The One Word a16z Left Out of "The Money Flow Is the Moat"

a16z published "The money flow is the moat" this week, and the argument landed: the durable businesses sit in the flow of value and take a clip. Visa processes $14 trillion annually and keeps 1.5%. Stripe powers millions of online businesses and extracts a percentage from each one. The railroads charged for moving the grain, not for owning the train.

I've spent years sizing up crypto companies with a single filter: if you don't have a moat, you don't have a business. So I'm with the thesis. The problem is the VC framing stops one layer too early. When the rails are open and programmable—which is the entire promise of crypto—access to the flow stops being scarce. Anyone can sit in it. And a moat everyone can build isn't a moat.

The missing word? Trust.

Volume You Rent vs. Trust You Earn

I was reviewing a DeFi protocol's pitch deck last month. Impressive hockey-stick volume chart. Transactions through the roof. Then I asked about their fee structure.

They'd dropped transaction costs to near-zero. The volume was real. It was also a coupon.

Fee-bought volume is mercenary. It showed up for the discount, and it leaves for the next one. Look at the customer acquisition playbook from every yield-farming summer: offer unsustainable returns, screenshot the TVL for the deck, hope you can build something sticky before the liquidity moves to the next farm.

Cheap fees aren't a moat. They're a crowd you rented.

Compare that to how Visa built its business. They didn't win on price. American Express often charged merchants more and still kept premium customers because of the trust embedded in the product: chargebacks, fraud monitoring, dispute resolution, the entire operational infrastructure that means a consumer can swipe without thinking.

The volume Visa processes isn't subsidized. It's earned. Because the fee isn't rent-seeking—it's the price of someone else eating the 3am fraud call.

When Bezos Meets Blockchain

a16z borrows Bezos for the kill line: "your margin is my opportunity." He's not wrong. Amazon obliterated retail margins and forced every competitor to justify their markup.

But here's what the crypto pitch decks forget: Bezos also built the returns department. Some of that retail margin was never rent. It was the price of trust. The ability to return something without a receipt. The guarantee that if the package doesn't show up, someone answers the phone.

Strip the fee, keep the speed, and you haven't disrupted the middleman. You've removed the airbag.

I watched this play out during the exchange blowups of 2022. FTX processed transactions cheaper and faster than most traditional platforms. They also commingled customer funds and evaporated $8 billion. Turns out the "expensive" incumbents like Coinbase were charging for boring things: segregated accounts, audit trails, insurance, regulatory compliance.

The margin wasn't friction. It was accountability. And when it disappeared, so did trust.

The Railroad Parallel Nobody's Discussing

Open rails don't kill the middleman. They just shift what's scarce.

In the 1800s, railroads democratized access to transportation. Suddenly anyone could move goods coast-to-coast. The infrastructure was open—if you had a cargo, you could pay to ship it.

But the companies that survived weren't the cheapest shippers. They were the ones who delivered reliably. They protected against theft. They had liability insurance. They built reputations for getting your grain to market intact, not just fast.

The flow of goods became abundant. Trust became the bottleneck.

Crypto is running the same script. When anyone can fork the code, deploy a contract, and sit in the transaction flow, the technology stops being the differentiator. You're competing on the one thing code can't replace: the trust that when something breaks, someone fixes it.

What "Open and Programmable" Actually Costs

This is where the philosophical purity of crypto meets the operational reality of running a business.

Decentralization is elegant in theory. In practice, it means there's no one to call when the smart contract drains your wallet. "Code is law" works beautifully until the code has a bug, and then you're $600 million short with no legal recourse and no customer service number.

I get why the narrative is seductive. Disintermediation sounds like liberation. Cut out the rent-seeking middlemen. Let people transact peer-to-peer. Keep the fees low and the rails open.

But what looks like rent-seeking often includes services people forgot they valued: dispute resolution, fraud protection, regulatory compliance, someone who answers when things go wrong.

The traditional payment processors aren't charging 2.9% + $0.30 because they're greedy. They're charging it because they're absorbing risk, maintaining infrastructure, and navigating a compliance landscape that would bankrupt most startups on the first audit.

When you compress that margin to zero, you're not eliminating waste. You're offloading liability onto the user. And users—especially the institutions you're trying to court—will pay to avoid that.

The Question You Should Ask Monday Morning

So before you compress the next fee, figure out which part is rent, and which part is the reason your customers sleep at night.

If you're evaluating a crypto investment, a DeFi protocol, or a blockchain business model, here's the filter I use:

What happens when something breaks?

• Is there a human I can call?

• Is there insurance that covers losses?

• Is there a legal entity that assumes liability?

• Is there an audit trail that satisfies regulators?

If the answer is "code is law" or "the community will figure it out," you're not investing in a business with a moat. You're investing in technology looking for a business model.

The money flow might be abundant. But trust is still scarce. And in a world of open rails, trust is the only thing left worth paying for.

Compress the wrong margin and you didn't build a business. You rented a crowd and inherited someone else's liabilities.

---

What to do next: If you're advising clients on blockchain investments or evaluating crypto platforms, add one question to your due diligence: "Who eats the loss when this fails?" If the answer is "the user," that's not disruption. That's risk transfer. And risk transfer doesn't scale.