Stripe is in the news (again). With another bump in their valuation, the company is now privately valued at $159 billion. The news comes when their new strategy is starting to consolidate. Their new stablecoin infrastructure isn’t just another crypto play—it’s a complete reimagining of how money moves in the digital economy.

The proof? In the last 24 months, they spent over $1.2 billion acquiring the pieces: Bridge for $1.1 billion and Privy for hundreds of millions. But here’s what most people missed: they’re also building their own blockchain called Tempo. This isn’t experimentation—it’s execution of a master plan.

When the company that processes nearly a trillion dollars annually makes this kind of bet, they’re not following trends. They’re creating them.

The Four-Layer Money Stack

Stripe’s approach breaks down into four distinct layers, each solving a critical piece of the stablecoin puzzle:

Layer 1: The Trust Layer

This is where most stablecoin projects fail. Stripe partnered with established stablecoin issuers like Circle (USDC) rather than creating their own token. Smart move. They’re essentially saying: “We’re not here to reinvent money—we’re here to make it work better.”

The trust layer handles the messy questions: Is this stablecoin actually backed by dollars? Who’s auditing the reserves? What happens if the issuer goes under? Stripe’s answer: Let the specialists handle stability, we’ll handle utility.

By partnering instead of competing, Stripe avoided the regulatory minefield of stablecoin issuance while gaining access to the most liquid and trusted digital dollars in the market. Circle’s USDC alone has over $35 billion in circulation and regulatory clarity in major markets.

Layer 2: The Infrastructure Layer (Powered by Tempo)

Here’s where Stripe’s secret weapon comes into play: Tempo, their proprietary blockchain infrastructure. While everyone was focused on the Bridge acquisition, Stripe quietly built their own high-performance blockchain optimized specifically for stablecoin transactions.

Tempo isn’t trying to be Ethereum or Solana. It’s purpose-built for one thing: moving money fast, cheap, and reliably. Think of it as AWS for stablecoins—highly available, globally distributed, and optimized for financial workloads.

The key insight? Existing blockchains weren’t built for the payment volumes Stripe processes. When you handle $800 billion annually, you need infrastructure that can scale without compromising on speed or reliability. Tempo gives them that control.

Ethereum processes about 1.2 million transactions daily. Stripe processes over 2 million. The math doesn’t work on public blockchains—especially when your customers expect 99.99% uptime and sub-second response times.

Layer 3: The Security & Compliance Layer (Enhanced by Privy)

The Privy acquisition solves the authentication puzzle. Stablecoins require private key management, but businesses can’t handle the security implications. Privy’s embedded wallet technology lets users interact with stablecoins using familiar login methods—no seed phrases or hardware wallets required.

This removes the biggest barrier to stablecoin adoption: user experience. Instead of managing private keys, users authenticate through methods they already know—email, phone, biometrics. Privy handles the cryptographic complexity behind the scenes.

This layer also handles Stripe’s regulatory moat. They’ve spent 14 years building relationships with financial regulators worldwide. That infrastructure now extends to stablecoins, handling KYC, AML, and cross-border compliance automatically.

Regulatory compliance isn’t just about following rules—it’s about building trust. When Stripe processes a stablecoin payment, it carries the same regulatory rigor as a traditional card transaction. That matters to enterprise customers who can’t afford compliance gaps.

Layer 4: The Experience Layer (Bridge + Stripe Checkout)

This is where the Bridge acquisition becomes critical. Bridge built the best stablecoin API infrastructure in the business—handling multi-chain compatibility, wallet management, and blockchain abstraction. But now it’s supercharged by integration with Stripe’s existing checkout infrastructure.

Imagine checking out on any e-commerce site. Today, you see “Pay with card” or “Pay with PayPal.” Soon, you’ll see “Pay with USDC” right alongside those options. Same familiar checkout flow, but settlement happens in minutes instead of days.

Bridge’s infrastructure handles the technical complexity—automatic gas fee optimization, transaction batching, seamless conversion between different stablecoin standards—while Stripe Checkout handles the user experience. The result? Stablecoins that feel as simple as credit cards.

The integration goes deeper than just adding another payment method. Stripe can automatically convert between traditional currencies and stablecoins at checkout, handle tax calculations across jurisdictions, and provide the same dispute resolution merchants expect from card payments.

Why Tempo Changes the Game

Building their own blockchain might seem like overkill, but it’s actually genius. Here’s why:

Performance Control: Stripe can guarantee transaction speeds and costs because they control the entire stack. No more waiting for Ethereum gas fees to spike or network congestion to clear. They’ve essentially built a private highway for money movement.

Regulatory Clarity: Operating their own blockchain gives Stripe more control over compliance. They can build regulatory requirements directly into the protocol rather than working around existing blockchain limitations. When regulators ask how transactions are processed, Stripe can show them the code.

Integration Depth: Tempo can be optimized specifically for Stripe’s existing products. Automatic reconciliation with traditional banking rails, seamless conversion between stablecoins and fiat, native integration with Stripe’s fraud detection systems. This isn’t possible when you’re building on someone else’s blockchain.

Economic Model: On public blockchains, transaction fees go to miners or validators. On Tempo, Stripe captures that value directly. At their transaction volumes, we’re talking about hundreds of millions in potential revenue.

The Numbers That Matter

Stripe processes over $800 billion annually. If even 5% of that volume shifts to stablecoins, we’re talking about $40 billion in new stablecoin transaction volume. That’s roughly doubling the current market.

But here’s the real kicker: Stripe’s international payment volume is growing 50% year-over-year. Traditional cross-border payments are expensive and slow—exactly what stablecoins solve. The addressable market isn’t just existing volume, it’s unlocking entirely new use cases.

Consider the math on cross-border payments:

- Traditional wire transfers: 2-5 days, 3-7% fees

- SWIFT payments: 1-4 days, 1-4% fees plus FX spreads

- Stablecoin payments via Stripe: Minutes, sub-1% fees

For a $100,000 international payment, businesses could save $2,000-6,000 in fees and get their money days faster. Multiply that across millions of transactions, and you understand why Stripe bet $1.2 billion on this infrastructure.

The Stripe Advantage

Other companies are building stablecoin infrastructure too. So why does Stripe win?

Distribution: They already have relationships with millions of businesses worldwide. Those businesses trust Stripe with their money. That trust transfers to stablecoins. When Stripe adds stablecoin support to their dashboard, millions of merchants get instant access.

Full Stack Control: From blockchain (Tempo) to security (Privy) to APIs (Bridge) to user experience (Checkout), Stripe owns every layer. That integration creates competitive moats competitors can’t easily replicate. It also means faster innovation—they don’t need to coordinate with external partners to ship new features.

Regulatory Relationships: Getting money transmitter licenses in all 50 states isn’t just expensive—it’s a multi-year process requiring millions in compliance costs. Stripe already paid that price and built those relationships. New entrants have to start from zero.

Network Effects: The more businesses use Stripe’s stablecoin infrastructure, the more valuable it becomes. Businesses that receive USDC payments through Stripe can more easily make USDC payments to their suppliers. This creates a self-reinforcing cycle that’s hard for competitors to break.

What This Means for Business

If you’re running a business that moves money internationally, pay attention. Stablecoins through Stripe could cut your settlement times from days to minutes and your fees from 3-5% to under 1%.

But more importantly, this infrastructure makes new business models possible:

Instant Global Payroll: Pay remote workers anywhere in the world, instantly, without traditional banking delays. No more explaining to contractors why their payment is “processing” for a week.

Real-time Revenue Sharing: Automatically distribute revenue to partners the moment transactions clear. Imagine affiliate commissions paid instantly instead of monthly.

Programmable Payments: Smart contracts that execute automatically based on business logic. Pay suppliers when goods are delivered, release funds when milestones are met, split payments between multiple parties automatically.

Micro-transaction Economics: Traditional payment rails make small transactions uneconomical. Stablecoins enable new business models around micro-payments, subscription tiers under $1, and pay-per-use services.

The Competitive Landscape

Stripe isn’t alone in this space. PayPal has PYUSD, Visa is building stablecoin settlement rails, and dozens of crypto-native companies are fighting for market share. But Stripe’s approach is different in three key ways:

Enterprise Focus: While others chase consumer adoption, Stripe built enterprise-grade infrastructure first. Their customers need reliability, compliance, and integration depth—not flashy features.

Infrastructure Investment: The $1.2 billion in acquisitions plus Tempo development represents the largest single bet on stablecoin infrastructure by a traditional fintech company. That level of investment creates capabilities competitors can’t match quickly.

Proven Execution: Stripe has a track record of taking complex financial infrastructure and making it simple for developers to use. They did it with online payments, payouts, and subscription billing. Stablecoins are the natural next step.

The Regulatory Wild Card

The biggest risk to Stripe’s stablecoin strategy isn’t technical—it’s regulatory. Stablecoins exist in a gray area in many jurisdictions, and rules are evolving rapidly.

But this is where Stripe’s 14-year head start matters. They’ve navigated regulatory challenges in dozens of countries, built relationships with central banks and finance ministries, and proven they can operate complex financial infrastructure compliantly.

More importantly, Stripe’s approach reduces regulatory risk by partnering with licensed stablecoin issuers rather than issuing tokens themselves. They’re the infrastructure layer, not the money issuer. That distinction matters to regulators who are still figuring out how to classify different crypto activities.

The Global Implications

If Stripe succeeds, they won’t just change how businesses move money—they could reshape global financial infrastructure. Stablecoins that settle instantly and cost almost nothing to transfer could reduce friction in international trade, enable new forms of economic cooperation, and extend financial services to underserved markets.

Consider the implications for developing economies where traditional banking is expensive and unreliable. A small business in Kenya could receive payment from a customer in Germany and immediately pay suppliers in China—all settled in minutes for minimal fees. That’s not possible with traditional correspondent banking.

This infrastructure could also enable new forms of programmable money. Smart contracts that automatically escrow funds, release payments based on real-world events, or split revenue between multiple parties. These capabilities exist in crypto today but require technical expertise most businesses don’t have. Stripe’s infrastructure makes them accessible to any developer.

What Comes Next

Stripe’s stablecoin infrastructure is still in early stages, but the pieces are falling into place rapidly. Bridge is already processing millions in monthly volume, Privy’s wallet technology is being integrated across Stripe’s product suite, and Tempo is handling initial transaction loads.

The real test will come when Stripe starts migrating significant portions of their existing payment volume to stablecoin rails. That migration won’t happen overnight—it requires merchant adoption, regulatory clarity, and user education. But when it does happen, it will represent one of the largest shifts in payment infrastructure in decades.

For businesses, the smart move is to start experimenting now. Understanding how stablecoin payments work, what the tax implications are, and how they integrate with existing financial processes will provide competitive advantages as this infrastructure scales.

For developers, Stripe’s API-first approach means stablecoin functionality will be as easy to integrate as traditional card payments. The same few lines of code that power e-commerce today will soon enable instant, global, programmable money movement.

The Bigger Picture

Stripe isn’t just building better payment rails—they’re building the financial infrastructure for the next decade of global commerce. That $1.2 billion investment, plus the development of Tempo, signals they’re all-in on making it happen.

The implications extend far beyond Stripe’s business. If they succeed in making stablecoins as easy to use as credit cards, they’ll accelerate adoption across the entire fintech ecosystem. Other payment processors will be forced to follow suit, traditional banks will need to offer stablecoin services, and businesses worldwide will gain access to programmable, instant, global money movement.

This isn’t just about payments—it’s about unlocking new forms of economic coordination. When money can be programmed, transferred instantly, and settled globally without intermediaries, entirely new business models become possible.

The question isn’t whether stablecoins will become mainstream payment infrastructure. Stripe just answered that with a $1.2 billion exclamation point. The question is how quickly the rest of the industry will catch up—and what new possibilities emerge when money moves at the speed of the internet.

When the dust settles, Stripe’s four-layer money stack might be remembered as the infrastructure that brought digital money to the masses. Not through speculation or hype, but through the same methodical, developer-focused approach that made them the backbone of internet commerce.

That’s how revolutions really happen: not with a bang, but with infrastructure so good that using anything else feels obsolete.

Hey builders,

I’ve been watching fintech companies make the same expensive mistake over and over again—and it’s costing them millions in rebuilds, years in delays, and massive competitive advantages to smarter competitors.

Let me tell you a story that’ll change how you think about fintech architecture...

Why Your Fintech App Needs DeFi-First Architecture (Before It’s Too Late)

Traditional fintech is hitting a wall, and it’s not the one you think.

While legacy banks struggle with decades-old COBOL systems, modern fintech companies are making the exact same architectural mistakes—just with shinier APIs and better UX.

The result? Costly rebuilds, feature limitations, and watching competitors eat your lunch as blockchain technology reshapes the entire financial landscape.

The Million-Dollar Retrofit Trap

Let me paint you a picture with real numbers.

Robinhood—supposedly the poster child of “tech-forward” fintech—took years to enable crypto withdrawals. Years! For what should be a basic feature that users expect from day one.

Why the delay? Their entire architecture was built for traditional securities trading. Adding true crypto functionality meant rebuilding core systems from scratch. We’re talking millions in development costs and countless missed opportunities.

PayPal? Same story. Launched crypto trading in 2020, but users couldn’t actually transfer their crypto until 2022. Venmo followed an identical timeline.

This isn’t coincidence—it’s architectural debt coming due.

The Four Pillars That Separate Winners from Losers

Pillar #1: Multi-Asset Flexibility (Or How to Stop Playing Catch-Up)

Blockchain isn’t just Bitcoin and Ethereum anymore. Your users expect access to hundreds of tokens across dozens of chains. Today.

Your architecture must handle:

- Multiple blockchain protocols simultaneously

- Dynamic asset onboarding without touching code

- Cross-chain interactions and bridging

- Varying transaction finality times (seconds to minutes)

Companies that hard-code asset support spend their entire roadmap just keeping up with user demands. Smart companies build flexible systems that adapt automatically.

Pillar #2: Non-Custodial by Default (The Regulatory Writing on the Wall)

This isn’t about crypto ideology—it’s about survival.

Europeans already have MiCA requiring segregated crypto assets. Americans are watching FTX fallout drive custody regulations that will reshape the entire industry.

DeFi-first architecture means:

- Smart contract wallets instead of traditional key management

- Account abstraction for gasless user experiences

- Social recovery mechanisms (no more lost seed phrases)

- Programmable spending controls built-in

Pillar #3: Direct Protocol Integration (Cut Out the Middlemen)

Most fintech apps use centralized exchanges as their blockchain gateway. This creates single points of failure and kills your margins.

True DeFi-first architecture connects directly to protocols:

- Uniswap for decentralized trading

- Aave for lending and borrowing

- Compound for yield generation

- Layer 2 solutions for scalable transactions

Direct integration unlocks composability—the ability to combine protocols like financial Lego blocks. Your competitors using centralized services can’t match this flexibility.

Pillar #4: Real-Time Settlement (Speed as a Feature)

Blockchain transactions settle in minutes, not days. Your architecture should match this speed:

- Event-driven systems responding to blockchain events

- Instant balance updates

- Real-time transaction status

- Automated settlement processes

Traditional batch processing feels broken to users who expect instant everything.

Why This Creates Unfair Competitive Advantages

Companies building DeFi-first architecture today will completely dominate tomorrow’s financial landscape. Here’s why:

Superior Economics: Direct protocol access eliminates intermediary fees. You offer users better rates while keeping higher margins. Win-win.

Innovation Velocity: New DeFi protocols launch weekly. DeFi-first companies integrate them immediately. Legacy systems need months of development just to evaluate them.

Instant Global Scale: Blockchain is borderless. DeFi-first architecture enables global expansion without traditional banking partnerships, compliance overhead, or geographic restrictions.

Regulatory Resilience: Non-custodial systems face fewer regulatory hurdles. Your users control their assets, dramatically reducing your compliance burden and risk profile.

Your Implementation Roadmap

Building DeFi-first doesn’t mean abandoning traditional finance. It means creating flexible infrastructure that handles both worlds seamlessly.

Start with these principles:

1. Design for multiple blockchains from day one (not as an afterthought)

2. Build modular systems that adapt quickly to new protocols

3. Prioritize direct protocol integration over centralized services

4. Plan for self-custody, even if you start custodial for regulatory reasons

The timeline matters: Companies making this investment now will define fintech’s next decade. Those that don’t will spend it rebuilding from scratch—if they survive the competitive pressure.

The Bottom Line

The choice is brutally simple: build for the future, or spend years rebuilding from the past while your competitors capture your market.

I’ve seen too many talented teams learn this lesson the expensive way. Don’t be one of them.

---

What’s your biggest concern about implementing DeFi-first architecture? Reply and let me know—I read every response and often turn the best questions into future deep-dives.

Chris Dixon’s “Read, Write, Own” gave us a framework everyone loves to cite: the attract-extract cycle. Platforms pour money into growth, cultivate

network effects, then flip the script—extracting value from the same users who made them valuable. Facebook perfected it. Google institutionalised it. Visa and Mastercard turned it into a $500 billion business model.

Blockchain technology promised to break this cycle. Community-owned networks. Aligned incentives. Value flowing to participants, not just shareholders.

But here’s what nobody wants to talk about: Are we actually escaping this trap, or are we building shinier versions of the same extractive systems—just with better marketing and a token?

I’ve spent enough time in this space to recognise when we’re telling ourselves comforting stories. The discomfort that follows is worth sitting with.

The Ethereum Question

Start with what we call the most decentralised smart contract platform: Ethereum.

Post-merge, the narrative is “decentralised enough.” But look at the numbers.

Lido commands roughly 30% of staked ETH. Add Coinbase, Kraken, and Binance, and you’ve got a handful of entities controlling the majority of validation power. This isn’t speculation - it’s on-chain data anyone can verify.

Does this make Ethereum centralised? Not quite. The protocol remains credibly neutral. Anyone can stake. The rules are transparent. No single entity can unilaterally change the system.

But here’s what decentralisation maximalists miss:

First uncomfortable question: If value accrues to large staking entities, have we escaped the attract-extract dynamic - or just rotated who does the extracting?

The L2 Control Problem

If Ethereum’s decentralisation is debatable, L2s make the question unavoidable.

Look at who’s building the dominant rollups:

• Base - Coinbase (a public company with shareholder obligations)

• Arc - Circle (issues the second-largest stablecoin)

• Tempo - Stripe and Paradigm (processes a meaningful chunk of internet commerce)

These aren’t anonymous cypherpunks hacking in basements. These are the same institutions that dominate traditional finance - the ones blockchain was supposed to disintermediate.

I work in this space. I believe many of these teams genuinely want to build better systems. But let’s be brutally honest about the structure:

Who controls the sequencer? Who decides transaction ordering? Who captures MEV?

These questions determine who extracts value. Right now, the answer is: centralised operators with “decentralisation roadmaps.”

“We’ll decentralise... eventually.” Sound familiar? It’s the exact language every platform uses during its attract phase. Twitter was going to be an open protocol. Facebook was about connecting people. Now they’re attention-extraction machines.

Second uncomfortable question: Are L2s “decentralised” because they settle to Ethereum, or are they centralised platforms with decentralised exit options?

The bull case says exit options matter - if Base gets extractive, bridge to Arbitrum. The bear case says network effects and liquidity make migration costly enough that exit rights are theoretical, not practical.

“In Brazil, we watched this play out with PIX. The instant payment system is technically open, but Nubank and a handful of banks captured the network effects. Theoretical openness, practical concentration. The pattern repeats.”

The Missing Layer: Blockchains vs. Payment Networks

Here’s where it gets interesting - and more uncomfortable.

A blockchain is a network with:

• Technical infrastructure (nodes, consensus, execution)

• Economic incentives (tokens, fees, staking rewards)

But a payment network requires layers that blockchains don’t provide:

• Rules governing relationships between parties

• Dispute resolution mechanisms

• Compliance and identity frameworks

• Merchant acceptance and consumer protection

Visa isn’t just infrastructure. It’s a dense web of rules, agreements, and guarantees layered on top of infrastructure.

Circle just launched the Circle Payments Network--explicitly a network on top of blockchain networks, providing the relationship layer that raw blockchain infrastructure lacks.

Fundamental question: If value increasingly moves to these overlay networks, does the underlying blockchain’s decentralisation even matter?

Think it through: You’re on a “decentralised” L2, but your payment flows through Circle’s network, which sets its own rules for participation, fees, and dispute resolution.

The blockchain is decentralised. The payment network on top is not.

The Layering Paradox

The emerging stack looks like this:

[Payment Networks: Circle, etc.] - Rules, relationships, compliance
              |
    [L2s: Base, Arc, Tempo] - Execution, settlement guarantees
              |
         [Ethereum L1] - Security, data availability

Each layer has distinct control characteristics:

• L1: Relatively decentralised (with concentration concerns)

• L2: Mostly centralised today (with decentralisation promises)

• Payment Networks: Corporate-controlled (with blockchain optionality)

The promise of community-owned payment systems assumes alignment across all layers. But what if the layers have conflicting incentives?

“Are we building decentralized rails for centralized applications? And if so, is that actually better than what exists today?”

Maybe it is. Maybe transparent extractive systems beat opaque ones. But let’s not pretend we’ve solved the problem.

The Case for Cautious Optimism

I don’t want to be purely cynical. Real differences from the web2 playbook exist:

Exit rights matter. Even imperfect exit options change game theory. Visa can’t let you migrate transaction history to Mastercard. Blockchain systems make this theoretically possible--and theoretical possibilities constrain behaviour.

Transparency changes incentives. When fee structures live on-chain, you can’t quietly raise take rates. The attract-extract pivot becomes visible, and visibility creates accountability.

Composability creates alternatives. If Circle becomes extractive, competing payment networks can plug into the same rails. Infrastructure is shared even when applications aren’t.

But I’m not convinced these add up to “escaping” the trap. More like “making the trap visible and somewhat escapable.” That’s progress. It’s not victory.

Questions Worth Asking

I started writing this expecting to arrive at a thesis. Instead, I arrived at questions:

1. What’s the minimum viable decentralisation for a payment network? Does every layer need decentralisation, or is one layer sufficient?

2. Who captures value in a layered system? If L1 is decentralised but L2 and payment networks aren’t, who actually benefits from “community ownership”?

3. Is “decentralise later” a legitimate roadmap or a marketing strategy? How do we distinguish genuine plans from perpetual promises?

4. Does blockchain optionality matter if switching costs are prohibitive? Exit rights only work if exit is practical.

5. Are we building for user sovereignty or institutional efficiency? These might not be the same thing--and we should stop pretending they are.

Final Thoughts

Chris Dixon’s framework is useful. The attract-extract cycle is real, and blockchain technology genuinely offers new tools for building networks differently.

But “offers tools” is not “has solved.”

The honest assessment: we’re in an uncertain middle. Some parts of the stack are genuinely more decentralised than their web2 equivalents. Other parts are just as concentrated—or more so.

The promise of community-owned payment systems remains just that. Whether it becomes reality depends on decisions not yet made by entities with their own incentives—ones that may not align with ours.

That’s worth naming. Not to be pessimistic, but to be honest about where we actually are.

The trap is well understood. Whether we escape it remains an open question.

What’s your read? Are L2 decentralisation roadmaps credible, or are we watching the attract phase of a familiar cycle? I’d love to hear your perspective in the comments.

How tokenized stocks will do for investing what stablecoins did for dollars

Stablecoins proved something most people missed: you can decouple access from infrastructure.

A guy in Nigeria can hold USDC backed by US government bonds in a simple wallet. No bank account needed. No brokerage. No correspondent banking relationships. No legacy rails. The token IS the access.

That is what makes it simpler. And that simplicity is about to reshape how the world invests.

The Stablecoin Playbook

Before we talk about where this is going, let us acknowledge what stablecoins have already accomplished.

The numbers are staggering: $240 billion in supply, with $10 trillion transacted in the last 12 months (adjusted for payments, source: visaonchainanalytics.com). This is not speculative crypto trading volume. This is real economic activity: remittances, cross-border B2B payments, savings in stable currency, and increasingly, everyday commerce.

But the numbers only tell part of the story. What stablecoins actually did was solve a distribution problem that traditional finance could not crack for decades.

Think about what it used to take for someone in Brazil, Nigeria, or Indonesia to hold US dollars:

• Open a local bank account with all associated requirements

• Find a bank that offers foreign currency accounts (most do not)

• Navigate complex regulations around foreign currency holdings

• Pay significant fees for currency conversion and maintenance

• Accept limited access and withdrawal restrictions

Or they could download a wallet app and buy USDC. Same underlying asset exposure. Fundamentally different access model.

This is not about technology replacing banks. It is about distribution channels evolving to match how people actually want to interact with financial products.

The Equities Opportunity

Now apply that same logic to equities.

Imagine a Brazilian using Pix to buy a stablecoin, swapping to USDC, then purchasing a token representing the S&P 500. From an asset perspective, there is little difference between this token and a traditional index fund. The underlying exposure is identical.

But from an access and distribution perspective? This changes everything.

Consider what happens when someone in Mexico wants exposure to US equities today:

1. Open a local brokerage account (KYC, documents, waiting periods)

2. Find a broker that offers international investing (limited options)

3. Navigate currency conversion through their bank

4. Pay custody fees, FX spreads, and transaction costs

5. Accept restricted trading hours and settlement delays

6. Deal with tax reporting complexity across jurisdictions

Each of these steps adds friction, cost, and exclusion. Not technical exclusion, but practical exclusion. The kind where the product technically exists but is so cumbersome that most people never bother.

Now imagine the alternative: Pix to stablecoin to tokenized S&P 500. Three steps. One interface. Available 24/7. Settlement in seconds.

Why Intermediaries Collapse

Here is what actually happens when assets move onchain:

The number of intermediaries collapses. No local custodian, no foreign broker, no correspondent banks, no transfer agents, no clearing houses in the traditional sense. The blockchain handles settlement. The smart contract handles custody rules. The token handles transferability.

The product becomes simpler and cheaper. Fewer middlemen means lower fees and less complexity. When you remove four intermediaries from a transaction chain, you do not just save their fees. You eliminate the operational complexity, the reconciliation errors, the settlement delays, and the regulatory overhead that each one introduces.

Demand increases exponentially. When you make something 10x easier to access, you do not get 10x more users. You get entirely new categories of users who were priced out before.

This is the part most traditional finance people miss. They look at the current market for international equities and think about taking share from existing competitors. But the real opportunity is in creating markets where none existed because the friction was too high.

> How many people in emerging markets want exposure to US equities but have never seriously considered it because the process is too complicated? That is the addressable market.

Derivatives Make Even More Sense

If tokenized equities seem like a natural evolution, derivatives and perpetuals are an even more obvious fit for blockchain rails.

Think about what a derivative actually is: synthetic exposure to price movements without owning the underlying asset. There is no physical settlement. No custody of actual shares. No dividend processing. No corporate actions to manage.

It is just math. A contract that pays out based on the difference between entry and exit prices.

And math runs better on programmable rails than on legacy infrastructure.

Traditional derivatives markets work through a complex web of prime brokers, clearinghouses, margin accounts, and settlement systems. All of this exists because trust between counterparties needed to be intermediated. But smart contracts can do this natively. Collateral can be locked programmatically. Settlement can happen automatically. Margin calls can execute without human intervention.

For emerging market users who want exposure to US asset price movements, onchain derivatives offer something that traditional markets simply cannot match: direct access without the infrastructure burden.

The Flywheel Effect

This creates a self-reinforcing cycle that traditional finance cannot easily replicate:

More access leads to more liquidity. When you make it easier for someone in Colombia to buy tokenized US equities, you are adding to the global liquidity pool. This is not regional liquidity. It is unified, 24/7, global liquidity.

More liquidity attracts more asset issuers. As tokenized asset markets grow, more traditional asset managers will see the distribution opportunity. Why limit yourself to investors who can navigate traditional brokerage infrastructure when you could reach anyone with a wallet?

More products drive more access. As the range of available tokenized assets expands, the value proposition for new users increases. First they came for dollar exposure. Then equities. Then bonds. Then alternatives. Each new product category brings new users into the ecosystem.

More users create more infrastructure. As user numbers grow, more developers build better interfaces, better on-ramps, better compliance solutions. This makes the next wave of adoption even easier.

This flywheel is already spinning in stablecoins. The infrastructure built to support USDC and USDT adoption is now the foundation for tokenized equities. The same wallets, the same exchanges, the same compliance frameworks.

The Emerging Market Perspective

I keep coming back to Brazil, Nigeria, Mexico, Indonesia. This is not random. These markets represent where the impact will be most dramatic.

In developed markets, the existing infrastructure mostly works. A retail investor in the US or Europe can access diversified equity exposure through low-cost ETFs with reasonable friction. The gains from tokenization are incremental: maybe faster settlement, maybe 24/7 trading, maybe some cost reduction.

But in emerging markets, the existing infrastructure often does not work. Or it works so poorly that most people are effectively excluded.

This is not about financial literacy or investor sophistication. Plenty of people in emerging markets understand the value of diversified equity exposure. They have phones. They have internet. They have money to invest. What they lack is infrastructure that makes investing practical.

Tokenized assets solve this at the infrastructure level. You do not need to build brokerage networks in every country. You do not need correspondent banking relationships in every jurisdiction. You do not need local custody solutions for every market.

You need a wallet and an internet connection.

The Scale of Opportunity

Let me put some context around the opportunity size.

Stablecoins achieved $240 billion in supply by solving the dollar access problem. This is meaningful, but dollars are just one asset class. A useful one, but limited in what they can do for long-term wealth building.

Equities are where wealth is actually created over time. The historical equity risk premium, the power of compound growth, the access to corporate earnings. This is how people in developed markets have built generational wealth.

Global equity markets represent roughly $100 trillion in market capitalization. Even a small percentage of this moving to tokenized rails represents an opportunity that dwarfs stablecoins.

But the real opportunity is not in tokenizing existing investments. It is in the new investments that become possible when access friction disappears.

How many people globally would invest in diversified equity portfolios if they could do it as easily as sending a text message? That is not a rhetorical question. That is the addressable market for tokenized equities.

What This Means

The future of financial products will be built on high-quality American assets tokenized and distributed globally.

This is not a prediction about technology. It is a prediction about distribution economics. The infrastructure that stablecoins built, the regulatory frameworks emerging around digital assets, the wallet penetration in emerging markets, all of this is creating the conditions for tokenized equities to follow the same trajectory.

The builders who understand this today will define how the next billion people invest.

I do not say this as an observer. This is exactly why I am building Awa.

We are creating a platform to enable builders in emerging economies to create savings, investments, and borrowing products on the shoulders of high-quality US assets once they are finally onchain.

The infrastructure is finally ready. The regulatory clarity is emerging. The user demand is already there, waiting for products that work.

Now it is time to build.

Final Thoughts

Stablecoins proved that the model works. You can take complex financial infrastructure and replace it with a token that just works. The $240 billion in supply and $10 trillion in transaction volume is the proof of concept.

Equities onchain is the next chapter. Same playbook. Bigger prize. More transformative impact.

The question is not whether this will happen. The question is who will build the products that bring it to market.

What do you think? Is the equities tokenization thesis overstated, or are we actually underestimating how fast this will move? I would love to hear perspectives from both sides.

If you found this valuable, consider sharing it with someone who is thinking about the future of asset distribution. And subscribe if you haven’t yet.

This reading list represents what amounts to a university course. I’ve spent two years following Nic Carter’s monetary analysis and infrastructure deep dives (including his Murmurations blog), Simon Taylor’s fintech lens, and the a16z crypto team’s systematic breakdowns. These writers don’t hype; they explain. They show you the rulebook, the incentives, and the failure modes. That’s what makes this collection worth your time.

How to use this list

• If you’re starting from scratch: Begin with Section 1 (Foundations & Primers) and move sequentially. The primers establish definitions and mental models you’ll need later.

• If you’re topic-focused: Jump directly to the section that matters—cross-border payments, regulatory frameworks, technical architecture.

• If you’re already conversant: Skip to Section 5 (Critical Analysis) for the sharpest debates and forward-looking questions.

Last updated: November 18, 2025

Make sure you subscribe to receive updated versions of this list in the future:

Subscribe now

1. Foundations & Primers

Core understanding of what stablecoins are and why they matter

Start here. These pieces establish the taxonomy, mechanisms, and historical context.

• Cryptodollars – Nic Carter (Castle Island Ventures, 2020)

  The foundational paper. Carter introduces “cryptodollars” as a concept and lays out the monetary and technical vision. Must-read.

• Stablecoins in 1,000 words – Matt Brown

  Perfect entry point. Clear, compact explanation of stablecoin types and basic mechanics.

• How to Think About Stablecoins – Alex Johnson

  Provides a mental framework for analysing value and use cases without getting lost in technical weeds.

• Understanding Stablecoins Banking History – a16z crypto

  Essential context. Shows how stablecoins relate to historical banking evolution—private money issuance, reserve backing, and trust infrastructure.

• Stablecoin Guide: What, Why, How – a16z crypto

  Comprehensive fundamentals guide covering the basics with the a16z team’s characteristic precision.

• A Stablecoin Primer – Jas Shah

  Another strong introduction to concepts, mechanisms, and ecosystem dynamics.

2. Infrastructure & Technical Architecture

How stablecoins work technically and their position in the payments stack

Once you understand what stablecoins are, you need to understand the machinery: blockchain rails, payment flows, trust infrastructure, and interoperability.

• Situating Stablecoins in the Payments Landscape – Nic Carter

  Carter’s infrastructure analysis is some of the best in the space. This piece maps where stablecoins sit among traditional payment methods—schemes, rails, and settlement layers.

• Yes We Really Do Want to Trust Crypto – Nic Carter

  The trust infrastructure underpinning stablecoin systems. Carter’s writing consistently cuts through the “trustless” rhetoric to show where institutional and technical trust actually matters.

• BVNK Stablecoin Infrastructure – Jas Shah

  Examines modern stablecoin infrastructure providers and their role in the ecosystem.

• Blockchains, Banks, Asset Managers, Fintechs – a16z crypto

  How traditional financial institutions are integrating with blockchain infrastructure. Shows the convergence happening beneath the surface.

• On-Chain Foreign Exchange and Cross-Border Payments – Circle

  Technical deep dive on implementing FX and payments on-chain. Gets into settlement mechanics and liquidity.

• Squaring the Circle – Luca Prosperi

  Technical analysis of stablecoin mechanics and market dynamics from another sharp voice in the space.

• The 10 Minute Guide to Tether – Luca Prosperi

  Comprehensive overview of USDT’s technical architecture and business model. Love it or hate it, Tether dominates—understand why.

• The Crypto Banking System

  Deep dive into how stablecoins function within the broader crypto financial system.

3. Real-World Applications & Market Dynamics

Use cases, adoption patterns, and market structure

This is where stablecoins leave the whiteboard and enter the real economy. Cross-border flows, emerging market dollarisation, corporate adoption, and market evolution.

Cross-Border & Emerging Markets

• Cross-Border Payments in ~1,000 words – Matt Brown

  Overview of how cross-border payments work today and where stablecoins fit into that machinery.

• It’s Time Local Went Global – Activant Capital

  The case for stablecoins in international commerce. Well-argued and grounded in payment system realities.

• Getting to the Crux of Stablecoin Utility – Josh from Parafi

  Analysis of how stablecoins power the DeFi ecosystem—collateral, liquidity, composability.

• Stablecoins: The Emerging Market Story – Nic Carter (at Castle Island)

  Deep analysis of adoption in developing economies.

• Making Sense of Tether on Tron – Liam Horne

  Why USDT on Tron became the dominant stablecoin for global payments, and how Ethereum can compete. Excellent on-chain data analysis.

• Dollarisation 2.0: How Stablecoins Unlock Emerging Markets – VIDEO – Nic Carter & Martin Carrica

  Deep exploration of stablecoin adoption in countries with unstable currencies. Carter’s monetary analysis shines here.

Industry Adoption & Market Analysis

• State of Crypto Report 2024: New Data on Stablecoins – a16z crypto

  Comprehensive market data showing stablecoin growth hitting all-time highs. The a16z team brings real numbers and trend analysis.

• Stripe Acquires Bridge: Stablecoins, Platform of Platforms – Ben Thompson

  Deep analysis of Stripe’s $1.1B acquisition signalling institutional adoption. Thompson’s strategic lens is always worth reading.

• Why Stripe Spent $1.1B on Bridge – Simon Taylor

  Simon’s financial infrastructure perspective on Stripe’s strategic stablecoin play. He’s one of the sharpest observers of fintech-crypto convergence.

• Stablecoins 2030 – Citi GPS Report

  Institutional research on the future trajectory of stablecoin markets. Big bank perspective with actual numbers.

• Crypto Neobanks – Bankless

  How crypto-native banking services are being built on stablecoin rails.

• PayPal PYUSD: Does It Really Matter? – Luca Prosperi

  Analysis of corporate stablecoin initiatives and their market impact. Healthy scepticism applied.

Market Structure & Evolution

• The Great Onchain Migration – Pantera Capital

  Vision for how stablecoins enable the shift to onchain finance. Forward-looking but grounded.

• The Stablecoin Duopoly is Ending – Nic Carter

  Analysis of changing market dynamics and new entrants. Carter tracking competitive shifts as they happen.

• Stablecoins Are Better – Simon Taylor

  Why stablecoins represent a superior payment infrastructure. Simon making the case clearly.

• Visa Crypto Report – Visa UK

  Payment network perspective on stablecoin integration. Useful to see how the incumbents are thinking.

4. Regulatory & Policy Landscape

Global regulatory frameworks and policy perspectives

Stablecoins operate at the intersection of technology, finance, and regulation. Understanding the policy landscape is non-optional.

• Statement on Stablecoins – SEC Division of Corporation Finance

  SEC’s official view on certain stablecoins not being securities. Critical for U.S. regulatory clarity.

• The CLARITY Act – Why It Matters, What to Know, and What to Do – a16z crypto

  Breaking down proposed U.S. stablecoin legislation. The a16z policy team does excellent work translating regulatory proposals.

• FCA – Regulating cryptoassets Phase 1: Stablecoins – UK Financial Conduct Authority

  UK’s approach to stablecoin regulation. Dense but important for understanding the British framework.

• The EU’s MiCA Framework: The One Regulation to Rule Them All – Patrick Hansen

  Comprehensive breakdown of Europe’s Markets in Crypto-Assets regulation—its genesis, goals, rules, and global implications.

• American Leadership in the Digital Financial Technology – White House

  Policy perspective on U.S. strategic advantages in stablecoins. Shows how policymakers are thinking about competition and innovation.

5. Critical Analysis & Future Perspectives

Challenges, criticisms, and forward-looking views

The best learning happens when you engage with criticism and uncertainty. These pieces push back, raise hard questions, and explore what might come next.

• Death To Stablecoins! – Luca Prosperi

  Critical take on stablecoin risks and limitations. Worth reading to stress-test your assumptions.

• The Last Word on Stablecoins and Free Banking – Nic Carter

  Historical perspective on private money issuance and its implications. Carter bringing banking history to bear on crypto debates.

• Nic Carter’s Bull Case For Stablecoins – VIDEO – Nic Carter

  Comprehensive argument for stablecoin adoption. Carter at his best—monetary theory, practical adoption, geopolitical implications.

• Will the real stablecoin please stand up – BIS

  Comparison of different stablecoin models and their trade-offs. Central bank perspective with technical depth.

• A Very Stable Future – Real-world Applications for Stablecoins – Ayo Omojola

  Comprehensive exploration of stablecoin use cases and future potential. Broad and forward-looking.

6. Essential Resources & Ongoing Coverage

Stay connected with ongoing stablecoin developments

The space moves quickly. These resources keep you current.

Podcasts

• Tokenized – Simon Taylor & Cuy Sheffield

  Regular deep dives on stablecoin developments and digital asset infrastructure. Simon Taylor co-hosts, which tells you the quality level.

Essential Blogs & Newsletters

• Murmurations , Nic Carter’s Old Articles – Nic Carter

  Carter’s infrastructure-focused blog covering DeFi, stablecoins, and crypto financial systems. Deep technical analysis with precision and clarity.

• Fintech Brainfood – Simon Taylor

  Weekly fintech insights with regular stablecoin and digital asset coverage. Simon’s one of the best at connecting traditional finance and crypto infrastructure.

• Dirt Roads – Luca Prosperi

  Critical analysis of DeFi, stablecoins, and crypto financial infrastructure. Luca’s work has been essential to my own understanding—he writes with precision and doesn’t shy away from complexity.

A note on curation

This is a living document. The pieces above earned their place through rigour, clarity, and willingness to engage with both potential and risk. If you’ve found an exceptional piece on stablecoins that should be included, let me know.

The writers I’ve highlighted—Luca, Nic, Simon, and the a16z crypto team—set the standard. They explain machinery, cite sources, map incentives, and acknowledge uncertainty. That’s the bar.

Start reading.

I spent three years losing thoughts across seventeen applications. Meeting notes lived in Notion. Article highlights scattered across Pocket, Instapaper, and browser bookmarks. Personal reflections fragmented between Apple Notes and Google Docs. Every time I needed to synthesise an idea, I’d open six tabs, copy-paste excerpts, and lose context switching between proprietary formats.

The breaking point came during a product meeting. Someone asked about a competitor analysis I’d written. I knew I’d captured detailed notes—I could picture the document. But I couldn’t remember which app. Twenty minutes of searching later, I found fragments across three services. None connected. None searchable together. The system had failed.

That failure forced a question: What if personal knowledge management borrowed from software infrastructure design?

Subscribe now

The Core Architecture

Modern software development solved the fragmentation problem decades ago. Version control tracks every change. Plain text files ensure permanence. Search spans entire codebases. The question was whether these principles could apply to personal knowledge.

The architecture I built has four layers:

Layer 0: Storage – Plain markdown files in a local folder structure. No proprietary formats. No vendor lock-in. Complete ownership.

Layer 1: Version Control – Git tracks every change with full history. Every note, every edit, permanently preserved. Branch experiments, revert mistakes, collaborate when needed.

Layer 2: Interfaces – Cursor (an AI-powered IDE) for primary interaction. Obsidian for visual browsing and graph views. Both read the same files; neither owns them.

Layer 3: Intelligence – Claude, an AI assistant configured with vault-aware context. Understands the entire knowledge graph. Helps synthesise across thousands of notes.

This isn’t a metaphor. I literally use an IDE as my note-taking interface. The system treats knowledge like code: versioned, searchable, refactorable at scale.

The Folder Taxonomy

The structure follows active project domains:

📂 Vault/
├── Areas/              # Active focus domains
│   ├── Work/           # Professional projects
│   ├── Personal/       # Life management
│   └── Writing/        # Content production pipeline
├── Readwise/           # Auto-synced reading highlights
├── Daily Notes/        # Daily capture
├── Templates/          # Reusable structures
└── Archive/            # Completed projects

The Writing subdirectory implements a five-stage pipeline:

1. Sketches – Capture every spark without judgement (no promotion criteria)

2. Long Ideas – Develop promising concepts (3+ related sketches → promotion)

3. Research – Accumulate supporting knowledge (outlined argument → promotion)

4. Drafts – Active article writing (complete first draft → promotion)

5. Ready to Publish – Production-ready content (final edit complete → publish)

Each stage has explicit promotion criteria. Nothing gets lost. Promising ideas surface naturally through structured review.

How Information Moves Through the System

Reading Integration Flow

Every morning, highlights from Kindle, PDFs, and web articles appear in my vault:

1. Highlight in any reading tool

2. Readwise syncs overnight to /Readwise/ folder

3. Notes appear as markdown with full source context and metadata

4. Claude synthesises insights across sources on request

No manual copying. Zero clipboard operations. Everything searchable and linkable.

Meeting Knowledge Flow

After every meeting:

1. Granola AI transcribes and summarises

2. Export to markdown with key decisions and action items

3. Import to vault using meeting template (includes attendees, decisions, links to projects)

4. Link to relevant project notes and people pages

The template enforces structure:

# [Meeting Title] - [Date]

## Attendees
- [Name] - [Role]

## Key Decisions
- Decision 1
- Decision 2

## Action Items
- [ ] Task → [[Linked Project]]

## Context Links
- Previous meeting: [[Link]]
- Related project: [[Link]]

Daily Thinking Flow

Each morning:

1. Create daily note from template (includes prompts: “What’s on my mind?”, “What did I capture yesterday?”, “What needs attention?”)

2. Review yesterday’s captures

3. Promote sketches through writing pipeline

4. Ask Claude to identify emerging patterns across the week

The daily template includes git commit automation. Every session ends with changes versioned automatically.

Writing Production Flow

When producing content:

1. Review Writing Pipeline Dashboard (shows counts per stage, recently updated items)

2. Select a draft from stage 4

3. Open in Cursor with Claude context

4. Ask Claude to synthesise research from vault

5. Write with AI assistance for structure and synthesis

6. Move to stage 5 when complete

Concrete Workflow: This Article’s Production

Let me walk through how this article came to exist:

Day 1 – Captured sketch: “Using Cursor for non-coding tasks”. Just a title. Saved to /Areas/Writing/1-Sketches/cursor-non-coding.md.

Day 3 – Promoted to Long Ideas after noticing three related sketches about AI tools and knowledge work. Created outline structure. Used Claude in Cursor:

Me: “Review my sketches about AI tools and knowledge work. What patterns emerge?”

Claude: “I found four sketches:
- ‘Cursor for non-coding tasks’ (you)
- ‘Context is all you need’ (you, 3 days ago)
- ‘AI-powered writing pipeline’ (you, last week)
- Highlights from ‘Building a Second Brain’ (Readwise)

Pattern: You’re exploring how developer tools apply to knowledge work. The common thread is treating knowledge like code—versioned, searchable, refactorable.”

Day 5 – Research phase. Claude helped analyse my vault structure, identify key workflows, and gather examples from actual usage. Created research document at /Areas/Writing/3-Research/pkm-system-research.md with 2,400 words of synthesised vault insights.

Day 7 – Moved to Drafts. Writing session in Cursor:

Me: “Help me structure an article about this PKM system. Audience: technical people who’ve felt knowledge fragmentation pain.”

Claude: “Structure suggestion:
1. Name the fragmentation problem (relatable friction)
2. Describe the architecture (layers, like a tech stack)
3. Show information flows (like data pipelines)
4. Provide concrete workflow example (numbered steps with specifics)
5. Address constraints and failure modes
6. Close with implementation path”

Wrote 3,200 words in one session with Claude helping synthesise examples from the research doc.

Day 8 – Final polish. This version.

The entire production happened inside the vault. Every stage versioned. Every decision preserved. Claude had full context because the context is the vault.

The Cursor Rules That Make It Work

The .cursorrules file at vault root contains instructions for Cursor’s AI:

# Git Automation
After each editing session:
1. Stage all changed files in /Areas/
2. Generate commit message describing changes
3. Commit automatically

# Vault Structure
Enforce:
- All notes include YAML frontmatter (tags, created, updated dates)
- Meeting notes use meeting template
- Writing pipeline files stay in correct stage folders
- Links use wiki-style [[brackets]]

# Privacy
Never commit:
- /Personal/Private/
- /Work/Confidential/
- Files marked with #private tag

This automation means I write, and the system handles versioning. No manual git commands. No deciding what to commit. The IDE does it.

The AI Configuration That Matters

Claude in Cursor has two operating modes, configured via prompts:

Thinking Mode – For problem-solving and analysis:

• Ask clarifying questions first

• Suggest multiple perspectives

• Map arguments to vault notes

• Identify gaps in reasoning

Writing Mode – For content creation:

• Understand the five-stage writing pipeline

• Synthesise research from vault

• Suggest structure before drafting

• Link to relevant notes as sources

The critical difference from generic AI chat: Claude has read every note in my vault. When I ask “what have I written about X?”, it searches 4,000+ markdown files and returns actual content, not hallucinations.

Where the System Shows Constraints

Learning Curve – Setting up Cursor with git integration and custom rules requires technical comfort. The first week felt slower than Notion. The payoff comes at scale.

Search Precision – Markdown files don’t have database-style metadata. Tag discipline matters. Inconsistent tagging degrades search quality.

Sync Complexity – Running this across multiple devices requires git discipline. Merge conflicts happen when editing the same file on laptop and phone. I’ve solved this by treating mobile as read-only (Obsidian) and editing only on desktop (Cursor).

Privacy Trade-offs – AI vault analysis means sending note content to Claude’s API. Sensitive material goes in .gitignored folders that Claude never sees. This works but requires constant vigilance.

No Real-time Collaboration – Unlike Notion, multiple people can’t edit simultaneously. Git merge workflows help, but this system optimises for individual deep work, not team real-time collaboration.

The Questions This Opens

If treating knowledge like code works at individual scale, what breaks at team scale?

Could a team vault with branch-per-person and merge-on-review work for shared knowledge bases?

What if AI didn’t just search existing notes, but synthesised new insights overnight—running reasoning chains while you sleep, proposing connections you hadn’t considered?

If plain markdown plus AI can match proprietary tools today, what happens when the AI gets better and the file format stays stable?

Leave a comment

Building Your Own Implementation

If you want to try this approach:

1. Foundation setup (Week 1)

   • Install Cursor (free tier works)

   • Install Obsidian (also free)

   • Create folder structure (start with /Daily Notes/ and /Areas/)

   • Initialise git repository at root

2. Configure AI context (Week 2)

   • Add Claude to Cursor (requires API key)

   • Create .cursorrules file with basic automation

   • Define operating modes for your needs

   • Test vault-wide search

3. Connect input sources (Week 3)

   • Set up Readwise for reading highlights (optional)

   • Create daily note template

   • Build meeting note workflow

   • Establish capture habits

4. Build your workflows (Ongoing)

   • Start with daily notes only

   • Add one workflow per week

   • Let the system grow organically

   • Watch for patterns that want automation

The key: start small. One folder. One template. One habit. The system’s value compounds with size and time.

What This Makes Possible

After an initial effor to migrate and converge all apple and notion notes to .md files, I have 4,000+ notes. Every book highlight. Every meeting. Every article idea. Every personal reflection. All in plain text files I own completely.

When I write, Claude can surface insights from notes I wrote years ago. Connections I’d forgotten. Patterns across domains I didn’t consciously notice.

The system gets smarter as it grows. Unlike Notion or Evernote, which slow down with scale, this architecture thrives on volume. More notes mean more context. More context means better AI synthesis. Better synthesis means better thinking.

The tools are ready. Plain markdown has been stable for 30 years. Git has been production-grade for 15. AI assistants became capable in the last two. The pieces exist. The question is whether the approach fits your knowledge work style.

For me, escaping the app prison meant adopting developer infrastructure for non-developer work. Your path might differ. But if you’ve felt the friction of fragmented tools and lost context, the architecture is worth considering.

This article lives in my vault at /Areas/Writing/4-Drafts/ and was produced entirely using the system it describes. The git log shows 47 commits over eight days of development.

Cursor proved this for code. AI browsers are trying to prove it for the web. But most are missing the critical piece that makes context management actually work, and ironically, Arc had already solved it before pivoting to Dia.

Subscribe now

How Cursor structures context

Look at Cursor’s interface. Three zones, each serving a distinct purpose:

Left: knowledge structure. Folders, files, Git branches. This isn’t decoration—it’s your spatial map. You see where everything lives. You navigate by location, not search. The structure persists across sessions, so you build muscle memory for where things are.

Centre: working artefact. The file, document, or terminal you’re actively editing. One thing, focused, unobstructed. Everything else fades. The centre is sacred—it belongs to your work, not the interface.

Right: AI interaction. Chat, agents, commands. This is where augmentation happens. Critically, the AI sees your context—which files are open, what you’ve selected, and where you are in the project structure. You don’t reconstruct context in conversation. It’s already there.

This structure works because it maps to how people actually manage knowledge work: you need to see what’s available (left), focus on one piece (centre), and occasionally invoke help that understands your working state (right).

Why it works beyond code

I use Cursor for two completely different types of work. Code, obviously—that’s what it’s built for. But also business knowledge: strategy memos, ideation documents, reports, and meeting notes. All written in Markdown, stored in a GitHub repo, managed like code even though it’s just text.

The structure holds for both. For code, the left panel shows my project files. For business docs, it shows my knowledge base structure—product specs in one folder, strategy docs in another, research notes nested by topic. The same spatial logic applies.

Version control is the unexpected benefit. Every change creates a commit. I can diff yesterday’s strategy memo against today’s. I can see when I added a section, who changed what, and why something was revised. That’s impossible in Notion or Google Docs without manual tracking or paying for history—and even then, it’s not structured.

The three-panel grammar works because context is always visible. When I’m editing a strategy doc and need to reference a product spec, I don’t search or switch apps. I see both files in the tree, open the second one, and work with both simultaneously. The AI chat sees both. I can ask it to cross-reference them, summarise differences, or check consistency—all without explaining what I’m looking at.

The browser: king of context

But here’s the thing: browsers are the actual interface layer where most knowledge work happens. You’re not just editing documents. You’re researching, reading, comparing sources, referencing multiple pages, and now—increasingly—asking AI to help you process all that information.

Browsers are the king of context. Or they should be.

AI browsers—Arc, Browserbase, SigmaOS—recognised this. They saw the Cursor pattern and tried to replicate it. Web page in the centre, tabs or history on the side, AI chat on the right, where you can reference pages and ask questions.

The structure makes sense. But they’re missing something critical: context management.

What Arc got right (and then abandoned)

Here’s the irony: Arc had solved context management for browsers before the AI era made it even more important.

Spaces. Profiles. Nested tab folders. The way you could organise tabs vertically, group them by project or context, pin persistent resources, and keep ephemeral research tabs separate—it was the missing piece. It was IDE-level context management for the web.

Most people don’t work with five clean tabs. They work with dozens. Some tabs are fixed—tools you access constantly, references you keep open, dashboards you check daily. Some are volatile—research tabs for a current project, pages you’ll close in an hour, temporary context that comes and goes.

Browsers traditionally treated all tabs the same: a horizontal strip at the top, chronological, flat. That doesn’t scale. You lose spatial memory. You can’t see the structure. You end up with 47 tabs and no idea which ones matter.

Arc’s vertical, nested, space-separated structure solved this. You could see your context at a glance. Work tabs in one space, personal in another, research for Project X in a folder, persistent tools pinned at the top. It wasn’t just organisation—it was spatial memory for the web.

The Dia pivot and what’s missing

I understand why they pivoted to Dia. Simplification for mass adoption. Solving deep infrastructure limitations from how they built the first browser. Those are real constraints.

But in chasing simplification, they’re abandoning the feature that mattered most—especially now, in the AI era.

When you’re working with AI assistance in a browser, context management becomes critical. You’re not just browsing one page. You’re referencing multiple sources, comparing articles, indexing research, and asking the AI to synthesise information across tabs. If your tabs are a flat, unsorted mess, the AI can’t help you. There’s no structure to reference.

The AI needs to see what you see: which tabs are part of your current project, which are persistent resources, which are temporary research, and how they relate to each other. That requires structure. It requires the nested, spatial, space-separated approach Arc had built.

What AI browsers need to learn from IDEs

IDE users figured this out decades ago. You can’t code with a flat file list. You need folders, nested organisation, and a persistent structure. You need to see where things are and how they relate.

AI browsers need the same. Honestly, every middle OS layer needs to learn that lesson. Tabs as a flat strip worked when browsing was linear. It doesn’t work when you’re managing complex research, cross-referencing sources, and asking AI to process information across multiple contexts.

The solution isn’t removing structure. It’s embracing it.

What Dia (or any AI browser) should incorporate:

• Vertical tab organisation. See your tabs as a structured tree, not a horizontal overflow.

• Nested folders. Group related tabs by project, topic, or context. Make the grouping visible.

• Spaces or profiles. Separate work contexts completely—work vs. personal, Project A vs. Project B.

• Persistent vs. ephemeral distinction. Pin the tabs that stay, let the rest flow. Make the difference visually clear.

• AI that sees the structure. When I ask the AI to “compare these three research articles,” it should know which ones I mean because it sees my tab organisation.

Arc built this. It worked. The fact that Dia is moving away from it—just as AI makes it more valuable—is the real missed opportunity.

Context is the interface

The pattern is consistent: when AI tools work, it’s because they make context persistent and visible.

Cursor works because you can see your file structure, your working file, and your AI interaction simultaneously. The AI sees what you see.

AI browsers should work the same way. Web pages in the centre, structured tab organisation on the left, and AI chat on the right that understands your browsing context.

But that only works if the context is actually structured. Flat tabs don’t give the AI (or you) enough information. You need spatial organisation, nested groups, persistent vs. temporary distinction, and space separation.

Arc proved this worked. Dia and other AI browsers should be doubling down on it, not simplifying it away.

Because in the end, all you need is context. But you need it structured, visible, and persistent. That’s what makes AI assistance actually useful—and what makes the interface disappear so you can focus on the work.

Leave a comment

Brazil processes more instant payments in a month than the US does in a year through a system that’s existed for 50 years. Pix hit 6.4 billion transactions in December 2024 alone—80% more than credit and debit cards combined. Over 160 million Brazilians send money instantly, for free, 24/7.

Yet while Brazil celebrates Pix and tests Drex, the blockchain economy is being built without us.

The US just created the rules of the game. The GENIUS Act established federal stablecoin regulation in July 2025. The CLARITY Act passed the House 294-134, defining which crypto assets fall under SEC versus CFTC oversight. Together, these laws create the regulatory clarity that unlocks institutional capital, cross-border settlement, and tokenised asset markets.

Brazil’s response? Double down on a CBDC we don’t need.

Drex solves problems Pix already fixed. Instant transfers? Pix does that. Financial inclusion? 76.4% of Brazilians already use Pix—far beyond any CBDC adoption globally. Low-cost infrastructure? The Central Bank built SPI as open, interoperable rails that work.

The real gap isn’t payment speed. It’s blockchain participation.

While US banks prepare to issue regulated stablecoins and tokenise treasury bonds, Brazilian institutions wait for Drex pilots to conclude. While American exchanges operate under clear market structure rules, Brazilian crypto companies face regulatory ambiguity. While the US positions dollar-backed stablecoins as digital infrastructure for global trade, Brazil builds a walled garden.

Abandon Drex. The resources poured into a centralised digital currency won’t help Brazilian institutions compete in an open, global tokenised economy.

The blockchain economy emerging from GENIUS and CLARITY isn’t about replacing national payment systems. It’s about tokenised securities, cross-border settlement layers, DeFi protocols, and programmable assets that operate across jurisdictions. Brazil needs regulatory frameworks that let our financial institutions compete in these markets, not just domestically but globally.

The path forward is clear: regulate direct access to on-chain assets. Create categories based on asset class, jurisdiction of issuance, and issuer. For assets issued abroad, place regulatory weight on distributors operating in Brazil—but make it easy for them to work with foreign assets. Fighting this trend only accelerates DeFi adoption as Brazilians route around restrictions to access global markets anyway.

Pix proved Brazil can leapfrog incumbents when we focus on the right problems. We skipped the card infrastructure buildout that took decades in developed markets and went straight to instant, free, open payments. That same leapfrog opportunity exists in blockchain—but only if we stop pouring resources into replicating what we already have.

The question isn’t whether Brazil should have a CBDC. It’s whether centralised infrastructure can compete with an open tokenised economy. Once American securities, assets, and funds live on public blockchains under clear regulations, any Brazilian with blockchain access can participate directly. Drex offers a walled garden. The US is building open rails that don’t require permission from central banks.

Brazil can either create frameworks to compete in that economy or watch capital flow to jurisdictions where the rules are already clear.

However, there is something even more important that most people in product neglect. The power of a well-designed feedback loop process. That’s when the inflow of customer support ticket information is digested in a way that creates constant high-quality inputs for the product team.

In every place I have worked, I have seen support teams flooding engineers with Slack messages, DMs, telegram, etc. More often than not, people think that reaching out to the developer directly they are helping the customer more than ever. The reality is you are not. You are not even helping yourself.

Product managers and customer support managers need to understand that the most valuable part of their work lies in the intersection of their teams. Sit down together and structure your information flow. I have a particular view on how to do it. By categorising customer inquiries into Questions, Requests, and Problems, you can create a more efficient feedback loop and enhance overall customer satisfaction. Let’s take a look.

1. Questions

Customers often reach out because they:

• Can’t find specific information about a feature.

• Don’t understand how a feature works.

• Disagree with something due to a lack of information - this will often be mischaracterised as a problem.

These scenarios usually stem from gaps in user experience or insufficient information available to the customer.

There are two main ways of dealing with question ticket types:

1. Improve UX/UI Design

   • Evaluate and Redesign Confusing Elements: Identify parts of your product that users find confusing and redesign them for clarity.

   • User Testing: Conduct regular user testing sessions to gather feedback on usability.

   • Simplify Navigation: Ensure that your product’s navigation is intuitive and user-friendly.

2. Enhance Information Availability

   • Customer Support Training: Equip your CS team with thorough knowledge and resources to assist customers effectively.

   • Comprehensive Help Center: Develop a robust help centre with FAQs, tutorials, and guides.

   • AI Bots and Virtual Assistants: Implement chatbots to answer common queries instantly.

2. Requests

These are instances where customers:

• Ask for new features or services.

• Request actions they can’t perform themselves (e.g., editing a shipping address).

Tips: Balance a quicker fix via CS flow vs a self-service UI flow on your application. The best solution depends on the frequency of occurrence.

1. Develop self-service options

   • Empower Customers: Add functionalities that allow customers to perform tasks independently.

   • User Interface Integration: Seamlessly integrate new features into the UI for easy access.

2. Equip customer support with specialised tools.

   • Internal Tools for CS: Provide platforms (like Admin Retool) that enable customer support to execute actions on behalf of customers.

   • Training and Security: Ensure CS teams are trained to use these tools securely and efficiently.

3. Problems

Problems are issues that hinder the customer’s ability to use your product effectively. While some tickets may initially appear as problems, they could be questions or requests upon closer examination. Genuine problems often arise from:

• Bugs: Errors in the system that need fixing.

• Incidents: External factors or partner-related issues impacting functionality.

Tips: Save all the time you can because you will need it to allocate to solving problems.

1. Allocate time for investigation.

   • Root Cause Analysis: Dive deep into reported issues to understand their origin.

   • Symptom Identification: Differentiate between user errors and actual system faults.

   There are two tricky types of problems:

   1. Easily identifiable but hard to fix.

      • Characteristics: Clear symptoms are reported by users, but solutions are complex.

      • Approach: Prioritize based on impact and allocate necessary resources for resolution.

   2. Hard to identify but simple to fix.

      • Characteristics: Symptoms are confusing or misleading, making diagnosis tough, but once identified, the fix is straightforward.

      • Approach: Improve diagnostic processes to detect these issues faster. CS should have a good troubleshooting guide.

2. Implement solutions

   • Bug Fixes: Log bug-related problems in your tracking system (like Linear) and prioritise them.

   • Incident Response: Activate protocols to manage and mitigate incidents promptly.

Conclusion

By effectively categorising customer support issues into questions, requests, and problems, you can tailor your response strategies to address each type efficiently. This not only enhances the customer experience but also streamlines your internal processes, leading to a more responsive and agile organisation.

In a nutshell, you can deflect and reduce tickets per each type by doing the following:

• Anticipate questions; These can be handled by the design team working on the UX/UI and by the help centre and cs training team to disseminate information accordingly.

• Automate requests; these should be evaluated, prioritised and shipped via your app (self-service to the user) or your CS admin tool (retool in my case).

• Save your time for problem-solving; that’s where your actual support work should be done. Streamline the flow between level 1 and level 2 to ensure serious issues are dealt with quickly.

Action Steps:

1. Disseminate the Framework: Share this categorisation with your CS team and other relevant departments.

2. Implement Training Sessions: Educate team members on how to identify and handle each type of issue.

3. Establish Feedback Channels: Create systems for CS to provide insights to product development and UX/UI teams.

Feel free to share your thoughts or experiences on streamlining the customer feedback loop in the comments below!

Since my first contact with the concept of e-money, I have been frustrated with it. To be honest with you, I couldn’t explain why until recently. Something about the regulatory burden of trying to describe and define all types of exchanges between parties seemed like an impossible task. Even more intriguing was that the name was pretty good – e-money is an electronic form of money, as e-mail is an electronic form of mail or e-commerce of commerce – it made more sense than the concept itself. It should be that e-money works just like cash in a digital world. A central authority issues it, has a set of parameters for verification purposes, has nominal value, serves as a payment instrument and after issuance, it’s free to circulate in the economy with few questions asked. Well, not exactly. But that was not the concept at first.

Let me start from Brazil, where I am from and have more experience. The Brazilian central bank started flexing its payment regulation to incentivise innovation. It was a smart move. New frameworks were created, new types of payment institutions (“PI”) were invented, and many entrepreneurs and investors decided to take their chances in Brazil’s highly concentrated payments and banking sector.

I was at Stone by the time, and we took direct advantage of this by becoming an Acquiring Payment Institution – but that is a conversation for another moment. At the same time, the central bank created the Electronic Money Issuer Payment Institution. As the name suggests, this was a specific PI focused on the activity of issuing electronic money. By hearing that, you could say: “It’s like the balance that appears on my bank app when I deposit or transfer money to it”. Well, not exactly. Bank deposits are slightly different.

The Difference Between E-Money and Bank Deposits

E-Money: Short for electronic money, e-money is digital value stored on electronic devices or remote servers. It’s issued against the receipt of funds and is intended for executing digital transactions or peer-to-peer payments. E-money can be accessed via digital wallets, online accounts, or smart cards, embodying the concept of ‘money on the move’.

Bank Deposits: These are funds placed into banking institutions for safekeeping. Deposits can take various forms, including savings accounts, checking accounts, and fixed deposits, among others. Unlike e-money, bank deposits are not specifically designed for rapid digital transactions but serve as a safe store of value, with the added potential for interest accrual.

E-Money was specifically conceived so non-bank institutions could participate in the payments and financial services industry without doing a bank’s core activity: receiving deposits from part of its customers while lending funds to another part. In economic terms, monetary base expansion is when a bank takes $100 as a deposit, lends $50, and keeps that $100 redeemable for the original depositor. After repayment, the new monetary base increased from $100 to $150 plus interest charged. Meanwhile, funds received by e-money issuers must be safeguarded either by segregating them in a bank account (or by securing them through an insurance policy or guarantee in the UK’s case). E-money issuers cannot use these funds for lending but can invest a small fraction of the total in low-risk assets like government treasury bonds.

As an innovation framework, e-money in Brazil started with a blank canvas but was constrained by a 100% minimum deposit ratio in a segregated account. However, each institution was allowed to design its own rules of engagement. This resulted in each institution writing very detailed rules and registering them as its own payment scheme. As time went by, Brazil had more than 100 newly registered payment schemes – each e-money issuer trying to come up with its own set of rules. The situation becomes even more complex to manage if you imagine that each of these payment schemes is usually interoperating among themselves.

From the liabilities standpoint, both bank deposits and e-money are claims against the institutions that are issuing the papers. So why would the central bank allow this? I don’t think It was a conscious decision, and its always an easy job to make claims after many years have passed.The brazilian central bank goal was to set up the industry for innovation to thrive, and you can say this was achieved in the past few years.

It is important to note that e-money issuers, most of the time, don’t have direct access to their respective central bank’s infrastructure, so they hold their assets and cash positions in banks, who are the de facto holders of central bank deposits.

Banks like Clear.Bank in the UK and Star Bank in Brazil are specialised in providing services to e-money issuers.

Why E-Money Became So Relevant In The Payments Industry?

The ascent of e-money as a dominant force in the payments industry isn’t coincidental; it’s a direct response to the evolving demands of the digital economy and consumer behavior. Several key factors contribute to its growing relevance and adoption:

Financial Inclusion: E-money has a pivotal role in extending financial services to unbanked or underbanked populations, especially in developing regions. By bypassing traditional banking barriers, e-money platforms offer access to financial services via simple mobile devices, bringing millions into the fold of the formal economy. E-money per se has no innovation, but the regulatory apparatus around it facilitated the boom of companies that innovated in the sector.

Regulatory Support and Innovation: Governments and regulatory bodies worldwide have begun to recognise the potential of e-money to enhance their digital economies. Regulatory frameworks are evolving to support the safe and responsible growth of e-money systems, fostering innovation while protecting consumers.

Impact on the Payments Landscape: E-money’s ascent has profoundly reshaped the payments landscape. It has introduced dynamism and flexibility previously unseen, challenging traditional banking institutions to innovate or partner with e-money providers. The rise of e-money also heralds a shift towards a less cash-dependent society, with digital transactions becoming the norm for a wide range of services and purchases.

The relevance of e-money today is a testament to its ability to address the needs of a digital and globalised economy, offering a glimpse into the future of financial transactions. But as we continue to move towards an increasingly digital world, new forms of digital assets circulating in new shapes of networks can change the entire landscape once again.

Introduction to Crypto dollars, Reais, Euros: Fiat-Backed Stablecoins

Diving into the digital finance revolution, one might ask, “What are stablecoins, anyway?” They’re not just another buzzword in the cryptocurrency frenzy; they represent a groundbreaking bridge between the traditional stability of fiat currencies and the innovative, borderless world of blockchain technology. Fiat-backed stablecoins, like cryptodollars, cryptoreais, or cryptoeuros, stand out as digital tokens pegged to the value of traditional currencies, ensuring they maintain a stable value akin to the fiat they represent.

Issued on blockchain platforms, these stablecoins carry the promise of global financial inclusion and efficiency by being easily transferable across borders without the typical constraints of traditional banking systems. They are designed to be as stable as the currencies they’re tied to, such as the US dollar, offering a less volatile alternative to traditional cryptocurrencies. This approach allows them to gain life “on-chain,” free to be traded on secondary markets and programmable into a myriad of applications, starkly contrasting with the geographically and jurisdictionally bound nature of e-money and bank deposits.

While our focus today zeroes in on fiat-backed stablecoins for their direct linkage to well-understood monetary values, it’s crucial to acknowledge the broader stablecoin ecosystem, including crypto-backed and algorithmic stablecoins. These variations promise a rich tapestry of innovation yet to be fully explored in future discussions.

Cross-Border Constraints: E-Money, Bank Deposits, and the Freedom of Stablecoins

When examining the cross-border constraints of e-money and bank deposits, it’s crucial to understand that both are fundamentally tethered to their monetary jurisdictions. This limitation stems from the regulatory oversight, the nature of the deposits, and the traditional banking infrastructure, which is inherently localised. Both e-money and bank deposits operate within the confines of national boundaries, influenced by specific countries’ regulatory frameworks and economic policies. This creates a natural barrier to their global fluidity and interoperability.

Regardless of their digital or physical form, e-money and bank deposits are confined by the monetary policies and regulatory oversight of the jurisdictions in which they are issued. This means that for both, expanding services across borders involves navigating a complex web of international regulations, banking partnerships, and compliance requirements. Retail banking, with its roots in serving local communities, particularly exemplifies this, as its model is built around national network effects that rarely extend beyond its home country’s borders.

In stark contrast, stablecoins — especially those pegged to fiat currencies like the dollar, euro, or real — transcend these traditional barriers by existing on blockchain networks. Once issued, stablecoins are not just digital representations of money; they become part of a global, decentralised ledger that knows no borders. This unique characteristic allows them to be freely traded on secondary markets, integrated into digital applications, and programmed for a wide array of financial transactions without the direct oversight of a central monetary authority. This is interoperability by design.

Unlike e-money or bank deposits, stablecoins can effortlessly cross borders, enabling instant and cost-effective international transactions. Their blockchain-based nature facilitates a level of interoperability and accessibility unparalleled by conventional financial instruments.

The ability to be embedded in smart contracts and other decentralized applications (dApps) further amplifies their utility, allowing for innovative financial services that can operate globally without the need for traditional banking infrastructure.

In the realm of international commerce, stablecoins have already begun to demonstrate their transformative potential. Take, for instance, the case of a small business owner in Brazil, importing goods from a supplier in Europe. Traditionally, this transaction would entail navigating multiple banks, incurring hefty fees, and waiting days for payment processing. However, by leveraging a fiat-backed stablecoin like the cryptoeuro, the Brazilian entrepreneur can instantly and cost-effectively send payment across borders. The supplier, in turn, receives the cryptoeuro and can choose to convert it to their local currency or keep it as a stable digital asset. This not only simplifies the transaction but also significantly reduces the costs associated with currency conversion and international money transfer fees.

Furthermore, remittances, a lifeline for millions of families worldwide, have seen a similar enhancement through stablecoins. Consider workers abroad sending money back home; by using stablecoins, they bypass traditional remittance channels, avoiding exorbitant fees and delays. A real-world example is a blockchain platform that enables Filipino workers in Japan to send remittances home in the form of a fiat-backed stablecoin, which their families can easily convert into Philippine pesos. This method has proven not only faster and cheaper but also more reliable, offering families immediate access to funds.

Navigating the New Risks Associated with Cryptodollars

As we peel back the layers of innovation surrounding stablecoins, it’s crucial to confront the risks associated with their use, especially considering their burgeoning role in the digital economy. One notable concern is the secondary price risk, which hinges on market pricing and confidence in the stablecoin issuer. Unlike traditional fiat currencies, whose value is relatively stable and backed by government entities, stablecoins rely on the trust that issuers have correctly managed and reserved the assets backing them. This confidence is paramount; a wobble in trust can lead to fluctuations in market pricing, potentially causing a stablecoin to drift from its peg, thereby affecting its stability and utility.

The reserves of the issuer present another critical point of vulnerability. The ideal scenario is one where each stablecoin in circulation is backed one-to-one by fiat currency or equivalent assets held in reserve. However, the reality is often more complex, compounded by the difficulty of auditing issuers, especially those operating outside the jurisdiction of the fiat currency to which their tokens are pegged. This geographical and regulatory disconnect can obscure the true state of reserves, raising questions about the issuer’s ability to maintain the peg in challenging times.

Moreover, the issue of redeemability casts a long shadow over the stablecoin landscape. Conceptually, holders should be able to redeem their tokens directly with the issuer, assuming a clear contractual relationship exists. Yet, in practice, many holders may find themselves at a disadvantage if the secondary market tanks. Issuers, particularly those with whom the holders have no direct relationship, may refuse redemption claims for myriad reasons, including compliance with Know Your Customer (KYC) and Anti-Money Laundering (AML) regulations. This situation underscores a critical gap in understanding and expectations: just because a stablecoin theoretically offers a mechanism for redemption doesn’t mean every holder will be able to utilise It Under All Circumstances.

What To Expect From The Future?

In the swiftly evolving narrative of digital finance, stablecoins emerge as both a testament to human ingenuity and a reminder of the complexities inherent in pioneering new monetary forms. They encapsulate the drive towards a more interconnected and accessible global economy, unshackled from the cumbersome constraints of traditional financial systems. Yet, as we chart this unexplored territory, we’re reminded that innovation, while exhilarating, carries its share of risks that must be navigated with vigilance and a forward-thinking mindset.

The conversation around stablecoins, particularly fiat-backed variants, is just starting. It opens up a broader dialogue on how we perceive, utilise, and govern digital money in an increasingly decentralised world. As we delve deeper into the potentials and pitfalls of stablecoins, our journey is as much about shaping the future of finance as it is about understanding the intricate balance between innovation and stability, freedom and regulation. In this dynamic landscape, our collective challenge is to harness the revolutionary promise of stablecoins while ensuring they serve as a secure and stable medium of exchange for all, heralding a new era of financial inclusivity and empowerment. How powerful can a public blockchain token representing a claim on a JP Morgan deposit be? Can you imagine a future where all bank deposits are on-chain, being traded and priced in real time? What impact would that have on the payments market? And other markets?

Long Reads

• I highly recommend reading Nic Carter’s Cryptodollar paper published in 2021. It’s probably the one document that made me realise what I was missing before.

• The EU is front-running the entire discussion on regulated stablecoins and has already published its first set of rules for the MICAr regime.

• This smaller piece from Coinmetrics breaking down Tether's meteoric rise and dominance among stablecoins is also quite interesting.

If you haven’t, read the 1st part of this series here: Beyond Cards and Codes: Unraveling Brazil's Pre-Pix Payment Maze.

We are now entering a new terrain, where the idea of PIX starts to materialise in the market through a relentless pursuit of efficiency, competitiveness and evolution by the Central Bank of Brazil. In this second part, I explore the main elements of PIX’s foundation, the past issues that were solved by the new system and the complexities of its creation.

The Issues

As we have explored in Part 1 – Beyond Cards and Codes: Unraveling Brazil’s Pre-Pix Payment Maze, the STR, created in 2002, is the backbone of the Brazilian Payments System, where money actually exchanges hands between institutions. But this infrastructure was already 20 years old. Although it performed quite well for the simple purpose of money flowing between large financial institutions, its design had a few limitations for further use cases:

• Not 24 / 7 – STR worked from 6:30am to 6:30pm on weekdays, skipping weekends and national holidays. Any other financial system working on top of it would have to work in the same time window or create a workaround to solve liquidity out-of-STR-hours.

• No confirmation message – the STR protocol didn’t have a proper message designed for the confirmation of deposit by the payee’s institution. Consequently, transactions given as settled were refunded hours later without any specific reason, creating complexity in the ecosystem and limiting the development of other payment flows and use cases.

• Protocol limitations – The protocol was designed to move money between two user accounts, two institutions or between a user account and an institution. But this was all. The old protocol was too rigid in terms of extensibility to allow the creation of other payment flows and secondary products.

• Infrastructure access – the Settlement Account, created as an instrument for Payment Institutions to access the backbone, was too complex to operate, and few institutions applied for it.

In addition, STR was only running a small portion of the TEDs in Brazil since all the banks were sending most of the transactions through SITRAF. Imagine what would have happened if, during its creation, Pix was optional. This was the exact case before.

To summarise, we can say that while TEDs had a great infrastructure for that time, the payment scheme above lacked more rigid rules to shape the evolution of the transactions in terms of REACH, TRUSTNESS and COST EFFICIENCY.

In an amazing technical analysis called Lessons From Pix: How to Build a Real-Time Payments Platform at its Full Potential, Mariana Cunha e Melo and Jonas Abreu give us a great summary of the issues the CB stood out to solve with the new design:

The Protocol Issue – It must be standardised but not over-specified. The communications protocol of many payment rails suffers from over-specification. Over-specification happens when, during the protocol design process, one establishes all the use cases to be supported instead of creating a backwards-compatible extensible structure that is able to represent all those use cases through extension. A familiar example in the world of payments is the ISO 20022 payment messages. Everything that goes into the payment needs to have a specific and already specified field in one of the messages. When another type of payment appears, instead of simply extending the message in a backwards-compatible way, you need to standardise the new type of payment and fully specify it in the protocol, which can take years to happen.

The need for standardisation is apparent. Without it, the many participants of the system cannot communicate with each other. But once the protocol is built to accommodate one specific need with specific fields and flows, it has to be extended by adding more specific fields and re-signifying old fields to evolve. The result is an intricate, inexpugnable web of old grammar with new semantics to decipher, and a big pile of unnecessary data travels in the system.

The Trust Issue – is twofold. From a regulatory standpoint, moving money around requires the guarantee that money will not disappear or multiply by mistake or by malicious design. For financial system regulators, that is an issue of systemic resilience and of currency stability. If a player is able to print digital currency or loses track of where someone’s money is located, the value of the national currency or the trust in the financial system may be severely affected. The risk of a systemic failure is at the heart of the Brazilian Central Bank’s concerns in regard to payment systems.

The second aspect of trust in this context is that connecting the financial system takes an arrangement between any two account service providers in the land. This requirement unfolds into several issues, from the complexity of business arrangements to the technical arrangements and issues, such as determining the source of truth.

The Issue of Reach – The issue of reach comes from the fact the payment system needs to connect any two accounts from any two account service providers. Solving this issue is the difference between having a walled garden where only a few member institutions can provide access to the payment service and having an ecosystem where every user knows that whatever economic transaction they are making, they can use the payment service to liquidate the price because everyone is reachable. That is what enables users to leave their houses with only one payment method because they know they will be able to use it for their every need.

Solving The Puzzle

However, solving that was no easy task. There were infinite paths to take, from promoting a new regulatory framework, allowing for a private infrastructure to take place, creating a public payment scheme but handing over the infrastructure, or building both but making them optional. A lot of trade-offs were on the table.

The final approach is, in my opinion, one of the strengths of PIX’s success over the last 3 years:

1. Create a new financial market infrastructure called SPI – Instant Payments System – another real-time gross settlement system, working on top of STR but 24/7. Link to SPI webpage here.

2. Create DICT – a centralised payment alias database that facilitates the identification of bank and payment accounts by linking personal keys like phone numbers, CPF (personal tax number) or emails to them. Link to DICT API specification here.

3. PIX, the scheme – a payment scheme created, incentivised and managed by the Central Bank, containing rules about how participants interact, what use cases are supported, what types of transactions can be charged and how, dictating minimal UX / UI parameters. Link to PIX rules here.

4. PIX, the brand … I know, I know, branding is just as part of a payment scheme as deposits and loans are part of banks. Some people say that Visa and Mastercard are brands first and networks second. I can’t disagree. The CB also did that, and quite well. They’ve chosen a name that was easy to say, easy to use and very catchy. They created brand guidelines, applications, DO’s and DONT’s. Link to the brand book here.

5. Direct and Indirect Access – From the start, the CB understood that access and reach were essential to full adoption. They created two types of access to make the system and scheme adoption mandatory to every institution in the S1 and S2 tiers but easily available to every other registered institution. In one shot, the CB gave full access and created another market for the BaaS – bank as a service – players.

   1. Direct Participants would hold accounts in SPI called Instant Payments Accounts, accessing them directly through the RSFN (National Financial System Network).

   2. Indirect Participants would be able to access all functionalities but through a Direct Participant, borrowing their access to the infrastructure and paying a fee for that.

This is what the new high-level architecture looks like after the SPI and DICT introduction.

I highlighted the Payment Institution E and Indirect Participant, accessing SPI through Bank C. To make this happen, Bank C would have to offer a transactional account with specific API functionalities designed for Indirect Participants. In this case, Payment Institution E exists in the PIX scheme, has its identifier and can be found by other institutions, but all its transactions are sent and settled through Bank C.

All other institutions in the drawing above are directly connected to SPI and have Reserves or Settlement accounts underneath – in the STR instance. They use their accounts in the STR system to feed their Instant Payment Accounts in the SPI instance. During STR non-working hours, institutions can borrow money from the central bank to solve liquidity issues, similar to what is done in STR.

This is how a set of transactions would look like:

• Payment Institution E is an indirect participant and uses Bank C as its direct participant

• Bank C uses its Reserve Account to top up its Instant Payment Account with R$ 1000

• Customer C (from Bank C) sends R$ 100 to customer D (in PI D) via PIX

• Customer E (from PI E) sends R$ 200 to customer A (in Bank A) via PIX

Real-Time Transactions Everywhere, Anytime… 

What if we invert STR and SPI? Could the entire market drink from a continuously available font? Imagine card transactions settling on weekends, securities too… One of the core characteristics of STR is its specifically built messages supporting 2nd layer infrastructure systems like SILOC, SITRAF or even BM&F for securities settlement. All those secondary systems were restricted by STR working hours. Building those messages into SPI could create second-effect changes of the same magnitude as PIX. Could that change in the near future? What are your thoughts?

Creating a new payment paradigm

Creating a new payment system like Pix and ensuring its adoption in a market dominated by entrenched players like Visa and Mastercard is a complex challenge. The incentives that card schemes have put in place over many years make them an integral part of the financial ecosystem. However, introducing a new system isn’t solely the domain of policy-making; private markets can play a significant role as well. Here’s a breakdown of the various aspects involved:

1. Forming a Market & Breaking Through:

   • Regulatory Push: Policy-driven initiatives can provide the initial momentum. A central authority’s backing can build trust quickly, streamline implementation, and ensure widespread collaboration. In PIX’s case, the Central Bank enforced adoption through mandatory implementation by S1 and S2 institutions.

   • Cost Advantages: If the new system offers cost savings over traditional card schemes (lower transaction fees, reduced infrastructure costs, etc.), it becomes attractive for both merchants and consumers, as was the case with PIX.

   • User Experience: Speed, convenience, and versatility in transaction methods (like QR codes and instant transfers) can make the new system more appealing. But not just that. The Central Bank also created a very compelling and detailed guideline of the UX UI for end users of PIX, mainly inspired by the UK Openbanking Initiative that has pioneered this approach – link to the UK Openbanking UX UI Guidelines here.

   • User Experience: Speed, convenience, and versatility in transaction methods (like QR codes and instant transfers) can make the new system more appealing. But not just that. The Central Bank also created a very compelling and detailed guideline of the UX UI for end users of PIX, mainly inspired by the UK Openbanking Initiative that has pioneered this approach – link to the UK Openbanking UX UI Guidelines here.

   • Integration: Ensuring that the new system integrates seamlessly with existing financial infrastructures is crucial. This reduces barriers to entry for banks and merchants.

2. Policy Making vs. Private Markets:

   • Policy Making: Government or central bank initiatives can quickly roll out and mandate the adoption of new systems. This is particularly true in markets where the public good is prioritised over commercial interests.

   • Private Markets: While private entities can develop innovative solutions, their challenge lies in achieving scale and trust. Collaborations or consortiums of private entities can introduce new systems, as seen with cryptocurrency-based solutions or platforms like UPI in India.

3. Externalities:

   • Positive: A new system can promote financial inclusion, reduce transaction costs, and boost economic activity by offering a more efficient method of payment.

   • Negative: Potential risks include cyber threats, initial resistance or distrust, and disruption for businesses that need to integrate new payment methods.

4. Incentive Contrasts:

   • Pix vs. Card Schemes: While card schemes might offer loyalty points, cashback, and credit facilities, Pix focuses on instantaneous transactions, lower fees, and universal accessibility.

   • Pix vs. TEDs/DOCs: Pix operates 24/7 and settles instantly, while TEDs and DOCs had time restrictions and could take longer. Also, Pix often has lower or no fees compared to these traditional transfer methods.

   • Pix vs. Boletos: Boletos are payment slips often used for bill payments. They’re not instantaneous and require manual processing. Pix, being digital and immediate, offers a significant advantage in terms of speed and convenience.

In summary, while policy-driven initiatives like Pix have an edge in terms of trust and widespread adoption, private markets can also introduce disruptive payment methods. The key lies in understanding the existing market incentives and offering significant advantages in terms of cost, efficiency, and user experience.

The Main Elements that Led to Pix’s Adoption

• Accessibility: Pix was designed to be inclusive. Being smartphone-based (though not exclusively), it tapped into the widespread use of mobile devices in Brazil, making it more accessible to a broader demographic.

• Speed: Instantaneous transactions, regardless of the hour or day, were a game-changer. Unlike TED or DOC, which had restrictions, Pix operates 24/7.

• Cost: Transactions through Pix are often cheaper, especially for personal use cases. Businesses also benefit from it, usually paying less than a TED for a transfer and less than a debit card transaction for an in-store or online payment. This economic advantage drove adoption among merchants and consumers alike.

• Online User Authenticated Transaction: As it’s based on account-to-account transactions, PIX transactions are always authenticated by the sender through its own bank’s app. Card transactions online, especially debit ones, were usually more risky as cards didn’t offer an authentication method. Even after PSD2, Brazil was still lacking functionality, and when present, the user experience was often very poor.

• Flexibility and Versatility: With various ways to transact (e.g., using a QR code, phone number, email, or Pix key), it appealed to a wide range of users and use cases. The way PIX’s protocol was designed allowed for the expansion of use cases without compromising the core functionalities of the system.

• Aggressive Promotion & Education: The Central Bank and other stakeholders carried out widespread campaigns to educate the public about Pix’s benefits, ensuring users understood and trusted the system.

Central Bank Roles

As highlighted by Mariana and Jonas in Lessons from Pix, a lot of the project’s success is based on the fact that PIX was a government-driven full-stack payments initiative:

The Brazilian Central Bank created the rail, its rules, and its systems. It is also the one who operates the system and provides the necessary infrastructure. BCB is the financial and payments market regulator. The fact it was the financial system regulator that drove the whole project had many advantages in Brazil’s case. Although it might be possible to make different arrangements work in other countries, there are few incentives for a market-driven solution to strive for global optimisation and build a truly open, cost-effective, and flexible rail. To facilitate the process of extrapolating Pix’s case to a broader set of circumstances, however, the role the BCB plays in the Pix rail may be identified in this paper more generally as a Payment System Operator or PSO.

Let’s detail the roles to make it clearer to our readers:

1. Regulator:

   • As the monetary authority of Brazil, the Central Bank is responsible for overseeing and regulating the financial system, ensuring stability and integrity.

   • Within the context of Pix, the Central Bank extrapolates the typical Scheme Institution role below to make some of its rules mandatory. Additionally, with its mandate of promoting competitiveness in the market, the CB can prioritise user benefits and market development instead of participants’ margins. That was the case when they made all P2P and P2B transactions free for all users. Participants may charge businesses when they are receiving payments through PIX, though.

   • It also mandates standards for interoperability, ensuring that all banks and financial institutions in the system can seamlessly transact with each other.

2. Payment Scheme Institution:

   • The Central Bank defines Pix’s operational, security, and business rules. This means it sets the guidelines for conducting, settling, and reporting transactions.

   • As the payment scheme institution, the Central Bank also manages risks within the system, ensuring Pix remains a safe and trustworthy payment method.

3. Settlement and Accounts AliasInfrastructure (SPI and DICT):

   • SPI is the core infrastructure that underpins Pix. It’s the backbone that processes and settles transactions in real-time.

   • The Central Bank is responsible for maintaining and overseeing this infrastructure, ensuring it remains robust, secure, and efficient. This is crucial given the need for instant settlements.

   • By being in control of SPI, the Central Bank ensures the high availability of the system and its resilience against potential cyber threats and technical glitches.

Here, you can see the difference between PIX, where the Central Bank plays multiple roles at once, and a Private Card Scheme, where multiple institutions are playing different roles:

By playing these roles, the Central Bank guarantees Pix’s smooth functioning and instils public trust in the system. Their involvement ensures that Pix adheres to the highest security, reliability, and efficiency standards. Moreover, by being both the regulator and operator, the Central Bank is uniquely positioned to swiftly adapt and innovate the system as required.

This multi-faceted role of the Central Bank is crucial for the success of an instant payment system like Pix. It ensures not just technical robustness but also systemic integrity and public confidence.

Concerns About Central Bank’s Roles

When a single entity, like the Central Bank of Brazil in the case of Pix, assumes multiple roles traditionally distributed among different organisations, concerns about potential conflicts of interest and concentration of power arise. Let’s break down these concerns:

1. Concentration of Power:

   • With the Central Bank acting as both the regulator and operator of Pix, it consolidates a significant amount of power and control over the payment system. This concentration can potentially reduce checks and balances.

2. Regulation vs. Operation:

   • As the regulator, the Central Bank’s role is to establish rules, oversee the system, and ensure fair competition. As the operator, its interest lies in the smooth functioning and adoption of Pix.

   • There could be situations where what’s best for the operation of Pix might not align with broader regulatory goals. For instance, a decision that boosts Pix’s performance might inadvertently sideline other payment methods or reduce competition.

3. Commercial Interests vs. Public Good:

   • The sustainability of the payment market as a business is very important. While the Central Bank’s primary role is to serve the public interest, managing a system like Pix could also present commercial conflicts, especially related to other private Payment Schemes and Infrastructures that are core businesses of publicly traded companies.

   • There might be concerns that decisions taken by the central bank could harm or even kill private profitable businesses.

4. Transparency and Accountability:

   • With multiple roles consolidated, there could be concerns about the transparency of decisions. It might be challenging to discern whether a decision was made with the regulator or operator hat on.

   • This blurring of roles can lead to accountability concerns. If an issue arises with Pix, the Central Bank would oversee and potentially critique its work.

5. Risk Management:

   • Having the same entity regulate, define the payment scheme, and operate the infrastructure means the Central Bank assumes all the systemic risks. A failure at any level could have cascading effects on the entire payment ecosystem and potentially the broader financial system.

To address these concerns, the Central Bank maintained stringent internal checks and balances and kept a very transparent decision-making process, while making sure that all participants – represented through the market associations – were heard. By doing so, it ensured that its roles did not conflict and that the integrity of the Pix system and the broader financial ecosystem remained intact.

The final result is the incredible adoption seen in the last years, with PIX now accounting for more transactions than credit or debit cards.

If you are like me and want to dive even deeper, please read the amazing technical analysis called Lessons From Pix: How to Build a Real-Time Payments Platform at its Full Potential from Mariana Cunha e Melo and Jonas Abreu.

Before diving into the state of the Brazilian payment landscape leading to the introduction of Pix, it’s essential to understand the various components that make up a payment ecosystem. Each component plays a distinct role, often intertwined, to ensure the smooth functioning of the system. In this specific case, understanding the complex set of incentives created through card schemes, bank conventions, and other dynamics is a key element to solving the puzzle of ‘what made Pix such a tremendous success’.

• A payment scheme is the “rulebook” of a particular payment method. It sets out the rules, standards, and procedures that dictate how transactions are executed, cleared, and settled. This encompasses the rights, responsibilities, and obligations of all participants.

• The payment infrastructure is the “machinery” that facilitates transactions. It includes the technological and operational frameworks, like hardware, software, and communication channels, ensuring transactions are efficient, secure, and reliable.

• A payment scheme institution is the “architect” behind a payment system. This entity usually designs the payment scheme and could also be involved in developing or commissioning the required infrastructure.

• Payment institutions act as intermediaries in the payment process. These entities are licensed to provide payment services to end users. Their roles vary, from facilitating funds transfers to managing electronic money.

• The regulator, often a central bank or a designated governmental body, oversees the entire payment landscape. It ensures the payment system’s stability, reliability, and fairness, setting standards and guidelines and ensuring compliance.

With these components in mind, let’s explore Brazil’s payment landscape before introducing Pix and see how these various elements interacted and evolved.

The Last Decade

Before the introduction of Pix, Brazil boasted an intricate patchwork of payment schemes in different flavours. Open schemes managed by government authorities like TED, open schemes managed by private companies like Visa and Mastercard and yet closed schemes managed by the banks themselves.

Card adoption surged over the five years leading up to Pix’s launch with a consistent CAGR of 17%. Though Visa and Mastercard retained their dominance, Elo – a homegrown initiative co-founded by leading banks like Caixa, Bradesco, and Banco do Brasil – rapidly gained traction using the issuing power of its shareholders. The industry was reasonably advanced, with NFC gaining traction, settlements centralised in a regulated system and the presence of mobile wallets. The Central Bank and CADE – the antitrust authority – were already fomenting a lot of competition and innovation through new regulation, attracting more investments to the market.

Brazil also had significant use of wire transfers and invoice-based payments outside of cards. There were two predominant wire transfer forms: TED (Express Wire Transfer) and DOC (Credit Transfer Document). “Boletos” were the Brazilian version of invoice payments, but they utilised a standardised barcode format readable by any bank, containing not just payee details but also payment terms like due dates, penalties, and interest.

Both TED, DOC and Boletos are in fact payment schemes, like credit and debit cards from Visa and Mastercard. They defined the rules, dynamics, economics, and structures behind each instrument.

Wire Transfers in Brazil – Why So Many Flavours?

TED – The Express Wire Transfer Payment Scheme

It was an open public scheme, which was open to new participants adhering to a set of rules and managed by a public institution – the Brazilian Central Bank. To send a TED or a DOC, the user had to fill in the same fields: Transaction Value, Bank Number, Bank Branch Number, Account Number, and CPF (Brazilian Personal Fiscal ID) or CNPJ (Brazilian Business Fiscal ID). TEDs were supposed to be cleared and settled in just a few seconds, but the protocol didn’t account for a confirmation message from the receiving bank, properly notifying the deposit of funds. This was basically done on an assumption basis after 120 minutes – the TED rule’s SLA for returning failed transactions to the original institution. Besides that, TED was a fast instrument, but it was also expensive for end users as we are going to see later on.

DOC – The Slower Wire Transfer

Meanwhile, the DOC was another scheme, for values up to R$ 5000 that was supposed to be cheaper than TEDs. The transactions were cleared by midnight and settled the next business day, which meant the money took at least one full working day to arrive at its destination. The reason for that latency is that this conversation starts to get tricky. Underneath each payment scheme were different Payment Systems, the so-called Financial Market Infrastructures – or FMI in short. Let’s dig into them.

STR – The Reserves Transfering System: A 1st Layer Financial Market Infrastructure

Launched in 2002, STR served as a real-time gross settlement system, acting as the foundation of the Brazilian Payments System (SPB). All financial entities held reserve accounts at the central bank, with STR facilitating transfers between these accounts. Think of the Central Bank as the bank of the banks, their reserve accounts are basically a checking account, and the STR is the internal transfer system that moves money between two banks. Transactions on STR were fast, often settling in seconds. Pricing was based on a cost-coverage model. Despite its efficiency, STR wasn’t intended for consumer or commercial transactions. It only operated on weekdays from 6:30 am to 6:30 pm, and transfers didn’t have a message from the receiving bank confirming the deposit of funds. Additionally, the official SLA defined by the TED scheme refunds in case of failure was two hours, and reinforcement was lacking. Coupled with the rising liquidity demands, banks felt the need for an additional layer atop STR. One that would reduce their cash needs, leading to the creation of SITRAF.

Example of 3 TEDs sent using STR:

1. User A sending R$ 30 from bank A to user B in bank B.

2. User B sending R$ 50 from bank B to user C in bank C.

3. User C sending R$ 30 from bank C to user A in bank A.

SITRAF – Funds Transfer System: A 2nd Layer Infrastructure. Redundancy For What?

While the STR was created and operated by the Central Bank, SITRAF was a product from CIP – the Interbank Payment Clearinghouse created by the Bank’s association FEBRABAN. SITRAF was a “near real-time” net differed settlement system that utilised small intervals of 5 – 10 minutes to clear transactions between banks, calculating net positions and reducing risk and the demand for cash in the system. Even though SITRAF aimed for efficiency, it came with its own cost. The banks created a private convention – a parallel set of rules – to guide TED transactions sent through SITRAF. Remember that the TED scheme already had a set of rules defined by the Central Bank and mandatory for every institution using TED.

Since transactions were usually charged from the user sending money – the Payer – a particularly contentious rule was introduced, creating the TIB (later RCO) – a fee paid by the payer’s bank to the payee’s bank – for operational costs reimbursement purposes. Reimbursing a bank for the ‘burden’ of receiving more deposits sounds almost unbelievable, but it is still true today.

In 2017 the TIB was set at an astonishing R$ 1,19 per transaction, more than 10x STR transactional fee. Even more interesting was that the private convention determined that any TED was subjected to TIB charges, regardless of the system used for settlement. Ultimately, banks might charge R$ 10 for a single TED, despite an additional and slower layer meant to ease liquidity constraints… Let’s put it together if it is unclear: they created another redundant, slower, more expensive infrastructure with an extra fee embedded as an operational cost reimbursement to limit their cash needs, and they managed to still charge everyone north of R$ 10 per wire transfer.

Example of 3 TEDs sent using SITRAF:

1. User A sending R$ 30 from bank A to user B in bank B.

2. User B sending R$ 50 from bank B to user C in bank C.

3. User C sending R$ 30 from bank C to user A in bank A.

4. Transactions are calculated in SITRAF and the resulting net position bank B needs to send R$ 20 to bank C.

Introducing Payment Institutions in STR

In the world of finance, accounts often play unique roles. Such was the case when the Central Bank created distinct accounts for banks and payment institutions. Settlement Accounts were the new addition to the system, designated for payment institutions, existing for the sole purpose of settling transactions. Imagine an account with very simple functions of sending money – both to other institutions and customers of those institutions – and receiving money from both types of senders. Meanwhile, Reserve Accounts allowed banks to conduct interbank operations and multiply currency, something that payment institutions could not do.

Besides that, Payment Institutions were also allowed to have bank accounts themselves. In fact, this was Stone’s main settlement mechanism for many years – every morning, we would send money to our account in Itau and spread the money with internal transfers to Itau’s customers.

Banks were not allowed to do that. They operated interbank operations through STR, where borrowing and lending money from each other was conducted directly in the system.

The Money Printing Machine for Stone

Stone’s venture into the STR (Brazil’s real-time gross settlement system) began with a simple need to test their newly opened Settlements account. A R$ 0,01 transaction was sent from Stone’s checking account in Itau to their STR settlement account, arriving a few seconds later. We repeated 9 more R$ 0,01 transactions and the tests were declared a success. We celebrated!

But the next morning brought a surprise. Itau had sent Stone R$11.90 for the TIB fee. After sending R$ 0,10 cents in 10 different transactions from one account – in Itau – to another – in the STR – we made R$ 11,90 minus STR operational costs of a few cents. So we repeated the tests, this time with 100 transactions. Repeating the test yielded another R$119.00.

Breaking the Convention

Stone’s team quickly realized what had happened. In their efforts to understand the system, they’d stumbled upon a loophole, inadvertently creating a money-printing machine for themselves. With deep knowledge of the rules, they knew this was a mistake and promptly sent the money back to Itau, explaining they weren’t adhering to the convention that created the TIB. The problem was, when fully operational, Stone’s STR account would make millions of transactions every month, creating an unbearable and artificially created cost.

The situation sparked confusion among the banks, leading to an endless exchange of emails, policies, guidelines, and documents. It took the intervention of the Central Bank to clarify that the TIB was not a mandatory fee. Stone had proven its point.

Emboldened by this victory, Stone began providing settlement APIs to other fintechs, rapidly becoming responsible for a significant portion of the STR transactions in a day. Realising they’d been outmanoeuvred, the traditional banks were forced to change the convention rules, reducing the TIB from R$1.19 to a few cents per transaction.

This tale of innovation, understanding, and agility paints a vivid picture of Stone’s entrance into a complex financial landscape. Through detailed study, clever innovation, and open communication with regulatory bodies, Stone challenged traditional banking norms and paved the way for a more cost-efficient and equitable system.

Boletos: A Unique Payment Method in Brazil

In Brazil, the “Boleto Bancário” is not merely a barcoded payment slip; it’s a financial instrument that carries a wealth of information and specific terms. A Boleto can contain a due date, clearly informing the payer of the payment deadline. If a payment is missed, the Boleto can contain calculated fines and interest, enforcing a penalty for late payment.

Additionally, Boletos can include other terms and conditions, such as discounts for early payment or specific instructions related to the purchase. These parameters are embedded within the barcode of the Boleto, allowing for automated processing by banks and payment institutions.

The adaptability of Boletos makes them suitable for various purposes, such as paying bills, purchasing goods, or even making loan payments. The ability to embed complex terms within a simple payment slip has turned Boletos into a versatile and widely accepted form of payment nationwide. Whether paid online, through a mobile app, or at a physical location, the Boleto system ensures a consistent and comprehensive payment experience, reflecting the multifaceted nature of financial transactions in modern Brazil.

SILOC

SILOC, which stands for “Sistema de Liquidação Diferida das Transferências Interbancárias de Ordens de Crédito,” is a net deferred settlement system used in Brazil.

In this system, financial transactions are not settled individually as they happen. Instead, they are accumulated over a specific period, typically a business day, and then settled all at once at the end of that period.

Here’s how it works:

1. Accumulating Transactions: Throughout the day, banks and other financial institutions record the transactions made between them. This includes all kinds of payments and transfers, such as interbank operations.

2. Netting Process: At the end of the specified period, all the transactions are netted out. This means that the payments and receipts between each pair of institutions are summed up to a single net value. For instance, if Bank A owes Bank B $100 and Bank B owes Bank A $60, the net value would be a $40 payment from Bank A to Bank B.

3. Settlement: Once the net values are calculated, the transactions are settled in a central location, typically the Central Bank. This might involve transferring money between the reserve accounts of the institutions involved. The deferred nature of the settlement allows for efficiency, reducing the number of actual transactions that need to be processed.

4. Reconciliation and Reporting: After settlement, reports are sent to the participating institutions, allowing them to reconcile their accounts and ensure everything is processed correctly.

The SILOC system provides efficiency and reduces costs for the financial system. Handling transactions in a net deferred manner minimises the immediate liquidity demands on banks and allows them to manage their funds more effectively. At the same time, the consolidated and well-regulated nature of the system ensures its reliability and robustness, contributing to the stability of Brazil’s financial infrastructure.

BPS – Brazilian Payment System Overview

After diving into the different pieces of this puzzle, it’s nice to look at the whole picture together. This scheme shows Brazil’s payments market layers through the last decade.

• Layer 0 – STR – where the money actually exchanges hands from institution to institution. Each participant institution has to open an account with the Central Bank. Payment institutions have settlement accounts, while banks have reserve accounts. SELIC serves as a central hub where federal government bonds are held, and transactions related to those bonds are settled.

• Layer 1 – Other financial market infrastructure – SITRAF, SILOC, Visa and Master legacy clearing houses and COMPE for checks.

• Layer 2 – The payment schemes – with their guidelines, networks, authorisation systems and institutions.

• Layer 3 – Market participants – mainly banks and payment institutions.

A second snapshot shows the addition of SLC – Sistema de Liquidacao de Cartoes – another piece of infrastructure created by CIP and made mandatory by the Central Bank for the settlement of card transactions. SLC was just a clearing layer to calculate the net position of all participants in the card schemes, using SILOC underneath to finalise the settlement process. The system went live in 2017, replacing legacy clearing houses from Visanet and Redecard.

After deciding to leave my last job, I started talking to many companies and investors in the UK and Europe to explore new opportunities and think about my next big step. Whenever I told my story and spoke about the products I built or the changes we promoted in the Brazilian market through Stone’s policymaking team, one topic always came up as a unanimous interest: Pix, the Brazilian Fast Payments system that took over the country by storm. Its adoption rate makes it one of the world’s most successful implementations of any payment system. However, I have a specific perspective on its success and the elements contributing to this outcome. So, I decided this was a great moment to start a personal blog, helping me organise my thoughts and explain more in-depth how the payments landscape has evolved in the country during the last decade.

This is it! I will start with a three-part series exploring Brazil’s payment landscape before and after Pix. The dynamic world of payments is ever-changing, and Brazil stands as a testament to innovation, growth, and transformation. From the complex web of traditional payment methods to the groundbreaking introduction of Pix, the country’s financial landscape is rich with lessons, triumphs, and glimpses into the future. In this comprehensive three-part series, we’ll guide you through Brazil’s intriguing payment history, unravel the core elements that made Pix a remarkable success, and venture into the exciting possibilities ahead. Whether you’re a financial enthusiast, tech aficionado, or curious explorer, this series promises a captivating ride through the maze of payments in Brazil.

• Part 1 – “Beyond Cards and Codes: Unraveling Brazil’s Pre-Pix Payment Maze” – I explore the fascinating world of Brazil’s payment systems before Pix. From the everyday ‘Boletos’ to the entangled world of TEDs and DOCs, we uncover the systems, schemes, and strategies that shaped Brazil’s financial ecosystem.

• Part 2 – “Pix Foundation: The Elements of Success” – Discover the secret recipe that led to the success of Pix. I’ll dissect the infrastructure, policies, and innovations that allowed Pix to take the country by storm.

• [Coming soon] Part 3 – “Future Discussions: From Nupay to Payment Initiation, What Does the Future Look Like?” – Fasten your seatbelts as we travel into the future of Brazilian payments. What awaits beyond Pix? I’ll delve into emerging concepts like Payment Initiation in the scope of Open Finance and explore interesting strategies some players are adopting, like Nupay from Nubank.