On any given day, the world's money transmitters keep billions of dollars sitting idle in correspondent bank accounts around the world. The cash is not earning. It is not moving. It is collateral for the next outbound payment to Manila, Lagos, or Karachi. Arf, a six-year-old fintech headquartered in Zug, Switzerland, thinks that model is ready to be retired.
The company sells a liquidity product to licensed financial institutions: short-term, revolving credit lines denominated in USDC that fund cross-border payments without requiring the sender to prefund a destination account [Crunchbase]. The credit lines are unsecured and settle on-chain. Instead of wiring dollars to a nostro account in advance and waiting days to recycle them, the institution borrows from Arf for the duration of the transfer, settles, and repays. The pitch is that working capital stops being dead weight.
The bet
Arf's wedge is the money transmitter and the small-to-mid-tier bank that moves real volume but cannot economically tie up cash in a dozen jurisdictions. The company says its rails support same-day settlement across more than 800 money transfer locations in Europe [Arf]. A strategic partnership with LuLu Financial Holdings, one of the larger remittance networks in the Gulf and South Asia corridors, is built around T-0 settlement [Arf].
The product sits on Stellar, and the working capital is USDC. It is a credit underwriter for regulated payment institutions, using a stablecoin as the settlement asset rather than a correspondent bank.
Why it could be big
Arf raised $13 million in a seed round backed by Circle Ventures, Hard Yaka, and the Stellar Development Foundation [The Payments Association]. Blockchain Founders Fund, Signum Capital, and United Overseas Bank Venture Management also participated [EIN Presswire][Dealroom.co].
| Metric | Value |
|---|---|
| Seed round total | 13 $M |
| MTO locations served (Europe) | 800 count |
The correspondent banking model that Arf is competing against has been shrinking for a decade. The competitive set, including BVNK, Conduit, and Bridge, suggests investors have decided this category exists [CB Insights]. Bridge alone was acquired by Stripe in 2024 (widely reported), which reset expectations for what a stablecoin-native payments infrastructure company can be worth.
The team and the merger
Arf was founded in 2019 by Ali Erhat Nalbant. The company has since announced a merger with Huma Finance, an on-chain credit protocol, according to StartupTicker. The combined entity pairs Arf's regulated liquidity business with Huma's tokenized credit infrastructure. Arf has also collected the PAY360 award for innovation in global payments three years running [PR Newswire].
What bears say, and what bulls answer
The credible bear case is regulatory. Selling unsecured credit lines denominated in a stablecoin to licensed financial institutions is a product that lives at the intersection of three regulators in every jurisdiction it touches. BVNK and its peers are racing through the same gauntlet [CB Insights]. The bull answer, visible in the cap table, is that Arf has chosen to base itself in Switzerland, has Circle and SDF as strategic backers, and is selling to counterparties who are themselves licensed. The merger with Huma also suggests the company is building toward a capital structure that does not depend solely on equity to scale the loan book.
What to watch
The next twelve months will turn on three things. First, whether the Huma combination produces a visible step-up in originated volume. Second, whether Arf names a second anchor customer in the Gulf or Southeast Asia at the scale of the LuLu partnership. Third, whether a Series A materializes, and at what valuation, given the comp set Stripe established with Bridge.
Correspondent banking is not going away tomorrow. But if a licensed money transmitter can borrow USDC for forty-eight hours, settle a payment to a counterparty in another time zone, and repay the line at a cost lower than the opportunity cost of prefunded cash, the math gets uncomfortable for the incumbent model very quickly.