The $100 trillion plumbing problem

$100 trillion.

That's what businesses pay each other every year. Globally.

The US alone moves north of $25T. Yet despite this scale, the infrastructure underneath it is embarrassingly antiquated.

ACH and wire handle over 75% of total volume and take 2-3 days to settle. Paper checks still account for $3T – 15% of volume. I authorized two this morning.

Cards, real-time payments, stablecoins? Less than 10% combined.

We can land a rocket on Mars. We cannot move $50,000 between two business accounts without a 3-day wait and an AP clerk to ensure it goes through.

That gap is the opportunity.

Think of the first issue as the map. The next four go deep on each layer of the stack — rails, infrastructure, middleware, and platforms.

Same format each time. The mechanics, the problems, the progress, what’s left to solve, and what it means for founders and operators building in this space.

Think of B2B payments as a five-layer cake. Each layer has distinct participants, distinct economics, and distinct problems.

Here's how a single transaction works in practice.

My fractional CFO firm sends a $10,000 invoice to a SaaS client. Client pays by credit card. One click. Done.

That click triggers a cascade across all five layers simultaneously, touching seven distinct parties, crossing multiple networks, and getting shaved by fees at almost every stop.

Corporates. SMBs.

They sit at the top of the cake.

Bridges (my firm) does the work. Sends the invoice. The client pays it. Neither party cares how mostly — until the fees get so large they can no longer be ignored. Fortune 1,000 retailers, airlines, and others are spending real money fighting interchange fees that cost them $150-200B a year.

Small firms like mine just want to get paid. On time. Without thinking about it.

Which is exactly why platforms exist.

Whoever owns the software owns the payment. That's the defining dynamic of this layer.

I wouldn’t have started my advisory business if I had to create every invoice and chase every payment. Anchor does it for me. And because it owns the workflow, it owns the payment as well.

It costs me $5 per invoice and ~$300 in fees which I’m happy to pay for convenience and certainty.

Same with Shopify that powers e-commerce. Toast for restaurants. Faire for retailers with brands. Payments are almost incidental — they're just what happens at the end of the workflow. But they're enormously valuable precisely because of that.

Platforms embedding payments create a second lock-in on top of the software itself. They capture the operational data and the transaction economics. Powerful combination.

This is where the last decade of fintech was built.

Stripe sits here. Adyen too. They abstract the complexity, handle compliance, and take a margin for doing so. Together they process $4T a year — roughly 3% of global GDP — powering payments for platforms like Anchor.

But at scale, they get expensive. A business processing $100M a year pays Stripe close to $3M in fees. That gap gave rise to a new layer of players — Rainforest, Finix — who let large businesses process payments at a fraction of the cost, without building their own infrastructure.

Payments was just the beginning. Stripe has since expanded into treasury and capital. Unit and Column built the middleware that let non-banks offer the full suite of banking products without a banking license.

The banking-as-a-service momentum was real. Then Synapse collapsed in 2024, and the fragility of the model was exposed overnight.

Layer 2 is where the real regulatory weight sits. These are licensed, supervised institutions.

They move slowly by design. That's a feature, not a bug. You want your bank boring and your settlement rails predictable.

But they also carry the deepest moats and the slowest pace of change. Meaningful improvement in how money moves requires working with these institutions. Not around them.

This is where Visa and Mastercard belong.

Contrary to common belief, they don't move funds — they route instructions and set the rules. And despite their brand dominance, card rails account for less than 15% of US B2B payment value. The bulk moves on:

The rails are moving.

The rails are the most durable layer of the stack. Card networks have been compounding for 60 years. But the cracks — cost, speed, opacity — are exactly where the next generation is building.

Thirty years ago, accepting payments required a bank relationship, a merchant account, and a payment gateway.

Stripe changed that with seven lines of code. Square (now Block) put a card reader in the hands of a food truck owner. Shopify gave a generation of brands a storefront and a checkout.

The results are real. Hundreds of millions of people and businesses transact faster, cheaper, and more reliably than they did a decade ago.

But the problems that remain are equally real.

It still costs $3 to move $100. Wire transfers settle in three days. $33B is lost to payment fraud annually. Nearly $7T is stuck in AR. Credit costs US businesses $3T in annual interest.

Small and medium businesses carry most of that burden.

The financial system is the most heavily regulated industry. For good reason.

The line between innovation and recklessness is thin — and the industry has crossed it more than once. Wirecard. Celsius. FTX. SVB. Synapse. Each one a reminder that moving fast in financial infrastructure has consequences.

The companies that have endured — Stripe, Block, Circle, Coinbase — learned this early. They worked with regulators, partnered with legacy institutions, and built compliance into the foundation rather than bolting it on at the end.

Software along isn’t enough to move the financial system forward.

That’s it for this week.

Please help me spread the word and forward this to your payments friends!

And let’s connect on LinkedIn in the meantime.

See you next Thursday,

The Payments CFO