Cash That Moves, Cash That Earns: Onchain for Treasurers

Speakers

Stablecoins for Cross-Border Payments and Treasury: What Actually Works Today?

• Fabricio Bezerra, Chief Digital Officer at ArcelorMittal Treasury
• Ioannis Kympritis, VP of Global Expansion at Hercle
• Gonçalo Correia, Institutional Lead EMEA at Utila
• Antoine Hello, Director of Business Development at Zama
• Maxime Coniglio, The Big Whale (moderator)

Corporate treasurers are arriving, with an ROI mandate rather than a crypto thesis

The composition of the room was itself the first signal. Of the 100-plus registrants from 66 organisations, 6% were corporates from outside the financial sector, a category that would barely have been represented a year ago. Maxime Coniglio framed it as the start of a third wave of adoption:

• First wave: speculation, driven by retail
• Second wave, still under way: financial institutions entering digital assets
• Third wave, now starting: multinationals and SMEs outside finance, with SMEs moving faster because they are less risk-averse and feel the same pain points as large groups (T+2 settlement, cut-off times, cross-border bank fees, heavy admin to open brokerage accounts)

ArcelorMittal illustrates how this wave starts. The group first looked at stablecoins in 2022 and stopped, because the available platforms did not meet its security bar. Its CFO relaunched the work last year with an explicit condition, as Fabricio Bezerra recounted: not because it is fashionable, but only for use cases with a clear ROI. The use case retained is international intercompany transfers. On technology, he has no reservations: whether a treasury self-manages wallets or delegates on- and off-ramping, “the solutions are there.” The hard part is internal: aligning compliance, cyber, internal control and, for US-listed groups, SOX, and explaining to a largely risk-averse team that nobody is moving 100% of the treasury onchain. In his words, it is change management.

The stablecoin sandwich: minutes instead of days, and basis points that add up

For intercompany settlements and cross-border payments, the application that reaches market fastest is the stablecoin sandwich: the correspondent banking chain is replaced by a blockchain settlement leg in stablecoins, connected at both ends to local real-time payment systems. Ioannis Kympritis quantified the two gains on the Europe–Nigeria corridor:

• Latency: around 13 minutes end to end, operational steps included, against two to three days for a SWIFT transfer over correspondent banking, when the payment is not blocked outright because the corridor is classified high-risk
• Cost: a Hercle client converting euros into naira had been executing at an average premium of 50 basis points over the interbank FX rate, before any local bank fees. On a €50 million transfer, that is €250,000 saved on a single transaction
• The caveat: the gain is corridor-dependent. It justifies specific flows, not a wholesale migration of treasury infrastructure

Security is no longer the bottleneck; governance and procurement are

Utila, which serves over 350 clients and processes more than $30 billion a month, answered the recurring fear of being hacked with its operating model: keys are split into MPC shares held by the client and by Utila, both required to sign, with a policy engine, approvals and quorums that can be set per team, wallet or vault. Gonçalo Correia used the $1.5 billion Bybit hack of 2025 to reframe the question. The keys were never stolen; the signers were manipulated into approving a transaction that looked normal. The question is no longer only “who holds your keys?” but “who can make your keys sign?”, which is a governance problem addressed by policies and quorums.

The real friction sits elsewhere:

• Procurement processes are designed for ten-year core banking contracts, not for innovation contracts
• Pilots prove the gain quickly (three days reduced to minutes), then stall because the decision is split across departments
• Fabricio Bezerra described a single custody contract circulating through legal, the DPO, cyber and ESG, each adding comments on something new, sometimes needing external consultants. One bank received ArcelorMittal's comments and went silent for seven months, “not ready yet”

Privacy on public chains is the missing piece for institutions

Transactions on a public blockchain are public by default: counterparties and amounts can eventually be reconciled with an address's history. Antoine Hello called this the elephant in the room, one the industry has worked around for two years with off-chain privacy. Zama's approach keeps the transaction on the public chain:

• Balances and amounts stay encrypted on Ethereum; a block explorer shows no figures
• The application (a custody platform such as Utila, for example) defines who is entitled to decrypt, so compliance can be set per jurisdiction
• Institutions get access to the liquidity and users of public chains without exposing their strategy or their counterparties
• On the yield side, Zama has launched the first confidential vault on Morpho with Steakhouse, while SG-FORGE runs its own vault with EURCV

Move cash with stablecoins, don't hold them: accounting and interoperability set the limits

ArcelorMittal does not want to hold stablecoins. It uses them as a transfer mechanism: in from fiat, across, and back out. The live example: cash collected late in the US, Mexico or Canada is worth more at group level than in a subsidiary, so the entity buys stablecoins, sends them to Europe (Paris, Luxembourg), where the cash can go into a tokenized money market fund, which counts as a cash equivalent. Stablecoins held in Europe would not. The Q&A confirmed where the limits are:

• Accounting: a TMS provider in the room noted that no accounting rule yet lets a treasurer carry stablecoins as a cash equivalent, which keeps corporates inside the sandwich. Deposit tokens, by contrast, can be booked in a TMS like any deposit
• Bank fees on the fiat legs: banks still charge the fiat ends of the sandwich, and some charge the maximum
• Depth: with USDC at around $70 billion, the same speaker argued that large TMS clients move comparable amounts daily. Gonçalo Correia's answer: liquidity follows volume, and providers will come once flows move onchain
• Interoperability: a treasurer wants to “send 10 million to this counterparty,” not choose between Solana and Ethereum and discover the recipient can only receive on the other chain. That layer is not there yet
• Yield: ArcelorMittal considers it premature. Utila reports a sharp rise in demand for tokenized money market funds and T-bills over the past six to nine months, including a payments company wanting to park funds six to eight hours a day with hourly yield and instant redemption

Integration is less of an issue than it looks: Utila connects to TMS and core banking systems, and Gonçalo Correia described reconciliation between the bank ledger and the onchain ledger as “a very solvable” technology problem. Longer term, Ioannis Kympritis expects the category to change nature within five to ten years, with stablecoins and tokenized deposits treated as currency on a unified ledger rather than as a routing instrument.

Banks are both the slowest link and the indispensable bundler

Every panellist came back to banks. Antoine Hello expects intra-group transactions to stay the first use case, with banks bringing treasurers onchain within 12 to 18 months, since they are the natural intermediary for these flows. Ioannis Kympritis sees them consolidating the services of providers like Hercle, Utila and Zama into one offering; intra-company treasury balancing is already live at scale among money transfer operators moving funds between European entities and emerging markets.

• ArcelorMittal's CFO wants a core bank as custodian, “to be on the safe side”; at EuroFinance the week before, Fabricio Bezerra found most banks not there yet
• Banks are moving beyond basic custody into stablecoins, payments and tokenization, under peer pressure comparable to what Revolut put on retail banking. Utila cited a well-known company that started with 10 to 30% of its treasury, saw 30 to 40% optimization, and immediately asked how fast it could scale
• Resistance depends on the business model: transactional banks are already adopting stablecoins and monetizing the flows, while yield-driven banks have no interest in faster settlement. Ioannis Kympritis's view is that tokenized money market funds and deposits let them recapture that liquidity, and that the banks positioned early will benefit most
• Adoption is faster where FX access is constrained or inflation is high. In Europe, a cash manager's reflex is that SWIFT already works and the TMS is stable. Fabricio Bezerra's suggestion: banks offering stablecoin transfers to retail clients, such as remittances through the banking app, would normalise the technology faster than any corporate pilot

Conclusion

Paris showed how far the stablecoin treasury debate has moved from whether to how. The stack is assembled: the sandwich delivers minutes instead of days and measurable FX savings on the right corridors, MPC custody with policy engines answers the security question, and confidential transfers on public chains are arriving. What holds corporates back is organisational (education, procurement built for decade-long contracts, custody agreements that take months to clear) and structural (no cash-equivalent accounting treatment for stablecoins in Europe, no interoperability between chains at the treasurer's level). ArcelorMittal's posture, using stablecoins to move cash rather than to hold it, is the realistic stance for a large European corporate in 2026. The pace will be set by banks: those that bundle custody, liquidity and privacy into a single offering will bring treasurers onchain, while transactional banks already monetizing stablecoin flows show the others what they risk leaving on the table. Fabricio Bezerra's advice to the room: keep building the railroad, because the train has already left.