Stripe confirmed on August 19 that it would acquire OpenRouter, a startup that routes AI requests to whichever model is cheapest or best for the job, concluding the deal FinanceFeeds covered earlier this week. Beyond the headline, the real story is that Stripe was buying something more specific than an AI asset. It was buying a routing layer.
A year earlier, Stripe had acquired Bridge, a stablecoin infrastructure company. Put the two together and a pattern emerges: the payments giant is assembling both the new financial rails and the intelligence needed to navigate them.
That combination is the real story in cross-border payments right now, and it is not the one the headlines usually tell. The common framing pits stablecoins against traditional banking rails in a fight one side eventually wins. The people building payment infrastructure describe something different, a future where both coexist, and where the hard problem, and the value, is deciding how money should move between them.
Stablecoins Became Infrastructure While No One Was Looking
The raw numbers around stablecoins are enormous and misleading. Reported stablecoin transactions run as high as $35 trillion a year, but most of that is trading, internal transfers, and automated on-chain activity, not payments for real goods and services. Strip those out, and a joint analysis by McKinsey and Artemis Analytics found that actual stablecoin payments in 2025 totaled $390 billion, a far smaller figure, and still just 0.02% of global payment volume.
What matters is the shape of that $390 billion, not only its size. Business-to-business payments made up the largest share by a wide margin.
Real-world stablecoin payments in 2025 were led by business-to-business flows at $226 billion, far ahead of consumer categories. Source: McKinsey & Artemis Analytics · Chart: FinanceFeedsB2B stablecoin payments reached roughly $226 billion, about 60% of all real-world stablecoin payment volume, and by some estimates that figure grew several times over year on year. The use case is unglamorous and exactly the kind that sticks: replacing slow, expensive correspondent-banking wires for international suppliers and invoices with near-instant settlement, and giving companies a faster way to manage cross-border treasury. Blockchain-based transfers can cut all-in transaction costs to a fraction of a percent, against the 2% to 7% that traditional wires can cost once fees, FX spreads, and intermediary charges are counted.
The pool of value those payments draw on has grown just as quickly. The total stablecoin market capitalization now sits at about $301 billion, dominated by Tether’s USDT and Circle’s USDC.
The stablecoin market has grown to roughly $301 billion, with USDT and USDC accounting for the large majority. Source: DefiLlamaInvestor Takeaway
B2B is where stablecoins have actually become infrastructure, at roughly $226 billion or about 60% of real-world volume, driven by near-instant cross-border settlement replacing correspondent-banking wires.
Not a Battle, a Coexistence
The instinct to frame this as stablecoins versus banks misreads where the industry is heading, according to Vadim Drozd, CEO of the payment orchestration platform FinteqHub, which routes transactions across both fiat and crypto rails for merchants operating internationally. In his account, the obstacle to stablecoin payments is not the technology but everything around it: regulation remains fragmented, banks treat stablecoin transactions inconsistently, and adoption is still far from universal. Building a payment operation around stablecoins alone, he argues, swaps one set of operational and regulatory risks for another rather than removing them.
The more likely future, in his view, is hybrid, stablecoins running alongside traditional currencies, banks, payment service providers, and local payment methods rather than replacing them. “The challenge will be deciding which rail to use, when and at what cost,” Drozd said. That reframes the whole problem. If every rail survives, the advantage goes to whoever routes across them best.
The Regulation Is Still Catching Up
Drozd’s point about fragmented rules is not abstract, and the past week made it concrete. On August 19, Comptroller of the Currency Jonathan Gould said the OCC now aims to finalize its GENIUS Act stablecoin rules by November, having missed the law’s July statutory deadline, so it can begin processing issuer applications early next year. Two days earlier, the Treasury Department proposed its own rule defining when a stablecoin counts as issued or sold in the United States and what licensing that triggers.
Those are steps toward clarity, but they also show how unsettled the ground still is more than a year after the GENIUS Act became law. Rules are arriving in pieces, across multiple agencies, on timelines that keep shifting, and the framework does not fully take effect until 2027. For a business trying to move money across borders today, that patchwork is exactly the kind of complexity that makes a single, hardcoded payment path fragile and a flexible one valuable.
Where the Value Moves
If rails proliferate and rules stay uneven, the orchestration layer, the software that sits above payment providers and routes each transaction, becomes the place where cost and reliability are won or lost. Drozd puts the potential savings high: intelligently routing transactions between fiat and stablecoin rails, he said, “can potentially reduce transaction costs by 75% or more,” framing that as his firm’s own estimate rather than an industry benchmark. Even discounted, the direction is clearly that the routing decision is where the money is.
Today, that decision is mostly made by hand. Routing still tends to rely on static rules that payment teams maintain manually, even as provider fees, availability, limits, and performance shift constantly. The fix Drozd describes is to make routing dynamic, using machine learning to weigh those moving variables in real time and choose the best path for each transaction, an approach he says FinteqHub is already applying to its own transaction flows.
This is where Stripe’s two acquisitions stop looking like separate bets. Bridge gave Stripe stablecoin rails; OpenRouter gives it a proven engine for routing requests intelligently across many options. Stripe’s own description of the OpenRouter deal, evaluating each request and sending it to the optimal destination based on price, speed, and reliability, reads almost exactly like the payment-routing problem Drozd describes, applied to AI models instead of payment rails.
Both moves point at the same destination: infrastructure that does not just carry a transaction but decides, intelligently, how to carry it. PitchBook has pegged stablecoins and agentic payments as Stripe’s two biggest priorities in emerging technology, and both run straight through that routing layer.
Investor Takeaway
As rails multiply and regulation stays uneven, the durable advantage shifts from owning any single rail to routing intelligently across all of them, which is the layer where cost and reliability are actually won.
What Comes Next For Stablecoins
The near-term trajectory, in Drozd’s view, is that the routing work stops being manual. He expects algorithms to take over much of the routine routing over the next few years, with payment teams shifting from maintaining rules to higher-level decisions about which partners, markets, and traffic to prioritize. It is the same shift Stripe is buying its way into at a larger scale.
None of this requires stablecoins to displace anything. It requires them to become one more rail among many, which the payments data suggests is already happening in B2B flows. The competitive question that follows is not which rail wins, but who builds the intelligence to move money across all of them. “The real opportunity lies in building the intelligence that decides how money should move between them,” Drozd said. Stripe, with a stablecoin business and a routing engine now under one roof, appears to be betting on exactly that.
