In partnership with |  |
Happy Monday, Fintech Takers! I’m in Atlanta this week for Emerge, the Financial Health Network’s annual event. I am very excited, and will be heads down, taking notes all week. Expect a report back on what I learned in Friday’s newsletter. I hope you had a restorative weekend and are ready for a big week! Let’s get to it. - Alex PS. You built payments for growth. But is it built for scale? Join me and WEX on May 27th to talk about what separates the payments stacks that hold up under pressure from the ones quietly becoming a liability. Was this email forwarded to you? Subscribe Here
Sponsored by Rain Agents that execute purchases on behalf of users aren't theoretical anymore.
But the infrastructure to support it still has a lot of catching up to do. That's where Rain comes in.
Rain supports single use virtual card issuance for AI agents acting on behalf of consumers or businesses.
For any fintech company or platform exploring agentic commerce, that means built-in controls to prevent fraud and runaway usage, no new integrations required, and no behavior change on the consumer or merchant side.
It's all funded from existing collateral balances and settled on trusted, global card rails.
The agentic economy is arriving whether your infrastructure is ready or not. Rain makes sure it is.
El Greco's View of Toledoca. (1599–1600).
3 FINTECH NEWS STORIES#1: ChatPFMWhat happened?ChatGPT is rolling out personal financial management capabilities: Today we’re releasing a preview of a new personal finance experience in ChatGPT to Pro users in the U.S. Now you can securely connect your financial accounts, see a dashboard of where your money is going, and ask ChatGPT questions grounded in your financial context – all while staying in control of your data. We’re starting with a preview to a smaller group so we can learn from real-world use, improve the experience, and expand thoughtfully.
So what?Well, we knew this was coming. OpenAI acqui-hired Ethan Bloch when it bought up his company, Hiro Finance, last month. That clearly signaled that OpenAI sees the same opportunity that Perplexity, Cash App, Revolut, Public, and plenty of other companies see: LLMs and agentic AI have the potential to finally make personal financial management (PFM) work for the masses. While this news isn’t exactly shocking, there are a couple of things about this announcement that I found particularly interesting: This new PFM functionality will, initially, only be available to ChatGPT Pro users, which costs a rather steep $100 per month. Now, on the one hand, I’m pleasantly surprised by this. Even the best PFM product can fail if it’s handcuffed to the wrong business model. Alignment with the end user’s best interests is incredibly important and the only fool-proof way to ensure that is to charge an explicit fee for the product. For the moment, OpenAI is doing exactly that, which I like. However, there are a few problems. First, $100/month is expensive. I know you get a lot more for that $100 than just the PFM capabilities, but still. Many of the leading PFM products in the market today charge less than $100 annually. Second, while OpenAI has a staggering 900 million weekly active users, roughly 95% of them are free users. Most of the remainder pay for the less expensive ChatGPT Plus tier ($20/month). Only a small fraction (maybe 200,000 at most?) subscribe to ChatGPT Pro. I would imagine that OpenAI will eventually push a version of these capabilities down to the Plus and Free tiers (the company says that 200 million users already ask ChatGPT personal finance questions each month), but that’s a dangerous proposition, for a couple of reasons. Nothing is free, so if free ChatGPT users are getting PFM capabilities, it’s because they have become the product. Additionally, by starting with the 0.02% of users that subscribe to Pro, OpenAI risks designing a product that doesn’t work for the majority of the market. Speaking of which … It was amusing watching tech people on Twitter react to this news. Many of them expressed skepticism about these new capabilities, specifically with regard to the user interface. They seemed to feel like chat was not the ideal way to parse complex financial situations (Sheel’s tweet is a representative example). Two things about this. First, these new PFM capabilities do go beyond chat. They include charts, graphs, and lists of transactions, embedded directly into ChatGPT. Second, OpenAI absolutely should not be listening to these people’s product feedback! A large percentage of tech founders, operators, and investors live in a truly bizarre bubble that is, and I can’t stress this strongly enough, not representative of the average adult consumer. PFM products constantly fall into the trap of building for folks living in Silicon Valley and, as a result, failing to build something that works for people living everywhere else. PFM powered by LLMs has the potential to finally escape this trap, but to do so, PFM builders need to stay off Twitter. The real opportunity for PFM powered by LLMs is to go from read to write. Don’t just answer my questions about my finances. Suggest actions and once you get my authorization, go execute them on my behalf. OpenAI clearly sees this “agentic PFM” opportunity. It talked about agentic use cases in its blog post, and it highlighted Intuit as the partner that would help it get there, writing, “For example, a user could go from getting a credit card recommendation to understanding their approval odds and submitting an application, or from asking about tax implications of a stock sale to getting a trusted tax estimate and scheduling a session with a live, local tax expert, powered by Intuit and all inside ChatGPT.” Fascinating. If I’m interpreting that correctly, it sounds like Credit Karma and TurboTax are coming to ChatGPT soon. I’d imagine that banks aren’t thrilled about this development. Many of them have made half-hearted attempts to embed PFM within their apps, but, from what I can tell, these efforts haven’t yielded great results. Now OpenAI is signalling that it wants to be the home for consumers’ financial questions, which would put it at the very start of many consumer finance workflows (the Credit Karma use case is a good example … if OpenAI is going to be steering credit card applications in the near future, that disintermediates banks from their prospective customers even more than they already are). Plus, with Plaid acting as OpenAI’s data aggregation partner, banks will be even more likely to see open banking as a threat to be stopped at all costs (I wonder if JPMC, with the benefit of hindsight, regrets making pay-by-bank the use case it tried to kill with its pricing, rather than PFM). And speaking of Plaid, I know the company is excited about this partnership (and its partnership with Perplexity), and I get why. It is gearing up for an IPO at some point, and partnerships with some of the most influential companies in the world are very valuable signals. However, if I could offer Plaid some advice, I would suggest toning it down a bit. A large part of Plaid’s success, up to this point, has come from the much-less-heralded work of traditional PFM providers (Monarch, Copilot, YNAB, etc.) that have built small, but highly loyal and engaged customer bases, all of which heavily contribute to Plaid’s overall transaction volume. These companies are scrambling to figure out how they can stay relevant in the age of LLMs and agentic AI, and I’m guessing they don’t appreciate Plaid’s overt public enthusiasm about Perplexity and OpenAI. OpenAI spends quite a bit of time in its announcement assuring us that these new PFM capabilities will keep users’ financial data secure, and that the company will fully respect users’ privacy. Personally, I don’t find these reassurances all that reassuring, given how data-hungry the AI labs all are and how little regulatory supervision they are currently receiving, but that’s just me. The question is how ChatGPT users, more broadly, think about these security and privacy risks, especially as OpenAI moves its PFM capabilities down to its Plus and Free tiers. My guess is that it will depend on how well the capabilities work. Consumers tend to be very open to sharing their data, in exchange for valuable products and services. However, if ChatPFM makes a bunch of dumb and frustrating mistakes (which seems very possible!) that trust will evaporate quickly and you will have a bunch of consumers suddenly wondering if OpenAI is liable for the damage caused by those mistakes (spoiler alert: it won’t be!)
#2: Selling Stablecoin-shaped Picks and ShovelsWhat happened?A bunch of stablecoin-related news! The Clarity Act finally cleared the Senate Banking Committee and will proceed to the floor for a full vote by the Senate: The Republican-led Senate Banking Committee on Thursday advanced long-awaited legislation that would create regulations for cryptocurrencies - a landmark step for the bill which has been bogged down by a dispute between crypto companies and banks. The bill, which was advanced with support from two Democrats, will now proceed to the full Senate, setting the stage for a fierce lobbying fight.
Anchorage Digital is pulling back on its commitment to support the growth of a specific stablecoin that it has economic interests in, and will instead focus on enabling the issuance of many different stablecoins: Anchorage Digital … says it will take a back seat to the Global Dollar stablecoin (USDG) consortium, which includes Robinhood and Kraken. USDG, which has a circulating supply of around $3 billion, is issued by Paxos Digital Singapore and supervised by the Monetary Authority of Singapore. [Anchorage co-founder and CEO Nathan] McCauley said that previously, Anchorage might have been boosting USDG specifically, but now the firm will take a more neutral approach. “I think one of the things you're gonna see from us is increased neutrality on the stablecoins. It just makes sense to be neutral and not specifically be pushing any one stablecoin.” Anchorage recently mentioned as many as 20 banks and tech giants are currently looking to issue stablecoins with the San Francisco-based custody firm.
And a stablecoin infrastructure company that helps fintech and crypto companies offer yield raised some money: Stablecoin infrastructure platform OpenTrade has raised $17 million in new funding. The strategic round brings the U.K.-based startup’s total funding to more than $30 million and gives OpenTrade “the runway to scale at the pace the market is demanding,” as the company said in a post on LinkedIn Wednesday. OpenTrade says it is working to address demand in three areas, including permissioned yield infrastructure for FinTechs and neobanks developing custodial stablecoin products, as well as “permissionless infrastructure for non-custodial platforms and asset issuers” in the decentralized finance (DeFi) space.
So what?OK, let’s see if we can connect all the dots here, starting with Anchorage. As you may know, Anchorage is the first federally chartered crypto-native bank, having received a national trust bank charter in 2021 (well before it was cool). In 2024, Anchorage became a founding member of the Global Dollar Network, a consortium of crypto and fintech companies — including Robinhood, Kraken, Galaxy, Visa, and Worldpay — that issued a stablecoin (USDG) using a shared-economics model that distributes float earnings back to network participants based on how much of the token they integrate and circulate. The stablecoin itself is issued by Paxos, but Anchorage has served as the consortium's US regulatory anchor. In 2025, Anchorage launched its own stablecoin issuance platform, allowing partner brands — fintech companies, banks, payment platforms, and crypto-native firms — to issue their own white-labeled, federally regulated stablecoins on top of Anchorage's OCC-chartered infrastructure, with Anchorage handling the regulated issuance, custody, and reserve management while the brands own the tokens, the distribution, and the customer relationships. This issuer-first strategy seems to have won out, internally, given Anchorage’s decision to step back from USDG in order to bolster its own neutrality. This decision mirrors a similar decision that PayPal made recently, with the launch of PYUSDx, a tokenization and issuance framework that enables third parties to create application-specific stablecoins backed by PayPal’s stablecoin (PYUSD). The question is: Why is this happening? Why are Anchorage and PayPal prioritizing stablecoin enablement for other brands, rather than continuing to drive the adoption of their own, proprietary branded stablecoins? I think the answer is simple: offering your own stablecoin is becoming a bad business, and enabling other companies’ stablecoins is becoming a good one. Consider what's about to happen. The Clarity Act, as currently written, bans yield that's "economically or functionally equivalent" to deposit interest paid solely for holding a stablecoin. But it preserves the right to pay activity-linked rewards, for trading, staking, transacting. Coinbase, which spent months lobbying for this compromise, has said publicly it got the "must-haves" it needed, which suggests to me that, assuming regulators stay friendly towards the crypto industry, this compromise would functionally allow Coinbase to continue doing what it’s doing. Yield, in some form, will be coming to virtually every branded stablecoin. This will be good for users and terrible for issuers. The historic business model of stablecoin issuance — take in dollars, hold them in Treasuries, pocket the float — only works when you don't pass the yield through. In a world where every branded stablecoin has to pay competitive rewards to win and keep customers, that margin gets arbed away fast. In this hypothetical (but very realistic) future, you’ll either need to get out of the stablecoin adoption game altogether (and start selling picks and shovels … like Anchorage and PayPal are doing), or you’ll need to find a better source for generating yield than simply buying Treasuries (or integrating with tokenized treasury funds like BUIDL to get those same treasury yields). That second option is where OpenTrade, which sources yield from a broader basket of tokenized real-world assets — short-duration credit and trade finance, alongside repo and Treasuries — comes in. I expect demand for its services (especially if they can be easily packaged in a way that dodges whatever yield restrictions end up coming out of Clarity) to grow significantly, for the same reasons that PayPal and Anchorage are moving into the stablecoin issuance business: Unless you’re Tether (and none of the normal rules apply to you), it’s about to become extremely hard to make money solely by offering your own stablecoin. #3:Prediction Market NonsenseWhat happened?[Deep sigh] Kalshi: Prediction market Kalshi today announced a $1 billion Series F round at a $22 billion valuation, led by Coatue, with participation from Sequoia Capital, Andreessen Horowitz, IVP, Paradigm, Morgan Stanley, and ARK Invest. This raise comes amid the start of institutional demand. Over the past six months, institutional trading volume has increased 800%. Kalshi will use the new capital to scale adoption across hedge funds, asset managers, proprietary trading firms, and insurance companies - unlocking access to trillions of dollars in capital. The company will continue expanding its product suite, including recently-launched block trading capabilities, upcoming risk products, and deeper broker integrations tailored to institutional demand.
The CFTC: The chief U.S. regulator of increasingly popular prediction markets says he'll treat them as financial products, arguing that most people aren't using prediction markets for "entertainment." Michael Selig, chair of the Commodity Futures Trading Commission, tells Axios in a rare interview that prediction markets and sportsbooks are "two separate things." "They're different models. The conventional sportsbooks and casinos are entertainment and they have a lot of authority to be able to kick people out when they keep winning," Selig said in an interview. "What you're seeing is markets versus entertainment," Selig said. "For those that want the discipline and integrity of a market, it's a better model. For those that want entertainment, the casinos might be the model for them."
Problem gambling: The National Council on Problem Gambling (NCPG) today announced a $2 million, two-year investment from prediction market platform Kalshi to support a strategic initiative focused on trader health and safety. NCPG has established a new membership subcategory: the Financial Services & Trading Subcategory. Kalshi will join NCPG as a Platinum-level member, becoming the first organization in the subcategory, a defining commitment to long-term customer safety from the financial sector.
Insider trading: For this campaign staffer, the method was simple. First, they'd receive a tip on an unreleased poll and compare it with the odds on a prediction market, like PredictIt or Polymarket. If the poll reported their candidate had a better chance of winning than the prediction markets, they'd use this edge to buy low-cost odds on their candidate — known as event contracts — before the poll was released. On prediction markets, the price of an event contract often mirrors the market's estimation of the probability of a given outcome — in this case the chance a candidate will win. So a contract selling for 20 cents means the market is pricing a 20% chance of success. Once the poll went public, the prediction market contracts shot up in value. The staffer would then sell their contracts at a higher price and make money.
So what?At some point in the future, a regulatory agency (The CFTC? The CFPB? Bueller? Bueller?) will take action to kick prediction markets out of retail investing and banking platforms like Robinhood and Coinbase, and we will all be like, “yeah, that makes sense. Surprised it took this long TBH.” And then my nightmare can be over and I can stop writing about prediction markets. But until then, yay! More prediction market news! All of these stories are interesting and dumb and infuriating for different reasons. Let’s quickly walk through them. First, we have Kalshi raising $1 billion at an $22 billion valuation and bragging about the surging interest that institutional traders are showing in the platform. This isn’t, at all, surprising. However, we should put that framing into the proper context. Institutional traders like prediction markets because those markets feed them a steady supply of unsophisticated retail traders who lose constantly. Why do retail bettors keep lining up to lose money on prediction markets? Well, according to the person in charge of regulating prediction markets, it’s not because they’re looking for “entertainment.” This assertion by Mike Selig is, obviously, absurd given the dynamics at play (sports contracts make up a majority of Kalshi’s volume!), and it is further undercut by our third story. Kalshi has invested $2 million in the National Council on Problem Gambling, becoming a Platinum-level member and establishing a new membership subcategory for financial services & trading (urrgghh) because … prediction markets are mostly used by disciplined institutional investors to hedge significant economic risks? Give me a fucking break. Kalshi bristles whenever anyone refers to the activity happening on its platform as “gambling,” and yet it just invested $2 million in the National Council on Problem Gambling. And Kalshi (and Mike Selig) love to point out how much better they are than traditional casinos, specifically because they don’t kick off bettors who are winning (this is bullshit, by the way … Kalshi empowers institutional traders to identify and, in theory, limit sharps on sports parlays), and yet both Kalshi and Selig have made it abundantly clear that they are going to crack down on what they refer to as “insider trading,” even though much of the activity that they designate as insider trading (especially in the area of political campaigns) doesn’t really fit the classic definition. Is it insider trading to trade on your decision, as a candidate, to drop out of a race? Is it insider trading to trade on the results of an unreleased poll that you’ve gotten early access to? Legally, it’s unclear. What is clear is that sharp retail bettors with an edge are bad for the institutional investors that Kalshi (and Selig) actually care about.
2 READING RECOMMENDATIONSI’m not sure I buy the premise of this piece, but it made me think. Slightly different take on the OpenAI news from Ron. BTW, Ron’s writing has migrated over from Forbes to Substack, which I am very happy about. Make sure you subscribe!
There are a TON of interesting questions being asked in the Fintech Takes Network. I’ll share one question, sourced from the Network, each week. However, if you’d like to join the conversation, please apply to join the Fintech Takes Network. How should on-chain prime money market funds (MMFs) like OpenTrade be regulated? Given the problems that TradFi prime MMFs caused during the great recession, it feels like this is an area that deserves some attention. If you have any thoughts on this question, reply to this email or DM me in the Fintech Takes Network!
Thanks for the read! Let me know what you thought by replying back to this email. — Alex |
