1. Podcasts
  2. Fintech Stacks
  3. 🚦Citi Beats and Still Bleeds. Visa Gets a Green Light in China. Stripe’s Privy Hands AI Agents a Wallet on XRP.
Fintech Stacks

14 Oct 2026 · 22m · 11 chapters

Episode artwork

🚦Citi Beats and Still Bleeds. Visa Gets a Green Light in China. Stripe’s Privy Hands AI Agents a Wallet on XRP.

In short

How financial “headline wins” hide costs, and how payments are shifting from regulated card networks to faster blockchain rails for AI agents.

Guest backgrounds

No guests are named in the transcript; it’s a two-host discussion.

Key claims

Citi’s Q3 results look great but the 40% profit jump is flattered by a prior-year Banamex impairment; sequential profit fell 10% (especially U.S. credit cards) due to a rewards/partner-payment arms race despite healthier consumer behavior. Visa’s China progress is only a “prepare” approval after nine years, likely tied to geopolitics, and it still can’t clear domestic transactions. Stripe’s acquisition of Privy enables embedded wallets for AI agents on XRP, but Ripple’s “13M agent payments” lacks independent verification and dollar-value context, so “25% of activity” may be vanity.

Notable examples

Citi net income $5.2B vs EPS $2.85; Banamex $726M impairment; card spending +11% while profits -8%; PBOC approval timing after Trump–Xi summit; Privy powering ~75M accounts; RLUSD stablecoin option; X402 Foundation consortium.

Written by AI. May contain mistakes.

Chapters

0:45

The Illusion of Financial Clarity

Exploring how financial results can be misleading despite positive headlines.

“But the actual game, the real mechanics of it, that's being played entirely in the shadows of the fine print.”
3:07

Citigroup's Q3 Earnings Report Analysis

In-depth breakdown of Citigroup's earnings and what the real numbers indicate.

“So they took this massive one-time accounting write-down last year.”
4:30

Market Reactions and Underlying Metrics

Discussion on why Citigroup's stock dropped despite strong reporting numbers.

“So people are spending more and paying it back more reliably.”
6:34

Visa's Long-Awaited Entry into China

Detailing Visa's attempts to break into the Chinese market and the challenges faced.

“It's a total margin pinch, and every single card issuer is feeling it right now.”
7:39

The Competitive Landscape in Chinese Payments

Examining the competitive environment Visa faces with local giants Alipay and WeChat Pay.

“Let's talk about the headline crossing the wire first.”
10:16

Geopolitical Factors Impacting Financial Regulations

Exploring how geopolitical dynamics affect Visa's regulatory challenges in China.

“Yeah, cash and plastic are just not part of the daily habit anymore.”
11:58

AI and Blockchain: The Future of Transactions

Introduction to Privy and how AI agents are changing the landscape of digital transactions.

“We are moving from the legacy plumbing of traditional banking straight to the blockchain, specifically looking at a phenomenon where AI robots are getting their own crypto wallets.”
14:00

The Limitations of Traditional Banking APIs

Explore the challenges faced by traditional banking systems for autonomous AI agents.

“And why wouldn't it just use a traditional bank API?”
18:17

Analyzing Market Metrics

Discuss the implications of transaction counts without dollar values in the financial landscape.

“And so they are aggressively hedging their bets by building the rails for this parallel borderless system.”
19:16

Understanding Financial Plumbing

Gain insight into the complexities of financial systems and their impact on transactions.

“Which leads me to a final thought I want to leave you with today.”
Show all 11 chapters
19:55

The Future of Autonomous Payments

Contemplate the potential impact of AI on traditional financial loyalty programs.

“Something for you to just mull over that wasn't explicitly spelled out in our sources, but it feels like the inevitable collision of everything we just discussed.”
A man walking through an airport, asking his phone a question

Ask about this episode

Get the gist, explore an idea or ask about something said.

Claude

On Fintech Stacks, what are the key points of this episode?

Searching your listening

Transcript

Automatic transcript. May contain errors.

0:00I mean, you know how it is. Usually when you look at a scoreboard at the end of a game, there is this expectation of absolute clarity, right? Oh, totally. The buzzer sounds and that's it. Yeah, exactly. You look up and the numbers tell you exactly what happened. 99 to 98. One team won, one team lost. It's just completely clean. Right. There's no ambiguity. But then you step into the world of global finance. You look at corporate earnings and suddenly that scoreboard is just entirely broken. Oh, it's shattered. It's the absolute definition of an optical illusion. Right. Because we are looking at a landscape today where a company can, you know, seemingly win the game, break records, and still get totally booed off the field by the crowd.

0:44Yeah, because you were looking at the bright lights of the headline. But the actual game, the real mechanics of it, that's being played entirely in the shadows of the fine print. Okay, let's unpack this. Because today's deep dive gives us three really fascinating looks at what happens when you peel back that bold font. We've got a lot of ground to cover. We really do. And if we look at the baseline here, whether we are talking about traditional banking or, you know, blockchain plumbing, every single win in finance right now comes with a hidden cost. Right. And I want to start right in the thick of traditional banking.

1:23Let's look at Citigroup's latest Q3 earnings report. Because if you just read the headline facts, you would think champagne corks were popping in the boardroom. Oh, on paper, it looked like a massive victory lap. It really did. I mean, Citi posted$5.2 billion in net income, and that was on$23.8 billion of revenue, which made it their best third quarter in a decade. A full 10 years. That's not a small thing. Not at all. And their earnings per share, their EPS, hit$2.85, which totally blew past the$2.65 estimate. Right. They beat the streak. Yeah. And their return on tangible common equity hit 11.5%.

2:02That's up 180 basis points from a year ago. So, I mean, here is my fundamental question. Let's hear it. If I am an investor, right, and I am looking at an 11.5 % return on equity and an EPS that just crushed estimates, I am expecting a massive stock rally. You'd think everyone would be buying. Exactly. But instead, Citi's stock dropped 4%. Like, did the market just decide the underlying quality of those earnings was terrible? Well, yeah. The market basically looked right past that top line beat and focused entirely on the baseline effect. Okay, break that down for me. Sure. So let's examine that 40 % year-over-year profit jump that drove all those initial headlines.

2:44The rocket ship number. Right, the rocket ship. That number is completely flattered by an accounting anomaly. It's a mirage. Really? How so? It looks so extraordinarily high only because we are comparing it to the year-ago quarter. And in that specific quarter last year, Citi took a massive$726 million impairment charge. Oh, wow. Okay. Yeah, it was related to Bonamex, which is their Mexican consumer banking business. Ah, okay. So they took this massive one-time accounting write-down last year. Which artificially lowers the floor. Exactly. So when you have a giant crater in last year's numbers, simply stepping onto flat ground this year looks like you climbed a mountain.

3:25That is actually the perfect way to visualize it. That's exactly how an impairment charge impacts year-over-year growth. Because the starting point was in a ditch. Right. So if you strip that one-time Banamex event out of the equation entirely, their net income only rose 17%. Which is, I mean, it's still a solid quarter. It is. But it's absolutely not the 40 % headline they were touting. And then you have to look at the metric that really had institutional investors heading for the exits. Which was what? This sequential growth. Net income actually slid 10 % from the second quarter of this very same year.

3:59Wait, they went backwards sequentially. They went backward. Okay, that explains the 4 % drop in the stock. And looking at the breakdown in our source material, that sequential bleed is heavily concentrated in the U.S. credit card division. Yep. The profit there fell 8%. But I've got to push back on the logic here because the underlying consumer behavior in this report looks incredibly healthy. It does look healthy on the surface. Right, because card spending actually rose 11 percent and charge-offs, people defaulting on their debt, dropped to 3.6 percent from 4.0 percent. So people are spending more and paying it back more reliably.

4:36Exactly. They are swiping more frequently and defaulting less often. That sounds like a golden goose for a bank. I mean, the only way profits fall 8 % in a high-spend, low-default environment like that is if the cost of acquiring and maintaining those cardholders is skyrocketing. And boom, that brings us to the actual mechanics of modern credit card economics. The real game being played. Right. The profit margin is just being devoured by higher rewards costs, partner payments, and credit provisions. Even though consumers are paying their bills, it is simply getting vastly pricier to defend those credit card lines from competitors.

5:14I think this is the exact moment where we need to point the finger directly at you, our listener. Oh, absolutely. Yeah, all those credit card rewards points that you meticulously optimize, you know, the rotating cashback categories, the airline miles, the airport lounge access. Everyone loves those. Of course we do. But you, the learner listening right now, you are actively squeezing the profit margins of some of the biggest financial institutions in the world. It's true. But to be fair, it is a completely self-inflicted arms race by the banks. How so? Well, the competition for the American consumer's wallet is just so fierce right now.

5:49Issuers are essentially trapped. Trapped by their own perks. Exactly. They have to offer these increasingly lavish perks just to prevent you from, you know, churning your account and jumping to a competitor's huge sign-up bonus. Right, because I'll just open a new card for the 80 ,000 airline miles and close the old one. Precisely. And when we talk about partner payments, paying the airlines or the hotel chains for those co-branded cards, the leverage has completely shifted. The airlines know they have the power. They do. They know exactly how valuable their loyalty programs are, so they demand a much larger cut of the swipe fees from the banks.

6:24So, yes, Citi won the top-line revenue game, but the market fixated on the mathematical reality that maintaining that revenue is becoming structurally more expensive every single day. It's a total margin pinch, and every single card issuer is feeling it right now. I mean, it completely explains the split decisions we saw across the broader financial sector this quarter. Oh, the Sunday NFL slate of banking earnings. Right. Because you look at J.P. Morgan, they completely missed their expectations. They posted 557 in EPS against a 590 estimate. A pretty big miss. Yeah. But meanwhile, Wells Fargo beat their expectations and also saw their card charge-offs drop.

7:03It paints a really vivid picture. You have a consumer base that is surprisingly resilient, but a financial industry that is paying a massive, massive premium just to keep that consumer's attention. So if you're a bank and you are trapped in this really expensive cage match for domestic consumer loyalty and your margins are getting compressed by airline partner payouts and all these cashback optimizers, what do you do? You inevitably have to look for a larger, untapped arena to grow. You have to go somewhere else. Which takes us straight to China. And Visa's attempt to finally enter that market.

7:37A very long attempt. Let's talk about the headline crossing the wire first. Visa just got a green light from Beijing. The People's Bank of China, the PBOC, approved their application to prepare. Keyword being prepare. Right. Now, Visa originally filed to become a bank card clearing institution in China back in August 2017. Nine years ago. A nine-year wait? That sounds less like a strategic expansion and more like waiting a decade at the DMV just to get a learner's permit. It really does. I mean, does an application to prepare actually mean they can start rolling out point-of-sale terminals in Shanghai next month?

8:13Like, can they finally swipe? If we connect this to the bigger picture, you need to put the confetti away. No swiping yet. No swiping. That permit does not even come with a car, so to speak. It's just the learner's manual. The PBOC operates on a highly deliberate, multi-tiered regulatory staircase. This approval is strictly a preparation stage. Oh. Yeah, there is no launch date attached to this, and Visa still cannot clear a single domestic transaction in China. Nine years, and they still can't clear a transaction? Why does it take so long? Because the regulatory bodies in China want to ensure that any foreign entity integrating into their financial plumbing cannot disrupt their domestic capital flows.

8:54Hence this agonizingly slow, step-by-step bureaucratic process. But looking at the competitive landscape from our sources, Visa is actually the last of the big three Western networks to make it to this specific step. Yeah, they are trailing behind. American Express got their preliminary approval way back in 2018. Right. MasterCard got theirs in February 2020. And notably, MasterCard did it through a joint venture with a local entity called Nets Union, rather than trying to just go it alone like Visa. A very smart strategic move on MasterCard's part. Clearly. Because Visa's been standing in this regulatory line the longest, and they are the absolute last ones to get to the front.

9:35And we really have to acknowledge what is waiting for them at the front of that line, too. It's not an empty room. Not at all. They aren't just competing with Amex and MasterCard for market share in China. They are stepping onto the home turf of the two absolute monopolies of the Chinese payment ecosystem. Alipay and WeChat Pay. Exactly. The final bosses. And for anyone listening who hasn't been immersed in that ecosystem, those two apps are literally the fundamental operating system for daily commerce over there. They run everything. Physical credit cards and point-of-sale terminals are practically archaic there.

10:08You scan QR codes through Alipay or WeChat for everything, From, you know, a high-end luxury purchase at a mall all the way down to a street vendor selling dumplings. Yeah, cash and plastic are just not part of the daily habit anymore. Right. So trying to convince a consumer base that is already perfectly optimized on mobile QR codes to suddenly adopt a Visa product that clears on a Western network, it feels like trying to sell fax machines to a population that exclusively uses iPhones. It's a massive technological mismatch. It is a paradigm mismatch, for sure. But we also have to look at the macro forces that likely dislodged this paperwork after nine years of just sitting there.

10:48Right. The timing is interesting. According to Reuters, this specific approval from the PBOC followed a diplomatic summit between Donald Trump and Xi Jinping. Yes. And executives from both Visa and MasterCard were notably in attendance at that summit. Which tells you everything you need to know. We are seeing geopolitics manifest in real time here. Yeah, we're looking at the raw facts of the timing, and it seems like a geopolitical thaw on the diplomatic stage often results in this very dry, bureaucratic movement behind the scenes. Exactly. A high-level meeting unblocks a regulatory choke point for foundational card networks.

11:24Even if FISA is just getting a preparation permit, merely getting a seat in the room in China is a massive geopolitical milestone. But it's so heavy, you know? It really highlights the sheer friction of the traditional fiat system. Oh, the friction is immense. You have one of the most powerful financial networks on Earth waiting nearly a decade for paperwork to clear. And they are highly dependent on diplomatic relations between heads of state, all just to attempt to compete in a highly fortified domestic market. It is slow, deeply regulated, and entirely bound by national borders. Right. So while Visa is playing this decade-long game of geopolitical chess just to get a learner's permit, there is an entirely parallel financial infrastructure being laid down that bypasses central banks and borders entirely.

12:11And it is moving fast. Very fast. We are moving from the legacy plumbing of traditional banking straight to the blockchain, specifically looking at a phenomenon where AI robots are getting their own crypto wallets. I know this sounds like absolute science fiction, but the software architecture for this is being deployed right now as we speak. Let's lay out the raw facts from the material. There is a New York-based wallet infrastructure company called Privy. They were founded in 2021, and they were quietly acquired by Stripe, you know, the massive payment processing giant, in June 2025. A huge acquisition that flew a bit under the radar.

12:49Totally. At the time of the acquisition, Privy was powering about 75 million accounts. What they specialize in is creating embedded wallets. And now they are bringing these embedded wallets to the XRP ledger, which allows apps and so-called AI agents to hold funds and execute transactions natively. It completely abstracts away the need for a user to manage complex cryptographic seed phrases. And just to understand why this matters for a second, look at the historical bottleneck of crypto adoption. The terrible user experience. Exactly. If you lose your 12-word seed phrase, your capital's unrecoverable.

13:26It's gone. Privy removes that friction entirely. They allow developers to build an app where creating a secure blockchain wallet happens in the background. It's as easy as logging in with a standard email address or, you know, a social media credential. Right, and I completely understand how that removes friction for a human user. But I want to dig into the AI agent aspect here, because the phrase giving a robot a wallet sounds like pure Silicon Valley marketing hype. It does sound like a buzzword salad. It really does. Like, how does a piece of software actually hold and spend money? And why wouldn't it just use a traditional bank API?

14:03Well, think about the limitations of our current banking infrastructure. OK. If you build an autonomous AI agent designed to act as your executive assistant, let's say you tell it, hey, book the cheapest flight to London, secure a hotel, and buy tickets to a conference. Right. That software needs to transact across multiple vendors instantly. Traditional banking APIs are incredibly clunky for autonomous software to navigate. Because they need all that human verification. Exactly. They require 16-digit card numbers, CVV codes, human identity verification, text messages with codes, and multi-day settlement times.

14:38Right, which breaks the autonomy of the AI. Completely. A blockchain wallet, however, is native internet money. An AI can hold a balance of digital dollars, execute a smart contract, interact with a vendor's API, and cryptographically settle a payment in three seconds. And it does all that without ever asking a human to pull out a piece of plastic. Exactly. Okay, conceptually, having internet-native money for internet-native software makes total sense. But we need to look at the actual adoption metrics provided in our source today, because frankly, they raise some serious red flags for me. Yeah, the numbers are a bit fuzzy.

15:14Ripple claims that 13 million agent payments have already been processed on the XRP ledger. And they are claiming this makes up roughly 25 % of the entire network's activity. And this is where we have to apply the exact same scrutiny we used on Citigroup's earnings report. Read the fine print. Right. You have to read the game tape, not the scoreboard. What's fascinating here is what is actually missing from that 13 million figure. Well, first off, the number is self-reported solely by Ripple. There is absolutely no independent on-chain verification provided in the text. None. But the truly glaring omission here is that there is absolutely no dollar value attached to that 13 million transaction.

15:56Precisely. Like, we know there were 13 million instances of software pinging the network, but we have no idea if they were transferring millions of dollars in actual economic value or simply moving fractions of a penny back and forth just to test the system. Furthermore, developers utilizing this Privy integration are not even forced to use the highly volatile XRP token to settle these payments. Oh, really? No. What else can they use? They have the option to settle in Ripple's RLUSD stablecoin. Which makes sense. For an AI agent trying to accurately price a hotel room, a stablecoin pegged to the U.S.

16:29dollar is vastly more practical than a wildly fluctuating crypto asset. Right, because you don't want the hotel room price to jump 10 % while the robot is booking it. Exactly. So bragging about a transaction count with absolutely no dollar figure attached, it's essentially like a wide receiver bragging about the number of catches he made in a football game while totally skipping over his total yardage. That's a great way to put it. If you catch the ball 13 million times, but you get tackled at the line of scrimmage every single time, you haven't actually moved the economic offense forward. Right.

17:00It is the definition of a vanity metric. Until these show us the actual capital flowing through those embedded wallets, calling it 25 % of network activity is basically just vibes with a decimal point. Vibes with a decimal point. I love that. But however, even if the current yardage is incredibly low, the fact that these players are on the field at all is a massive signal for where financial plumbing is headed. Because if you zoom out from the fuzzy math of the current metrics, the underlying infrastructure is being heavily capitalized by the most serious players in global finance. Oh, the heavyweights are all in.

17:32Yeah. Our sources mentioned the X402 Foundation. This is a consortium dedicated to building standards for these exact types of autonomous web payments. And look at who is sitting at that table. It's quite the guest list. You have the new guard, like Stripe and Ripple, sitting right next to the old guard, Visa and MasterCard. They all see the demographic shift in how transactions will be initiated over the next decade. They aren't blind to it. And even Citigroup, the same traditional bank we analyzed earlier that is currently bleeding margins on credit card rewards, they are simultaneously moving on chain.

18:05Yep. They are actively rolling out new token services and blockchain pilot programs in the UAE and Japan. So the legacy players know that the current system of plastic cards and airline miles is reaching its mathematical limit. And so they are aggressively hedging their bets by building the rails for this parallel borderless system. Exactly. It's incredible how tightly all of these seemingly disparate threads tie together. It really is. We started out today by questioning the illusion of the headline. And I think this deep dive into the material proved three very distinct realities about the modern financial system.

18:39Let's recap them. First, a bank generating record top-line revenue isn't actually winning if the underlying cost to defend their credit card market share is eroding their margins from the inside out. The fine print always wins. Always. Second, a massive nine-year diplomatic effort on the geopolitical stage can result in a mere preparation permit, which really highlights just how slow and fractured the traditional fiat system truly is. It's agonizingly slow. And finally, the infrastructure for autonomous AI payments is actively being laid down by the biggest companies on earth, even if the current transaction metrics are heavily inflated by marketing hype.

19:15I think the overarching takeaway for you, the listener, is just an appreciation for the sheer systemic depth of every transaction you initiate. It's easy to take it for granted. It is. The next time you strategically use a credit card to optimize a cashback reward, or you read about a company expanding overseas, or you use a piece of AI software to automate a digital task, remember the incredibly complex, highly regulated, and immensely costly financial plumbing that has to operate flawlessly just beneath the surface to make that instantaneous moment possible. It is never just a simple swipe. Which leads me to a final thought I want to leave you with today.

19:57Go for it. Something for you to just mull over that wasn't explicitly spelled out in our sources, but it feels like the inevitable collision of everything we just discussed. I'm intrigued. We talked extensively about how banks like Citi are structurally trapped. They are bleeding money to maintain these lavish credit card rewards programs. Airport lounges, concierge services, inflated points multipliers. Right, the whole arms race. And they do all of this for one reason, to psychologically entice human loyalty. It is a game entirely designed to make a human being pull their specific piece of plastic out of a physical wallet.

20:33Yes, human psychology. But if Privy and Stripe are successful and these embedded wallets become the standard, what happens when AI agents take over a significant portion of our daily purchasing? Wow. Okay. Yeah. If you delegate your subscription management, your travel booking, and your online shopping to an autonomous piece of software, how does that completely upend the multi-billion dollar concept of credit card rewards? That's a fascinating question. Right. Like, would an autonomous AI agent care about exclusive access to a Delta Sky Club? Obviously not. Exactly. Would a piece of software be swayed by the prestige of holding a heavy metal credit card?

21:17Or will it just mercilessly, mathematically route every single transaction through whatever network offers the absolute lowest fraction of a cent in settlement fees? It's going to go with the math every time. If the robots are the ones actually executing the spending, the entire scoreboard for human financial loyalty might not just be an illusion. It might become completely obsolete. That is something to think about. Thank you so much for joining us on this deep dive. keep questioning the fine print, keep looking past the headlines, and we will catch you next time.

From the publisher

Earnings season is open, and Citi managed to beat estimates and still get sold. We also dig into Visa’s nine-year wait for a yes from China’s central bank and what Stripe’s Privy is really doing with AI agents on the XRP Ledger. Short on hype, long on numbers.


More from Fintech Stacks

All episodes