Ant International announced an AI-native product stack for global businesses spanning payments, accounts, foreign exchange, treasury, and growth operations. In a September 18 press release distributed through PR Newswire (news provided by Ant International), the company called it its largest-ever product upgrade, with close to 100 key products across four business pillars. Its industry-first description is Ant International’s claim, not an independently established finding.
Ant International says it is combining payment, account, FX, treasury, and growth functions under a FinAI stack. Its stated foundation is a two-layered security architecture and two proprietary models: the payment-focused Antom 3-in-1 Transformer and the FX and liquidity forecasting model FalconTST. Neither source provides an independent audit of the security design or model performance.
A separate report from FX News Group recorded the launch, two-model design, and interoperability components. Its account does not independently validate Ant International’s usage, accuracy, or cost claims. Peng Yang, Ant International’s CEO, told more than 1,100 executives at its VOYAGE merchant event in Shanghai that a future fintech leader needs “a future-ready trust infrastructure” alongside foundational FinAI models and connected AI-native solutions.
Antom Transformer Combines Three Payment Data Types
Ant International says its Antom 3-in-1 Transformer processes sequential, tabular, and graph data in one payment model. The company says it has more than 10 billion parameters and processes 90 trillion data points annually. Its cross-modality attention architecture is intended for fraud detection, scam prevention, abuse prevention, and payment-success-rate optimization. These are stated use cases, not externally tested outcomes, and the release provides no independent benchmark.
The company also said chargeback rates among leading large-language-model clients of a newly launched subscription service fell by as much as 87%. The release names neither clients nor a baseline rate and provides no third-party measurement. It is a company-reported maximum result, not evidence that every Antom merchant will experience that reduction.
Ant International also reported that 89.5% of clients of its main merchant payment service deployed FinAI solutions in the preceding 12 months and that AI helped complete 81.4% of payment tasks. It separately put AI-agent deployment among Antom merchants at 89.5%. Neither source supplies an external methodology, customer list, or independent verification. The two 89.5% figures are company-reported deployment rates, while 81.4% is the company-reported share of payment tasks completed with AI help; neither establishes an industrywide adoption rate.
FalconTST Extends AI Claims Into FX and Cash Flow
For foreign exchange and liquidity, Ant International is relying on FalconTST, a time-series forecasting model trained on data from finance, retail, energy, travel, and other economic sectors. The company says FalconTST has more than 8.5 billion parameters and has reached more than 93% accuracy in FX and cash-flow forecasting. It also says real-world cases reduced corporate FX hedging and allocation costs by 30% to 60%. The release does not define the comparison set, the individual cases, or the forecasting methodology behind those figures, so they remain company claims rather than vendor-neutral benchmarks.
The rollout links these forecasting assertions to enterprise treasury software. Ant International named WorldFirst for Enterprise Treasury Autopilot among the products in the set, alongside the conversational Antom Autopilot for the merchant payment lifecycle, the SME natural-language agent Wyn, and WorldFirst for Enterprise Account for Agent, or AFA. The product names clarify the company’s aim to use AI across both merchant-payment and treasury workflows.
(Related: China’s Tech Giants Race to Embed AI Agents Into Enterprise Workflows)
AMP and WhaleRTP Provide the Stated Interoperability Layer
Ant International says the FinAI stack is underpinned by the Alipay+ Agentic Mobile Protocol, or AMP, and WhaleRTP, a blockchain-based wholesale settlement platform. The company describes AMP as an open payment protocol for digital wallets, super apps, smart devices, and other mobile interfaces, designed to let AI agents make payments. Since AMP’s April 2026 launch, Ant International says 10 Alipay+ wallet partners and seven acquirers have supported its Phase 1. It further says AMP can cut agent-to-wallet integration steps by 50% and enable agent-to-agent settlements as small as $0.000001; those are company estimates and capabilities, not independently verified results.
In this release, Ant International lists both AMP and WhaleRTP as underpinnings of the FinAI stack. Ant International says WhaleRTP interoperates across blockchain assets from 23 financial institutions. It also says the platform processed 45% of its cross-border volume in 2025, reduced working-capital requirements by 60%, and raised interest income by 23%. Ant International supplied those historical figures, but neither its press release nor the FX News Group report provides an audit or financial-institution-level breakdown.
(Related: Ant Opens a Payment Protocol for AI Agents Across Asia)
Global Rollout Plans Rest on Ant International’s Reported Network
The scale Ant International cites for its network is substantial but should be read as company-provided network data. It says its network covers 53 digital wallets and 10 national QR schemes, connects 150 million global merchants with more than 2 billion consumer accounts, and processes over 25 million average daily transactions across 210 markets. These figures describe the company’s reported network; they do not demonstrate the number of customers using each newly announced product or the commercial effect of the AI functions.
Ant International says the products are scheduled to roll out in global markets during fall and winter 2026. That is a stated rollout plan, not confirmation of broad availability, customer uptake, regulatory approval, or revenue impact. The announcement instead establishes a defined product agenda: payment data through Antom 3-in-1 Transformer, forecasting through FalconTST, agent payments through AMP, and settlement through WhaleRTP. The more consequential measures of the launch, whether its accuracy, savings, and adoption claims persist beyond the company’s release, are not answered by the two reports and will require evidence beyond this announcement.