EMVCo (the technical body behind Visa, Mastercard, Amex) is creating global standards for "agentic payments."
This is the biggest change in card payments since "tap to pay"
Here's how it works 🧵
Right now, AI agents are phenomenal at finding things to buy.
- Power users are starting to default to their research
- Can compare complex options and summarize
- And when people click through conversion is 2x to 5x higher
But...
There's no agreed way for payment to happen
- There's countless protocols
- x402 for agents accessing other tools
- ACP and A2P from Open AI and Google
- Visa and Mastercard have their own approaches
This is the XKCD standards problem playing out in real-time: "There are 14 competing standards.
Let's create a universal standard that covers everyone's use cases. There are now 15 competing standards."
EMVCo are the people who set the standards for how debit and credit cards work globally
Now they're creating a "passport" for authenticated agents.
- Use your Face ID
- FIDO token generated (like a passkey)
- Given to agent
Agents present a uniform cryptographic token that proves "I'm a good bot, authorized to spend."
The merchant firewall and fraud systems see the signature, verifies it came from a legitimate issuer, and lets it through.
For authentication, they're extending FIDO/SRC standards:
- You authenticate once with your face to "bind" an agent to your card, - Set spending limits, - Then the agent presents a delegated token at checkout. - No 3DS challenge needed (!!!)
This prevents the agentic commerce world from fracturing into incompatible silos.
It ensures an agent built on Microsoft/OpenAI can pay a merchant using Adyen, authenticated by an issuer on Marqeta.
Critically it creates a new liability framework (!!!)
- If merchants use the EMV standard, issuers take fraud risk.
- If they allow random bots, merchants eat the cost.
We're about to see a new transaction category emerge: "Agent Present."
- Not card-present.
- Not card-not-present.
- Agent-present.
With its own interchange rates, its own fraud rules, its own liability shifts.
Are standards boring? Absolutely.
Are they critical?
More than almost anything else happening in payments right now.
PS. Notice the complete lack of stablecoin mentions?
Me too.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
JPMorgan clients can now swap JPMD for USDC on Base.
That sentence should break the internet.
JP Morgan payments moves $ trillions PER DAY
It dwarfs the entire stablecoin industry.
This is how 1000x more dollars go onchain 🧵
Picture the actual flow:
- A JPMorgan institutional client holds JPMD (bank deposit token).
- They need to transact with a Coinbase customer holding USDC (stablecoin).
A corporation could
1. Move JPM Coin from JPM closed loop to Base 2. Swap JPM Coin for USDC 3. Receive USDC in their base wallet 4. Send that USDC to a 3rd party, or swap it for another bank deposit token
This is the baby step towards going open loop.
Banks have tokenized deposits in closed loop
- Citi
- HSBC
- Deutsche
- JP Morgan
But now those walls have doors that open onto public blockchain rails.
Base becomes the trading floor where closed systems meet open ones.
Stablecoins solve a bottleneck in the internet economy.
20th-century money is too slow, expensive, and infrequent for the demand of internet-scale payments.
This is a pattern that repeats in history.
🧵
1. The Industrial Revolution
- The Royal Mint couldn't create coins fast enough
- The shortage led to widespread counterfeits
- The new wage economy demanded more coins
- So factories with high quality machinery made their own
The Royal Mint accepted this before eventually USING that technology themselves 50 years later
2. The Railroad Boom
- The centralized banking system couldn't provide local capital
- Delaying western expansion and railroad build out
- States passed "free banking" laws
- Local banks set up with reserves at the state
This was tolerated until the 1860s where national charters and centralized money printing and control
The bank lobby is furious about stablecoin yield under the GENIUS Act. They're calling it a "loophole" that needs closing.
But here's what they're missing: We've seen this movie before. And it built an entire generation of fintech companies.
🧵
The GENIUS Act prohibits stablecoin issuers from paying interest directly to holders. Banks claim issuers are skirting this by paying third parties (like exchanges), who then offer rewards or yield to users.
Treasury estimates this could drain $6.6 trillion from bank deposits.
But let's reframe what's actually happening here.
Stablecoin issuers earn yield on reserves (mostly T-bills at 4%+). They keep some, pass most to distributors. Distributors use some for operations, spend some on customer acquisition through rewards.
The world's first 50% stablecoin IPO just happened. Crypto exchange Bullish received $1.15bn in USDC.
This quietly changes everything about how public companies can raise capital.
What actually happened:
• Bullish (NYSE: BLSH) closed their IPO on August 14th
• 50%+ of proceeds came as stablecoins ($1.15bn total)
• Settlement across 8 different stablecoin types
• Majority minted on Solana, custodied by Coinbase
Why this matters - IPO's become more global:
- Traditional IPO is single currency, single jurisdiction
- This IPO had USD + EUR stablecoins from US, Europe, Asia
- The stablecoins also settle instantly (not T+2)
Nubank's results are INSANE. Every other bank CEO must look at these and be like... HOW?
Here's the breakdown...
* 122.7 million customers (+4.1M net additions)
* $3.7 billion revenue (+40% YoY)
* $637 million net income (+42% YoY)
* $12.2 monthly revenue per active customer (+18% YoY)
* $0.80 cost to serve per customer
* 83.2% monthly activity rate
That's a benchmark every other organization in finance should print out on their wall. Only webank in China (with 494m users) can beat.
The unit economics *almost* don't make sense:
- $0.80 cost to serve each customer
- $12.20 revenue per customer/month
- That's 15x return 🤯
Most banks struggle to hit 3x - That's the benefit of self-owned technology and a branchless servicing model.
Geographic Split:
* Brazil: 107.3M customers (60% of adult population)
* Mexico: 12M customers (13% of adult population)
* Colombia: 3.4M customers (10% of adult population)
That says to me, the newer markets are taking longer to penetrate. Where's the next growth engine coming from? Not many 200m + populations around 👀