top of page

Payhawk Series C Lands $280m to Build an AI Finance OS

3 hours ago
4 min read
Payhawk Series C Lands $280m to Build an AI Finance OS

Payhawk has raised a $280m Series C led by Growth Equity at Goldman Sachs Alternatives, an all-equity round that values the London-headquartered spend management company at $2bn and takes its total funding past $515m. The Payhawk Series C, announced on 18 November 2026 and still subject to standard regulatory approvals, is earmarked for a single objective: turning an expense and card platform into what the company calls a finance operating system for the AI era.


The raise roughly doubles the $1bn valuation Payhawk reached in March 2022, when a $100m extension took its Series B to $215m and made the Sofia-founded business the first Bulgarian unicorn. Arriving at $2bn around four and a half years later places Payhawk among the higher-valued European finance-software companies, though still well below the $44bn that US rival Ramp commanded at its July 2026 round.


Every existing backer joined, including QED, Greenoaks, Lightspeed, Hedosophia and Bek Ventures. New investors Galdana Ventures, Cogito Capital, Inveready and Invus also came in, alongside the entrepreneurs Tomas Okmanas and Didac Lee. James Hayward, a managing director within Growth Equity at Goldman Sachs Alternatives, is joining Payhawk's board. That the round is all equity, with no debt facility attached, is itself notable in a sector where later-stage fintechs increasingly lean on credit lines to fund growth.


What will Payhawk build with the new capital?


The money is aimed at a three-layer system Payhawk argues is more than a collection of sharper tools. The first layer is AI agents that perform financial work; the second is payments infrastructure that lets those agents move money; and the third is enterprise controls that define what an agent may do and at which point a person has to make the decision. Payhawk says 77% of its customers already use its agents every day across financial control, procurement, travel and reporting.


Co-founder and chief executive Hristo Borisov frames the shift as a move away from manual checks and reconciliations and toward what he calls zero-touch finance, in which neither the finance professional nor the employee handles administrative work. He argues that AI on its own cannot deliver that change, and has to be paired with payment infrastructure and enterprise controls to be useful inside a large organisation.


How does a finance operating system differ from smarter finance tools?


Borisov's case rests on a distinction between connected systems and a changed operating model. ERP platforms, banks, procurement tools and treasury products are all becoming more capable as they adopt Model Context Protocol servers, which make their data easier for AI to reach. Payhawk's position is that making each application individually smarter does not reorganise the finance function; an operating system, it says, changes how the whole function works. That difference grows more consequential as software moves from recommending an action to carrying it out.


The practical expression of that argument is the division of labour between the three layers, with the control layer deciding when an agent proceeds autonomously and when it must hand a decision back to a human. It is the governance layer, rather than the agents themselves, that Payhawk is betting enterprises will pay for.


Why does Payhawk's independence from card interchange matter?


Most corporate-card providers earn a large share of revenue from interchange, the fee charged each time a card is used, which ties their growth to how much customers spend. Payhawk instead prices on subscription, which it says aligns its incentives with helping customers spend less rather than more. The company reports that it turned cash-flow positive in the third quarter of 2026 and still holds 70% of the capital it raised in its 2022 round, leaving this Series C as expansion capital rather than runway.


That independence also shapes the product. In the US, Payhawk's Financial Controller Agent works with a customer's existing business cards, so finance teams can adopt it without switching providers or reissuing cards. It is a pointed contrast with the interchange-funded model that US competitors built their businesses on.


How does Payhawk compare with Ramp, Brex and other rivals?


The raise lands in a market that is expanding and consolidating at the same time. Ramp reached a $44bn valuation in 2026 on an interchange-led, free-card model; Brex was acquired by Capital One for about $5.15bn; and European challengers Spendesk, Pleo and Qonto compete on local compliance and multi-entity controls, with Spendesk reporting that it reached profitability during the year. Across the category, vendors spent 2026 wiring AI agents into the ledger, moving the contest from expense capture toward automated financial operations.


Payhawk's pitch for that contest is its European footprint, operating across 33 countries from 12 offices, which exposes it to multiple currencies, banking systems, payment rails and regulations. The company frames that complexity as a durable advantage: infrastructure built to handle fragmented European finance from the outset is, on its account, harder for a US-first rival to replicate at speed.


Why This Matters to FinanceX Readers


For finance leaders, the signal in this round is less about the valuation than about where agentic AI is heading inside the finance stack. The competitive question is shifting from which platform captures receipts fastest to which one can be trusted to execute payments under enforceable controls. Payhawk's subscription model and cash-flow-positive position give it a different risk profile from interchange-dependent rivals, which matters for any CFO weighing a multi-year commitment to a finance platform.


The open question for readers to watch is whether the operating system framing proves to be a genuine change in how finance runs, or a well-funded repositioning of spend management, a verdict that will rest on how much real decision-making enterprises are willing to hand to agents.

 
 
bottom of page