Stablecoin Affinity Cards - Amazon Sues Perplexity & Robinhood Earnings

Hear me out: Stablecoin-Co-Branded Cards 

This week I showed a Fortune 500 CMO a stablecoin card who got wildly excited and said, 'I could do a co-brand for everything with this.' 

Co-branded cards (like Airline or hotel rewards credit cards) are a money-printing machine for brands with high-affinity customers. Giant brands have used them for decades to manufacture loyalty and generate high-margin revenue.

This model is costly, relies on a banks underwriting, and risks brand damage if it fails. For example, Apple Card's partner Goldman Sachs faced an $89 million CFPB fine and allegations of sexism due to their "black box" credit model, for unequal credit limits to female spouses, causing a PR disaster.

Fintech didn’t fix this. Neobanks, Robinhood, and BNPL providers are innovating with new card types, but there are still 1,000s of consumer brands that found co-branding too complex or a bad business case.

Which leaves a gap big enough to drive a blockchain through.

So, Stablecoins?

I feel your cynicism, dear reader. So put the pitchfork down and go with me here.

(Need a primer on Stablecoin Linked Cards? Click here)

Co-branded Cards. Loyalty Programs that Print Money?

If you’re big enough to matter to the banks and can handle the reputation risk. Then yes.

Co-branded cards are fundamentally loyalty programs disguised as financial products.

The concept evolved in three main era’s

• 1980s, The concept was pioneered by airlines. They had a massive, loyal customer base (frequent flyers) and a high-value, perishable reward (miles/seats). They partnered with banks to "sell" these miles, creating a card that turned everyday spending into travel. Hotels quickly followed suit.

• 1990s-2000s, Brands with high-affinity, high-frequency purchase cycles (like department stores and warehouse clubs) launched cards. The most famous "walled garden" example is the Costco card (first with Amex, now with Citi), which for a long time was the only credit card accepted in-store, forcing adoption.

• 2010s-Present, E-commerce and digital ecosystems. The Amazon Prime Visa is a perfectly example, integrating into the Prime membership ecosystem and driving loyalty with a simple, high-value proposition (5% back). 

The banking-as-a-service / credit-as-a-service era has delivered a lot of credit innovation in co-brand adjacent areas like

• Niche brands. Too small for Chase or Amex. Think the "Cleveland Cavaliers" card or a "creator" card.

• Fintechs Targeting the Underserved: CaaS was (and is) a powerful tool for companies targeting "thin-file" customers. The Chime credit-builder is a classic example.

• Embedded Lending (BNPL): This is arguably the most successful CaaS story. Companies like Affirm and Klarna, and even "white-label" BNPL providers, transaction linked lines of credit

We now see cards for everything from non-profits (donating a percentage of spend) to pet stores. What we don’t see is the giant name brands. This is limiting the potential of co-brand cards especially for the mid-sized merchants, those outside the top 20, but national or regional. You have a market where Fintech attacked the whitespace, and Co-brand at best, works by exception not the rule.

The Apple Card Disaster

Apple Card had its first major public crisis in 2019, shortly after launch. It centered three main issues.

• The "Sexist Algorithm" Allegation: A viral Twitter thread from tech entrepreneur David Heinemeier Hansson alleged the card's algorithm was "sexist." He received 20x the credit limit of his wife, even though they filed joint tax returns and she had a higher credit score.

• The Brand Damage: Apple co-founder Steve Wozniak chimed in, saying he had a similar experience (he got 10x the credit of his wife). The public narrative wasn't "Goldman Sachs has a bad algorithm"; it was "Apple's new card is sexist."

• The "Black Box" Problem: When customers called to ask why their limits were different, customer service (run by Goldman) had no answer. They infamously blamed "the algorithm," which was a "black box" they couldn't explain.

This was a nightmare for Apple, a brand built on transparency and user-first design. Regulators launched an investigation. While Goldman was eventually cleared of unlawful discrimination, the New York regulator pointedly stated that the bank's "deficiencies in customer service and a perceived lack of transparency undermined consumer trust.

Goldman, after losing over a billion dollars (because Apple demanded no-fee, mass-market approvals that led to subprime-level losses), the entire partnership was hit with an $89 million fine from the CFPB for compliance failures.

You get the revenue, but you give up control. You tie your brand to a partner whose core business, risk management, requires them to reject your customers. It's a fundamental misalignment that puts your brand reputation in the hands of a call center and a FICO score.

Apple's demands included:

1. No Fees: No annual fees, no late fees, and no foreign transaction fees, stripping Goldman of standard bank revenue streams.

2. Mass-Market Approval: Apple wanted the card to be accessible to its entire customer base, not just premium users. This reportedly pushed Goldman to approve customers with FICO scores below 660, leading to the worst loss rates on credit card loans among all major U.S. issuers—even worse than some subprime lenders.

The fundamental tension?

Brand Control vs Bank Unit Economics. 

The Apple Card was the worst of all worlds here, where the bank had poor unit economics and the brand had lost control of the customer experience. Surely then, Fintech fixes this? Not exactly. Fintech does something else.

Fintech hasn’t disrupted co-branding, but it has expanded credit.

The 2010s BaaS/CaaS wave (think Stripe Issuing, Marqeta, and the "Sponsor Banks" like Bancorp, Stride, and Cross River) was about access, not premium rewards.

It was a debit-first revolution that allowed neobanks (like Chime) and tech companies (like Lyft and DoorDash) to easily issue debit cards for payouts and basic banking. It was about embedding payments.

When "Credit as a Service" (CaaS) emerged, it still didn't compete with the Chase/Amex model. Here’s why the big bank partnerships have endured, which we can call the "Three B's."

1. The Balance Sheet (This is the #1 Reason)

• The Problem: A premium co-brand card portfolio is a multi-billion-dollar loan book. When Amex and Delta launch a card, Amex is committing its balance sheet to fund $10B, $20B, or $50B+ in consumer credit.

• The BaaS/CaaS Limitation: The 2010s model ran on "Sponsor Banks" (e.g., small community banks in Utah or South Dakota). These banks did not have the capital or the risk appetite to fund a $50B loan portfolio for a brand like Starbucks or Apple. They were built to handle transaction volume, not massive credit lines.

2. The Economics

• The Big Bank Model: As the CFPB noted, big banks pay billions to their partners. In 2022, they sent over $28 billion to their co-brand partners. This includes massive signing bonuses, marketing commitments, and a rich share of the revenue.

• The BaaS/CaaS Limitation: A BaaS provider and its small sponsor bank cannot compete economically. They can't pay a $1B signing bonus to an airline.

3. The "Black Box" (Risk & Compliance)

• The Big Bank Model: For a global brand like Marriott, the bank's "black box" (their underwriting, risk, and compliance department) is a feature, not a bug. It means Amex is legally responsible for KYC/AML, the Truth in Lending Act, Fair Credit Reporting, and managing delinquencies and fraud. 

• The BaaS/CaaS Limitation: This model still has the "Apple Problem." The brand (e.g., Apple) still has to partner with a bank (e.g., Goldman). Even if it's a "fintech" provider, that provider still has a sponsor bank in the back that controls the underwriting. A loyal customer still gets rejected. The brand still gives up control. 

The 2010s BaaS model didn't solve this; it just put a friendlier API in front of the same fundamental problem. “Credit as a Service” isn’t a solution for giant brands, and it's incredibly difficult to implement, even as a Neobank, because you’re at the mercy of your bank partners risk appetite.

The US has 1000s of banks, and many stepped up to do the hard work on risk management to solve unmet needs, and that’s why the new products have emerged.

The Gap: Brand Control Without Bank Constraints

But this leaves a gap. What if we could have more brand control over the financial product? With the modular building blocks stablecoins provide, could we get more creative in the financial product offering, and how that shows up to consumers? 

Stablecoin-linked Co-Branded Cards

What if you could launch a card where the only underwriting question was: 'Does my customer have $50 in their stablecoin wallet/vault?’

Your customer pre-funds a stablecoin wallet (a bit like a Starbucks app wallet), and they could spend from that balance using a card. 

This is what non-custodial stablecoin cards do. The card is a debit card linked directly to a user's on-chain digital wallet. If they have the USDC, the transaction is approved. Instantly. 

The customer sees your brand, your card, and you could build some incredible rewards/activations around that card and spend. This is prepaid but 100x better. Because they’re spending from their cash-like balance, not something on your P+L. 

Imagine

• Paying a reward of 3 to 10% of balances held in your card as “points” that can be spent with your brand (or partners). Funded by stablecoin yield, not your P&L.

• Give discounts to users who spend with that stablecoin on your platforms

• Letting users borrow against their balance of your stablecoin to make purchases elsewhere.

All with the revenue engine of a co-brand card, where as an “issuer” of that card “swipe fees” (interchange) flip from being a cost center to a revenue line item.

As a CMO, I’d think about if this is a lighter way to do affinity cards, or a new way to expand customer loyalty. 

Stablecoin Card Affinity: A New Option for Brands.

For 40 years, only airlines and hotels could print money with co-brands. Stablecoins just made that possible for everyone. The global co-branded credit card market hit $14.6 billion in 2024 and is growing at 9.8% annually toward $25.7 billion by 2030. But I there’s a ton of whitespace that could accelerate that.

There are almost too many choices for how you build an affinity program with stablecoins. By default, you typically have much more control because stablecoins are, by nature, modular. But the choice is on risk appetite. How rewarding do you want this affinity card to be, and how much economic risk are you willing to expose users to.

Two Paths Forward

1. The Stablecoin Debit Card

• Competes with: Debit and prepaid cards

• You monetize: Yield on float + interchange fees  

• Risk profile: Low (fully collateralized by user deposits)

• User experience: Prepaid without the stigma

Stablecoin debit cards solves the "Co-Brand Trap" and gives brands a simple, profitable, and safe way to launch a co-brand card.

2. The Collateralized Credit Card

• Competes with: Premium credit and wealth management

• You monetize: Yield on deposits + interchange + borrow fees

• Risk profile: Higher (depends on collateral volatility)

• User experience: Spend yield, keep principal

The Collateralized Credit Card model is wild:

1. The "Vault": The user deposits a yield-bearing asset, like a liquid-staked eETH token (currently earning ~9% yield).

2. The "Borrow Mode": When the user swipes the card, they don't sell their asset. They instantly borrow fiat against it (at ~4% APY).

The user is living off the spread. They are spending their yield without ever touching their principal. This brings the "borrow-against-your-assets" strategy—a trick the ultra-wealthy have used for decades to avoid taxes and compound wealth—to anyone with a smartphone.

The health warning: In a bear market, that 9% yield could flip to 2%, and the 4% borrow rate could spike to 8%. Living off the spread works until it doesn't.

This gives you a new way to build a co-branded program with credit if you wanted to. But as a chief compliance officer I’d scream will someone, please, think of the risks!

New Cards - New Risks - New Providers to Manage Them

Now, this sounds amazing. But ask your chief compliance officer and watch their face drain of color. Four words that end the conversation: 'What about hacks?’

Let’s assume we’re talking pure DeFi here. If you’re not working with intermediaries or providers who help you (or your customers) manage this risk. You’d face:

• Technical Risk: "What if the smart contract gets hacked? In the old model, the bank eats the fraud. In this model, the customer's funds could be gone. Forever."

• UX Risk: "What if the customer loses their 'private key'? Same problem. Their money is gone. They will call your customer service line, and you'll have to say 'Sorry, we can't help.'"

• Economic Risk: "What if the stablecoin 'de-pegs' and its $100 becomes $80?"

• Regulatory Risk: "Isn't this the 'wild west'? Are we going to get sued by the SEC?"

This is why most brands stop. They assume they have to be DeFi security experts. But here's what they miss: You don't have to be a DeFi security expert any more than Apple had to be a bank.

You partner with the stack that has already solved this. 

Example stablecoin providers.

• Stablecoin infrastructure: Companies like Rain or Bridge (via Stripe) can help issue cards to end users in 10+ countries. They partner on the back end with countless other 3rd parties to make it so the brandintegration becomes one API.

• Wallet-as-a-service / Custodians: Fireblocks, Dynamic, Turnkey, Privy help manage the UX risk, by removing the complexity of managing keys, and building failsafes and fallback models if a key is compromised. 

• Vaults & RWAs: Morpho, Veda, and a handful of others help users deposit stablecoins, earn yield, or borrow against that yield. They also offer tokenized “Real World Assets” like US Treasuries. 

• Issuance-as-a-service / Bring your own stablecoin: Trust companies and chartered issuers help turn regular cash into stablecoins or tokenized deposits for banks and non-banks with “mint” and “burn’ capabilities (and fees). 

This provider ecosystem means a brand can have many more dials to tweak if they want to control the experience. Want to issue your own stablecoin? Or maybe you don’t want to touch crypto at all? Or you want to do a fancy loyalty/redemption thing. It’s all theoretically possible. 

The control issue Apple had is gone, but in its place is an ecosystem that’s less mature and battle-hardened. Here’s how I’m thinking about the new set of tradeoffs.

It’s not easy, it's not obvious, but it’s possible.

What should you do?

That health warning is real. Yield in DeFi is just a real-time price for risk.

But that CMO I mentioned wasn't excited about yields. She was excited about control.

The old "Three B's" (Balance Sheet, Bounty, Black Box) that locked brands out? Stablecoins rewrite all three

1. The New Balance Sheet: In the Collateralized Credit model, the protocol is the balance sheet. Aave and Morpho are global, permissionless liquidity pools. 

2. The New Bounty: In the old model, the "bounty" (rewards, miles, cashback) is a massive liability on the brand's P&L or a huge rev-share given to the bank. Now it could actually be an incentive payment and stablecoin yield.

3. The New Black Box: The old model forced brands to accept a bank's "black box" of underwriting.

The new model gives the brand a choice.

• For the Stablecoin Debit card, it partners with the "Solvable Stack" to abstract the complexity, giving the user a simple, safe, "Fintech at the Front" experience.

• For the Collateralized Credit card, the "black box" becomes a feature. The brand's job isn't to hide the tech; it's to curate it—to be the trusted guide that vets and selects the safest, highest-yield protocols for its community.

This is the future of co-branding.

It’s a model where the brand has true control, not just over the UX, but over the economics. 

It’s a model that turns a loyalty program from a liability into a yield-generating asset.

And it's the first model that truly breaks the "Co-Brand Trap" for good.

ST.

4 Fintech Companies 💸

1. Ornn - The Compute Commodity Exchange

Ornn produces a single index for Nvidia H100 compute price per hour for things like AI training and inference compute. GPUs have a massive supply/demand imbalance. By creating a reference price for contract settlements, buyers can “lock-in” a future price of compute per hour

🧠The dollars spent on compute will exceed the dollars spent on oil over the next 10 years. (Says Don Wilson of DRW Trading). Every CFO building data centers or trying to budget for compute, could find this useful. There will be challenges as new breakthroughs in GPUs happen a lot faster than breakthroughs in say, types of energy discovered or types of wheat. But I like the idea of how oil futures helped us manage the industrial economy, and compute futures helping us manage the intelligence economy. Intuitively, that makes sense.

2. ZAR - The Uber for Stablecoins

ZAR is building a network of cash-for-stablecoin agents in the global south. It arms agents with a simple QR code. Users can take cash to any agent and receive dollars in return to a crypto wallet. That wallet can then have a stablecoin-linked card issued. Meaning the local user avoids the risk of hyperinflation and can spend anywhere Visa is accepted.

🧠Building a bottom-up agent network is audacious. The aim is to bring transparency by having agents publish FX rates on the platform that users can pick from (like Uber’s ride price estimate). Merchants can apply directly in the app, and then users can find the store in the map displayed to them. There’s also an ambassador scheme, where those who recruit new merchants get a share of the platform fee. 

3. Loon - The Canadian Stablecoin Dollar (CADC)

Loon is a Canadian Dollar stablecoin that is fully reserved $ with 1m CADC TVL and $200m of transactions historically. It operates under Canadian regulations and frameworks, It publishes monthly attestations of being 100$, 1:1 backed by Canadian dollars directly.

🧠Since its dollars and not treasury I wonder what their yield looks like. There isn’t yet a GENIUS Act equivalent in Canada, but like a lot of the G20, they’re reacting quickly and expect to publish something next year.

4. Leverage - Institutional-grade trading for consumers.

Leverage is a mobile app aimed at consumers that allows up to 40x leverage across DeFi marketplaces and protocols. To date the app has more than 8,000 users who have traded $600m in notional. 

🧠 Is this the Interactive Brokers of crypto? IBKR has dominated the pro-sumer space in brokerage at the top end and Robinhood for the mid-level. But these companies often aren’t pushing the boundaries of price or leverage in crypto markets. There was a gap in the market for something along these lines. As all these things, caveat emptor.

Things to know 👀

1. Amazon is suing Perplexity to Stop its Agentic Browser Buying Stuff 

Per Bloomberg: “Amazon.com Inc. is suing Perplexity AI Inc. to try and stop the startup from helping users buy items on the world’s largest online marketplace” The accusation is of computer fraud by “failing to disclose” that the agentic browser (Comet, an alternative to Chrome with agents baked in) is shopping on a persons behalf in violation of its terms of service. Perplexity argues Amazon is simply bulling and intends to launch its own competitor browser and tooling to Perplexity. Amazon says the focus is on disclosure.

🧠Agentic browsers are a giant security risk to merchants. Amazon has a point. To a merchant a browser driving itself looks like the type of bot a fraudster uses to create fake items or buy stuff with stolen credentials. Agentic Browsers are also wildly easy to hack with “prompt injection.”

🧠Large merchants potentially have the most to lose in agentic commerce. If an agent can find you a better price or product somewhere else, the value of being the everything store diminishes. Browser-based agents are hostile to merchants' customer relationships without some way of retargeting that user.

🧠Google is said to be working on a browser agent, I imagine they’d spend more time with big merchants before launching it. And they have a ton of users already in Google Chrome.

🧠The outcome of this lawsuit could shape how all of agentic commerce goes. If Amazon wins, it will create a need for all agents to disclose which human they’re shopping on behalf of securely. The internet hasn’t agreed how we do that yet. 

🧠We need a clean protocol for agent to merchant comms. Google’s A2P and OpenAI/Stripe’s ACP are the frontrunners here. There’s also the card networks, and the messy job of stitching all of this together (which is what PayOS does).

2. Robinhood delivers 100% YoY Revenue growth

Robinhood’s latest earnings had some incredible numbers, $1.27B Revenue, +100% YoY, $556M Net Income, +271% YoY, $0.61 EPS, +259% YoY, $742M Adjusted EBITDA, +177% YoY, 3.9M Gold Subscribers, +77% YoY, 26.8M Funded Customers, +10% YoY, $333B Platform Assets, +119% YoY

🧠 A jump in crypto activity drove the earnings beat, but that’s the least interesting part. The platform assets are up nearly 120%, and Gold subscribers at nearly $4m. This is a much sticker, long term form of revenue.

🧠 So yes, Robinhood is benefiting from a bull market cycle, but it’s also continuing to diversify. It’s now a full-blown Neobank that competes on one side with Chase Sapphire, and on the other side with Coinbase and then brokerage. A very unique animal.

🧠 They’re going deep on prediction markets, and I’ll be honest, I don’t love how specifically sports are being correlated with “investments.” There’s some serious consumer harm risk here in blurring these lines. Overall, I like prediction markets as a data engine (as I’ve written before), but a sportsbook is a sportsbook.

Good Reads 📚

1.The Political Compass for Human Flourishing

The historical political compass of authoritarian/liberal vs conservative/socialism is out of date for our new reality. Today its power/impotence, vs extraction/generation. The extractive impotent version of society is Wall-E, Memecoins, limbic capitalism (sounds familiar), techno optimism is the leadership that extracts that value. Meanwhile the impotent generative world view is the regressive, backward looking activist.

🧠 The picture below communicates it perfectly. I suspect most readers, and most of us, want a future where tech enables human flourishing. But the way we’ve designed our economies haven’t allowed us to progress that way. 

🧠 We need a new game. We need a way less impotent version of optimism. We need to make it less fashionable to be horrified by the world. 

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—strong opinions weakly held

(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