Stripe Wants to Run Your Entire Money Stack

If you run a business and you take payments online, there's a good chance Stripe already touches your money. Maybe you signed up years ago because it was easier than dealing with your bank's merchant account paperwork. Maybe a developer set it up and you mostly see the dashboard when you're reconciling at the end of the month. Either way, the company you thought of as your payment processor has quietly become something much more ambitious — and a recent conversation between Stripe's President of Product and Business, Will Gaybrick, and a16z General Partner David George makes that ambition very legible.

The short version: Stripe wants to run your entire revenue operation. Taxes, banking, fraud detection, global expansion, and soon, transactions made on your behalf by AI agents you may not have hired yet. Whether that sounds exciting or alarming probably depends on how your last vendor relationship went.

From payment processor to financial operating system

Gaybrick's framing is direct. Stripe, he says in the a16z conversation, has "inverted" its value proposition — from a payments company with add-ons to a multi-product platform where everything orbits financial infrastructure. The product count has grown to somewhere between 25 and 30 named products, with hundreds of features underneath. The average AI company, according to the conversation, uses 11 different Stripe products.

That's a lot of surface area for one vendor. If you're a small business owner who learned the hard way what happens when a single supplier has too much leverage over your operation — when your POS goes down on a Saturday, or your payment processor puts your funds on hold during a dispute — the consolidation Stripe is describing cuts both ways. One dashboard, one integration, one relationship. Also one point of failure, one contract renewal, one company whose priorities may or may not align with yours in three years.

Gaybrick is clearly aware of this dynamic, even if the conversation doesn't dwell on it. His most grounded example of Stripe's expanding value is fraud detection. AI companies offering free trials discovered, painfully, that users were gaming the system — signing up repeatedly under different accounts to avoid paying. Gaybrick describes Stripe building a fraud-detection pipeline, using its network of transaction data to identify and block abusive accounts. He says ElevenLabs told Stripe it's now blocking 2,000 such accounts per day using these signals. That's real value. It's also Stripe positioning itself as indispensable infrastructure at every layer of the revenue stack, not just the payment moment.

The Minions question, and what it means for your headcount assumptions

Here's where it gets interesting for anyone who pays people to build or manage things.

Internally, Stripe built what they call "Stripe Minions" — AI coding agents that take a one-shot instruction and build software autonomously, then go through automated testing before a human reviews the result. Gaybrick told the a16z audience that these agents now account for roughly 7,000 pull requests per week and approximately 30% of all of Stripe's code changes — figures he stated in the interview and which have been reported by Investor and Founder Interview Transcripts.

To translate that out of software-world language: a pull request is a discrete chunk of new code being added to a product. Thirty percent of all new code at one of the most productive software companies in the world is being written by machines that don't take breaks, don't need onboarding, and don't have opinions about the sprint planning meeting.

Gaybrick's argument is that this productivity explosion is an opportunity to build more, not to lay people off. "The best way to optimize your cost structure is to grow more," he says. Stripe's internal tools — including an AI knowledge assistant called Kai — have made sellers 20% more productive, and Gaybrick's response to that number is not "we need fewer sellers" but "we need a lot more sellers."

That's a genuinely interesting philosophy, and I find myself believing it more for Stripe's situation than I would for most. A company with decades of unmet user requests, operating at global scale, can plausibly absorb every efficiency gain into new product territory. But it's worth sitting with what this means for the businesses Stripe serves. If you're a regional distributor who uses two or three pieces of software that each took a team months to build, you should know that timeline is compressing fast. The competitive landscape for your industry's software is about to get more crowded, not less. The upstart that couldn't afford to build what you needed last year can probably build it now.

Checkout pages are going away, and that's not a neutral fact for everyone

Gaybrick's most striking prediction in the conversation is also the most concrete: checkout pages will disappear.

Imagine you run a boutique clothing retailer with an e-commerce presence. Right now, your checkout page is a relationship moment — it's where you collect the email, offer the loyalty program sign-up, show the upsell, build the cart. It's friction, yes, but it's your friction. It's designed around your brand and your customer experience.

Stripe's vision is a world where payment credentials are stored with services like Stripe Link, and the purchase happens at the point of intent — on the product page, or even delegated to an AI agent the customer sent to shop on their behalf. Gaybrick says Stripe has already launched tools to enable exactly this kind of agent-driven purchasing.

The efficiency argument is real. Nobody actually enjoys checkout forms. But for merchants who have spent years optimizing that final step — tweaking button colors, A/B testing copy, building post-purchase email flows — the checkout page going away isn't just a UX improvement. It's a restructuring of where customer relationships get built. If an AI agent is doing the buying, it's not reading your brand story. It's comparing price and availability and moving on.

That's not a reason to oppose the change. It is a reason to think now about where your customer relationship actually lives, because the checkout page might not be load-bearing for much longer.

Stablecoins: genuinely useful plumbing, genuinely early

The stablecoin section of the conversation is where Gaybrick is most clearly speaking to a different audience than most Buzzrag readers. The pitch is straightforward: existing national payment systems are fast and cheap in their home countries — India's UPI handles the vast majority of its transactions at low cost, Brazil's Pix has grown rapidly — but there's no equivalent for moving money globally. Stablecoins, Gaybrick argues, solve a political coordination problem that fiat currency can't: they give the global economy a shared rail that doesn't require every country to agree on whose system wins.

The concrete example he offers is Felix (formerly Félix Pago), a remittance company built on stablecoins that has grown to between 5 and 10% of remittances on the US-Mexico corridor — the largest such corridor in the world — in just a few years. That's not theoretical. That's market share.

For most small businesses in the US, this is not an immediate operational issue. If you're processing domestic transactions, stablecoins are still an experiment, not infrastructure. But if you sell digital goods or services internationally — if you have customers in markets where wire transfers are expensive and slow and card acceptance is spotty — Gaybrick's argument that stablecoins let you reach roughly 150 countries where Stripe's traditional fiat coverage reaches roughly 60 is worth paying attention to.

Stripe has integrated stablecoins directly into Stripe Treasury, so you can hold a stablecoin balance the same way you'd hold a USD balance. The friction for you as a merchant is lower than dealing with crypto directly. Whether that's actually easier than it sounds will depend heavily on your accountant's comfort level and your own appetite for learning a new settlement workflow.

The token economy question nobody's fully answered yet

The conversation ends with a concept that's more future-facing than anything else: the "token economy." Gaybrick uses "token" to mean AI tokens — the units of computation that AI services sell — and argues that as more economic activity runs through AI agents, tokens and dollars will blur together. You'll pay for AI services the way you pay for compute or electricity: in small, metered transactions, possibly without any human logging in to confirm them.

His case for micropayments is more persuasive here than the old content-paywall arguments were. If an AI agent is doing a task that requires pulling data from 12 different services, nobody wants to set up 12 monthly subscriptions. The economic model has to change to fit the workflow. "Microtransactions will just be necessary for that economy to exist," Gaybrick says.

He's probably right about the direction. The pace and the specific mechanics are genuinely unknown, and Gaybrick says as much — this is a space where the foundational building blocks are still being assembled.

The honest version of this piece is that most of what Gaybrick describes is real infrastructure development, not vaporware. Stripe is building products that work. The questions worth asking aren't whether this future comes — it's coming — but who it serves most cleanly at each stage, and whether the small business owner writing a $4,000 check to Stripe every month has the same leverage in this relationship that Stripe's startup customers always assumed they had.

The vendors who promise to simplify your financial operations and the vendors who end up owning them are sometimes the same vendor. Worth knowing going in.

---

Dorothy "Dot" Williams covers small business and entrepreneurship for Buzzrag.