Money's switchboard: why VCs are pouring money into payment orchestration

Hello friends, and welcome to The Unsophisticated Investor! Brought to you by Scott & Rob, the founders of Shuttle!

If you want to invest alongside the VC funds who've backed breakout companies like Revolut, Asana, JustEat, Bolt, Lets Get Checked, Loom, Runna, Charlotte Tilbury, Deel, Aircall, AngelList, Carta, TransferWise and many more, regardless of your knowledge, network or net worth, then sign up using the link below.

Now, let’s get into it 👇

Primer's $100 million Series C last week is the most visible sign of something the rest of the payments industry has spent two years quietly working towards. A look at payment orchestration, the technology underneath it, and a category with roughly a dozen serious competitors and no clear winner yet.

Last week, London-based Primer announced a $100 million Series C, oversubscribed, co-led by Belgium's Sofina and Peak XV Partners (formerly Sequoia Capital India), with Tencent (rare on a European cap table) joining alongside Balderton, Accel, ICONIQ and Speedinvest. The company, founded in 2020 by Gabriel Le Roux (ex-Braintree) and Paul Anthony (ex-PayPal), has now raised $170 million in total.

What payment orchestration actually does

Orchestration is, in essence, load balancing for money. Five years ago, a mid-sized European e-commerce merchant would integrate with a single payment service provider, most commonly Stripe, Adyen or Checkout.com, which handled card networks, fraud screening, currency conversion and regulatory plumbing in each jurisdiction. If approval rates dropped in a country or for a card type, the merchant lived with it or rebuilt the stack. Adding a second provider was a multi-quarter engineering project.

An orchestration layer sits between the merchant's checkout and that ecosystem of providers. The merchant integrates once, via a single API. Underneath, the orchestration platform maintains connections to dozens of PSPs, acquirers, alternative payment methods (Klarna, iDEAL, Pix, UPI, and so on) and risk and fraud tools. When a transaction arrives, the platform decides in real time which provider to route it to. If that provider declines, the platform retries elsewhere within milliseconds. If approval rates in Brazil drop with one acquirer, traffic shifts to another. If a customer pays with an Italian card, the transaction routes to the acquirer with the strongest Italian approval performance.

The economics are straightforward. Approval rate improvements of one to three percentage points, on businesses processing billions in annual volume, translate into nine-figure revenue uplifts. Dynamic routing to the lowest-cost provider that will still approve a transaction typically delivers a further 10 to 30 basis points of margin. The orchestration platform takes a share; the merchant keeps the rest.

The category leaders capture a great deal of data in the process. Primer publicly references customers including GetYourGuide, Dialpad, Rail Europe and Printful, and states that its platform captures more than 400 data points per transaction and manages more than 95% of customer payment volume on average. Berlin-based Payrails publicly works with Puma, Vinted and Careem. Yuno counts McDonald's and Avianca as customers. The dataset built up by these platforms is the foundation for what comes next.

The AI overlay

Primer framed the raise around its AI roadmap rather than its payments infrastructure. Its AI agent, Companion, launched in 2025 alongside Global Accounts and currently surfaces contextual insights and answers complex queries about a merchant's payment performance. The capital is intended to evolve Companion into something more autonomous: running experiments, optimising routing logic, and executing operational decisions inside parameters the merchant has set. The round will also fund US expansion.

The reason payments are an unusually good fit for agentic AI is the clean outcome signal on every transaction. Was the payment approved? At what cost? Did the customer convert? Each transaction is a labelled training data point. A routing experiment run on Monday morning produces results by Tuesday. Most enterprise software workflows have outcomes that are slow, ambiguous and hard to instrument. This is why payments is the natural first surface for production-grade AI agents in financial services.

The harder problem is data. AI agents are only as useful as the data they run on, and most enterprise payment data lives in fragments: different formats, different schemas, different reconciliation cycles across each provider. The orchestration layer is, almost incidentally, the place where this data gets unified. As Peak XV's Aakash Kapoor put it in the press release, "As payments enter a new architectural era, that depth of context becomes critical for AI agents to make decisions." The investor framing is that the company solving orchestration also ends up owning the dataset on which financial AI agents will eventually be trained. That is the version of the thesis that, plausibly, persuaded Sofina and Peak XV to lead the round at the price they did.

The scale of the opportunity

Mordor Intelligence puts the global payment orchestration market at $2.65 billion in 2025, growing to $7.27 billion by 2031. SkyQuest projects $9.34 billion by 2033. Global Growth Insights, the most aggressive, models $44 billion by 2035. The ranges are a clear indication that the category is still finding its size.

The more useful framing is the size of what sits beneath orchestration. McKinsey's global payments revenue forecast is in the region of $2.5 trillion. Global e-commerce is expected to reach somewhere between $6 trillion and $8 trillion by 2027. Between 60 and 80% of large merchants now operate multi-PSP setups, up from a small minority five years ago.

The orchestration layer is the thin slice on top of all this, a small percentage of a very large number, and the part of the stack that gets to charge for being clever rather than for commodity processing.

The structural tailwinds keep arriving. Cross-border B2B payment value is projected to grow 40% by 2028. Stablecoins moved $32 trillion in 2024 and are forecast to take up to 20% of cross-border flows. Each creates new payment methods someone has to orchestrate. The category is structurally short of solved problems.

The competitive map

Three layers of competition are worth distinguishing.

First, the platform giants. Stripe, Adyen, Checkout.com, Worldline and ACI Worldwide increasingly offer orchestration features as part of their core products. The argument they make to merchants is straightforward: a business already integrated with one of them does not need a second integration to add routing logic. This is the most serious structural threat to standalone orchestration, because vendor neutrality is harder to defend when your largest provider also runs the routing layer.

Second, the dedicated challengers, where most of the venture money is concentrated. Primer is the most visible after last week's round. Gr4vy, founded in 2020 by ex-PayPal executive John Lunn, has raised $27 million. Berlin-based Payrails, founded by three ex-Delivery Hero engineers in 2021, has raised $52 million total, with Andreessen Horowitz and EQT Ventures joining HV Capital in a $32 million Series A this June. Yuno, the LATAM-origin platform, has expanded into the Middle East. Spreedly, the Durham-based incumbent founded in 2008, supports more than 140 gateways. After that comes a long tail, APEXX Global, IXOPAY, BridgerPay, CellPoint Digital, Juspay, and others.

Third, consolidation. Mangopay acquired the Irish no-code orchestration platform WhenThen in 2025. More deals of this shape are likely as larger payments groups try to defend against the dedicated challengers by acquiring capability rather than building it internally.

The honest reading is that roughly a dozen serious competitors are chasing the same problem with broadly similar technology. The eventual winners will be decided by some combination of distribution, AI execution, and which of the incumbents decides to acquire whom.

Why Europe is unusually well represented

For a sector where US capital and US scale usually dominate, the orchestration cohort is heavily European. 

The structural reason is that Europe is the natural market for orchestration. The US landscape is comparatively concentrated: Visa, Mastercard, a handful of dominant acquirers, and a thinner stack of alternative payment methods. Europe is the opposite. More than thirty local payment methods are in active use (iDEAL, Klarna, SEPA, BLIK, and so on), twenty-seven EU regulatory regimes plus the UK and Switzerland, different 3-D Secure implementations across jurisdictions, and significant cross-border traffic on every large merchant's books. Operating at scale in Europe forces the orchestration problem on the merchant. American merchants can defer it for longer.

This is the same dynamic noted in our earlier edition on defensible fintech business models in the AI era: Europe's regulatory and structural complexity creates moats for the companies that build through it. The PSD2 regime, the PSD3 update arriving over the next eighteen months, the EU's instant payments mandate, and the gradual evolution of the digital euro all add complexity. Each layer is a reason for a merchant to want an orchestration partner who can absorb it on their behalf.

The risks worth flagging

Three headwinds are worth noting.

The first is the platform bundling threat already described. If Stripe, Adyen and Checkout.com decide that orchestration is a feature rather than a product, and price it accordingly, the standalone vendors lose pricing power on every renewal.

The second is that AI claims in payments are easy to make and hard for outsiders to verify. Primer Companion has been in production for roughly a year. Its actual performance against rules-based routing is not something an outside investor can audit. Across the category, the gap between marketing claims and production behaviour is wide, and the discipline of measuring agentic systems against the rules-based baselines they replace is still immature.

The third is that the category is overcrowded relative to the realistic size of the winner cohort. Two or three of these companies will end up being valuable. The rest will sell to a larger payments group, get acqui-hired, or quietly wind down. Picking the survivors is not obvious from the outside, and the people who claim it is should be treated with caution.

What this looks like from where we sit

The AI overlay is what moves this from a known fintech category into something that may attract a different kind of valuation multiple. Primer's Sofina-led round, with Peak XV alongside, reads less like a payments raise and more like a vote that the orchestration layer is becoming the financial AI dataset. That framing has consequences for who buys the eventual winners and at what price.

The dispersion of outcomes will be wide. The category-average return on a dozen $50–150 million bets, several ending with strategic acquirers, a few failing, is probably reasonable but unspectacular. The right two picks will be a different number entirely. The winners are not yet decided. The work of picking them is happening now, in private rounds, before the public market gets a look.

What we’ve been working on at Shuttle

• Final stages of our fundraise 💰

• Completed our FCA Supercharged sandbox application ✅

• Adding new features and improvements to our platform 🧑‍💻

The Unsophisticated Investor is brought to you by Scott & Rob, the founders of Shuttle. We’re both sick of private markets being a playground exclusive to the ultra-wealthy so we started a company to challenge the status-quo. Shuttle’s singular focus is to unlock private markets for Millennial and Gen Z tech professionals and help them build wealth through the highest performing private market opportunities.

Scott & Rob
Shuttle Co-Founders