell set up entirely inside a conversation with an AI agent. Donors choose Ā£5, Ā£10 or Ā£15 a month, then the agent asks for their bank details and sets up a Direct Debit mandate on their behalf. The checkout was built inside the FCAās AI Live Testing programme, which GoCardless joined earlier this year.
š„” Takeaway: This dovetails in neatly with the previous story. Most of the agentic commerce work so far has been built on cards, with Visa and Mastercard handing agents tokenised card credentials to spend with. GoCardless is making the case for bank rails instead, and Hiroki Takeuchi put it plainly: āWeāve long argued that account-to-account payments are the natural foundation for agentic commerce.ā For recurring payments that argument holds up well. A Direct Debit mandate is already a standing permission that the payer can see and cancel whenever they like, which is close to the scoped, revocable permission everyone says agents need, and it comes with the Direct Debit Guaranteeās refund protections already attached.
The pilot itself is small. This is a charity donation of Ā£5 to Ā£15 a month, and the agent is collecting bank details from a person sitting in the chat rather than spending money on its own. That is deliberate. GoCardlessās own research says 64% of UK consumers are open to AI managing their recurring payments, as long as they keep control over limits or final approval. GoCardless is in the FCAās second AI Live Testing cohort alongside Barclays, Scottish Widows, UBS and Experian, and the FCA expects to publish its evaluation by early 2027, which is likely where the first UK ground rules for agent payments will come from. That will probably matter more than the transaction itself.
Pay by bank has struggled to win at the ecommerce checkout in the UK (and in many markets to be honest), mostly because the bank redirect is a confronting UX experience. An agent doesnāt mind a redirect, so I could see agentic payments ending up as the way account-to-account finally gets a foothold against cards, which is exactly what the UKās National Payments Vision has been asking for.
š¦ Nubank is in talks to buy Monzo, Sky News
š The Rundown: Sky News reported last week that Nu Holdings is in early-stage talks to acquire Monzo, in a deal that could value the UK digital bank at between Ā£8 billion and Ā£10 billion. Monzo has hired Morgan Stanley and Qatalyst to advise, and any deal would likely be a mix of cash and stock. The other option on the table is a new funding round to pay for expansion into mainland Europe, and Sky has since reported that Advent is among the buyout firms that have held early talks about a minority stake. Neither company would comment.
š„” Takeaway: Monzo was last valued at Ā£4.5 billion in a secondary share sale in October 2024. Since then it has grown to 16 million customers, and its latest results showed revenue up 39% to Ā£1.7 billion with adjusted pre-tax profit of Ā£172.6 million. For a profitable bank that is still growing that quickly, the Ā£8-10 billion range doesnāt look like a stretch. On Nuās side it works out to somewhere between a sixth and a fifth of its own $65.5 billion market cap, which is a big bite, but a digestible one if a chunk of it is paid in stock.
Nubank has built its way into every market so far. Mexico and Colombia were greenfield, and the US launch we covered a few weeks ago came alongside the Nu Global push. Buying Monzo would be a different gear: a UK banking licence, 16 million customers and a team that has already run a full-stack licensed neobank in a market where Nubank has no presence at all. Revolut took around four years to go from application to a full UK banking licence, so buying direct access rather than waiting for it is a much quicker route if Nubank is serious about going global.
Thereās a nice bit of symmetry (and irony) if this deal were to come together. Monzo spent years trying to crack the US and walked away earlier this year, while Nubank went the other way and is now in the US on its own terms. A deal would put both of those chapters in one proverbial book, and it would also take one of the more obvious London listing candidates off the table for the foreseeable future. In some ways, this would be a sad end to a golden era for UK fintech that has yet to see an IPO.
I genuinely donāt know if this gets over the line. Talks at this stage often fall over, and Monzo has a perfectly good fallback in a funding round at a similar valuation. If it does, Iāll be very interested to see what Nubank does with a UK bank.
š° Paymob raises $35M in a pre-Series C round, Wamda
š The Rundown: Cairo-founded Paymob announced a $35 million pre-Series C round last week, co-led by Mubadala and the EBRD, with British International Investment, Global Ventures and DPI Ventures participating. No valuation was disclosed.
š„” Takeaway: Paymob serves more than 390,000 merchants across Egypt, the UAE, Saudi Arabia and Oman, and its revenue has tripled in 18 months. The Gulf is doing most of that work. GCC revenue is up seven times and is now close to half the total, helped by the UAE retail payments licence it picked up in January 2025 and the 20,000 UAE merchants it has signed since.
An Egyptian payments company moving its centre of gravity to the Gulf makes a lot of sense. Egypt gives you scale and a big SME base, but also a currency that has been sharply devalued, while the Gulf gives you dollar-pegged revenue and merchants who will pay more for acceptance. Some of the money is going into agentic commerce products, which seems to be a line item in every payments round at the moment.
š° Benford raises ā¬5M to rebuild the financial audit, Tech.eu
š The Rundown: Oslo and London-based Benford announced a ā¬5 million pre-seed round last week, led by firstminute capital with Global Founders Capital and Sondo participating.
š„” Takeaway: Benford is super interesting. It plays in the full-stack space that I think is only going to get hotter as AI continues to eat up more of the workload. More specifically, Benford is a licensed Norwegian audit firm that carries out statutory audits itself, rather than selling software to auditors. Its platform connects to a clientās ERP, subledgers and bank accounts and runs the audit all the way through to the signed opinion. It already has live clients, and it plans to expand into Sweden and then the European mid-market, which it says the Big Four are stepping back from.
Owning the licence and the service means Benford keeps the whole fee rather than just a software subscription. It also means carrying the liability and the staffing of an audit firm, and Benchās collapse at the end of 2024 is a reminder that running a service business on a software companyās cost base is harder than it looks. But maybe this time is different. If the AI-led automation holds up, this is a much bigger prize than another tool for the incumbents.
š§ Taktileās CEO Raised $190M Betting Most Bank Jobs Wonāt Exist in 10 Years, Fintech Leaders (September 22)
Miguel Armaza sits down with Maik Taro Wehmeyer of Taktile, whose platform automates credit, fraud and compliance decisions for financial institutions, to talk through the failed first product and the 99 bank rejections it took to get there.
š§ Under the Hood of Fintech Products, Fintech Takes (September 23)
Alex Johnson and Jas Shah dig into why bank digital onboarding still falls short after a decade of work, and why personal finance management remains one of the hardest categories in consumer fintech. Add this one to your playlist.

ā¤ļø Show Some Love For FR
š§ Feel free to reach out if you want to connect. Iām @alantsen on Twitter, or you can D.M. me directly by clicking the button below ā“
Ps. If you like what Iām doing with FR, please share it on your social disinformation network of choice. Iād also appreciate it if you forwarded this newsletter to a friend who might enjoy it.
š What did you think of this weekās issue of FR?
I Love It! ā I Like It ā Not Bad ā I Donāt Like It ā Itās Awful





Airtel Africa seems to be the better investment option because they will be improving their finances by selling this stake in Airtel Money.