Prefund Split: USDC-versus-fiat prefunding plans for Visa Direct payout firms

I could not find a product that does exactly this: daily P50/P90/P99 prefunding forecasts for each corridor, with recommended top-ups and tracking of forecast misses, sold to treasury at remittance, payroll and B2B payout firms. The horizontal treasury systems (Kyriba, GTreasury, HighRadius, Trovata, Atlar, Nilus) forecast cash at the level of the entity, bank account or currency. They are not modeled around payout corridors, partner balances, failure rates or stablecoin float. Industry write-ups say outright that tuning buffers for each corridor needs analytics and forecasting that firms mostly build themselves or do in spreadsheets. The opening is a specialist, payments-native layer, and the Buffer Backtest is a strong way in. Quantifying how much capital could have been released at the same failure rate is a CFO-legible ROI number that horizontal tools do not produce. Atlar's starting price of about 3,000 euros per month gives a reasonable reference point for pricing. There are three main risks. First, substitutes: Ripple ODL, Tazapay's no-prefund payouts, Visa's stablecoin prefunding and Thunes stablecoin funding all reduce the size of the buffer problem. Second, Ripple owns GTreasury and could bundle corridor forecasting with its rails. Third, the biggest payout firms often have in-house data science teams. The dependency on Timestral is a concern. I could not find any public evidence that Mistral has released a time-series model by that name. Even if it exists, the moat is the corridor data model, the integrations with ledgers and bank feeds, and the calibration and backtesting methodology, not the foundation model. Open models such as Chronos make the model layer easy to copy. The product should be built to work with any model rather than depend on one.

• Kyriba

  An enterprise treasury management system with cash positioning, bank connectivity and forecasting. It is a general corporate tool and is not built for corridor-level prefunding at payout firms. It is the platform a large remittance company's treasury team would most likely already have.

• GTreasury (now owned by Ripple)

  A treasury platform covering liquidity management, cash forecasting, payments and netting. Ripple bought it in October 2025, so Ripple could combine treasury forecasting with its no-prefunding payment rails. That makes it the most dangerous strategic adjacency.

• Ripple Payments / On-Demand Liquidity

  A substitute rather than a forecaster. It moves value through XRP as a bridge so that firms do not have to prefund in destination markets at all, which competes directly with the idea of optimizing prefund buffers.

• Atlar

  An AI-native treasury platform (cash management, payments, forecasting, reconciliation) with payments-adjacent customers such as Mangopay and Zilch. Pricing starts at about 3,000 euros per month. Its forecasting is general cash forecasting across entities and currencies, not percentile-based buffers per corridor.

• Trovata

  Bank-connected cash visibility and AI cash forecasting for treasury teams. It is generic and not built for payment firms.

• Nilus

  A cash and liquidity tool for finance teams: visibility, forecasting, reconciliation. Not specific to payment firms. (This CB Insights page lists it alongside Atlar.)

• HighRadius Treasury

  Enterprise AI cash forecasting and positioning across multiple banks. Horizontal, not corridor-aware.

• Thunes

  A payout network that publishes guidance on corridor liquidity and offers stablecoin prefunding to its members. It could build forecasting into its own network instead of buying it.

• Cybrid

  Payment orchestration with prefunded liquidity pools and automated top-ups for stablecoin remittance builders. It handles top-ups but does not appear to offer probabilistic forecasting.

• Tazapay

  Sells no-prefunding, per-transaction stablecoin-funded payouts. This is another substitute that removes the need for buffers in the corridors it covers.

• Visa Direct stablecoin prefunding

  A Visa pilot announced in September 2025 that lets firms prefund Visa Direct payouts with stablecoins treated as cash. It lowers the cost of holding a buffer, which shrinks the value of optimizing it. (This is a crypto news site's report, not a Visa source.)

Checked 2026-10-16