APAC AI Fintech: 8 Companies Compared
A side-by-side analysis of 8 AI-native fintech companies in APAC — funding, AI differentiation, M&A readiness, and who is likely to acquire them.
The APAC fintech market generated $9.3 billion in deal activity across 763 transactions in 2025. Within that, AI-focused fintech deals pulled $7.2 billion in the first half of 2025 alone — a run rate that would nearly double the full-year 2024 figure. That gap between AI and non-AI fintech valuations is widening.
This page maps eight of the most significant AI-native fintech companies operating in Asia Pacific today: who they are, what their AI actually does, what they are worth, and who is likely to buy them. It is intended as a reference for investors, corporate development teams at financial institutions, and AI fintech founders who want to understand where they sit in the landscape.
The comparison is structured around dimensions that matter for M&A and strategic decisions — not just funding totals, but data moats, license stacks, exit pathways, and acquirer theses.
Why AI Fintech in APAC Is the Highest-Stakes M&A Vertical in 2026
Three structural forces are converging.
The unbanked population problem creates genuinely differentiated AI. Western credit AI is marginally better than FICO scores. APAC credit AI is categorically different — because for hundreds of millions of adults across Indonesia, Vietnam, the Philippines, and India, there are no bureau scores at all. Companies like ADVANCE.AI, Kredivo, and Perfios have spent a decade building credit models using telecom usage, e-commerce behavior, social signals, and behavioral biometrics. That proprietary data — and the models trained on it — cannot be replicated by an incumbent bank in two years or five.
Regulatory licensing has scarcity value. A full digital banking license in Indonesia, a payment institution license in Singapore, a virtual bank license in Hong Kong — these are not freely available. Any financial group trying to acquire a capability comes with the license if the target holds one, but change-of-control approval requirements mean licensing creates friction and scarcity. Companies with clean, transferable regulatory stacks are worth more.
Japanese and Korean financial groups are the most aggressive acquirers in the region — and they are not done. Mizuho’s acquisition of Upsider in 2025 for approximately 46 billion yen was not an isolated event. It reflects a multi-year thesis: Japan’s megabanks are systematically acquiring AI capability that their internal technology teams cannot build. MUFG has been investing across Southeast Asia for years. KB Kookmin and Shinhan are expanding into Indonesia and Vietnam. The demand side of the M&A market for AI fintech in APAC is active and well-capitalised.
The Comparison: Eight AI Fintech Players
The following eight companies were selected based on: genuine AI-native differentiation (not AI marketing), disclosed funding of $50M or more, active presence in APAC markets, and strategic relevance to M&A acquirers.
| Company | Country | Founded | Total Funding | Valuation | Sub-vertical |
|---|---|---|---|---|---|
| Airwallex | AU / HK | 2015 | ~$900M | $8B | AI Payments / Treasury |
| Toss (Viva Republica) | South Korea | 2013 | ~$900M | $10–15B (pre-IPO) | AI Banking Super App |
| WeLab | Hong Kong | 2013 | ~$700M | $1B+ | AI Digital Banking |
| Advance Intelligence Group | Singapore | 2016 | ~$700M | $2B+ | AI Credit / KYC / BNPL |
| Kredivo Holdings | Indonesia | 2015 | ~$660M | ~$1.6B | AI Credit / BNPL / Digital Bank |
| Thunes | Singapore | 2016 | ~$350M | $1B+ | AI Cross-border Payments |
| Perfios | India | 2008 | ~$263M | ~$1.2B | AI Credit Infrastructure |
| Tongdun / TrustDecision | China / SG | 2013 | ~$246M | ~$1B | AI Risk / Fraud Detection |
AI Differentiation Tier: What Each Company’s AI Actually Does
The most important lens for M&A valuation is not how much AI a company claims — it is whether the AI is the product or decoration on top of a product.
Tier 1 — AI is the product (remove AI, product ceases to exist)
Perfios built the credit infrastructure layer for South and Southeast Asia. Its CAM AI system, launched in 2025, reduces multi-day credit assessments by 85% by autonomously extracting and reconciling data from GST filings, bank statements, and financial documents. It processes over 5 million transactions per month across 50+ financial institutions. The company aggregates and analyzes financial data in real time from 7,500+ financial institutions. There is no Perfios without the AI — the product is the decision engine.
ADVANCE.AI runs three distinct AI engines: eKYC and fraud detection for 500+ banks and fintechs, a BNPL credit scoring system that approves credit in under 60 seconds for unbanked Southeast Asian consumers, and an AI-driven P2P lending operation in Indonesia. The models are trained on over a decade of alternative data for populations that have never had a credit bureau score. The IP compounds with every loan decision.
Tongdun / TrustDecision uses graph neural networks to detect fraud rings — not individual fraudulent transactions, but the coordinated networks behind them. Its behavioral biometrics layer analyzes device fingerprinting and typing patterns in real time. The explainable AI output (a regulatory requirement for many Asian banking regulators) is a technical capability most incumbents do not have.
Kredivo evaluates 2,000+ data points for credit approval decisions in Indonesia — telecom usage, e-commerce transaction history, social signals, and bank statement analysis. Its net charge-off rate of 3.8% compares to a 5.6% Indonesian bank average. The model’s edge is in the data, not the algorithm, and the data compounds with every loan.
Tier 2 — AI materially transforms an existing product
WeLab repositioned explicitly as an “AI-first” bank following a strategic partnership with Google in 2025. Its AI Investment Research Agent uses Gemini models to analyze wealth management datasets. Its AI-powered FX service was marketed as Hong Kong’s first. AI is now embedded across lending, wealth management, and daily banking. However, WeLab’s core banking product predates the AI layer — the bank would exist without it, just with lower differentiation.
Toss deploys AI credit scoring for 8–10 million Korean adults without traditional credit history, real-time ML fraud monitoring across 30+ million monthly active users, and behavioral financial recommendation engines. AI is deeply integrated. But Toss is also a payments super app, a digital bank, an insurance comparison platform — AI is a core capability layer, not the sole product.
Airwallex has built genuine AI agents: an Expense Submission Agent that collects receipts from email and WhatsApp and matches them to transactions, an Expense Policy Agent that validates compliance in real time, and an AI Assistant that executes multi-currency account setup via conversational commands. These are real AI products built on top of what started as a payments infrastructure company.
Thunes is cross-border payment rails across 130+ countries. Its AI layer is primarily in compliance: ML-driven AML and sanctions screening integrated into real-time payment routing. The AI reduces false positive rates in compliance screening — meaningful but not the core value proposition.
Data Moat Analysis
The most durable competitive advantage in AI fintech is proprietary data — specifically, data that incumbents cannot buy or recreate by switching on a model.
Highest moat — 10+ years of alternative credit data for unbanked populations: Perfios, ADVANCE.AI, Kredivo. These companies have been underwriting credit for consumers and businesses that do not exist in any bureau. That data — millions of credit decisions with outcomes — trains models that improve continuously. A Korean bank acquiring Kredivo is not just buying a portfolio; it is buying a decade of Indonesian consumer behavioral data that cannot be constructed from scratch.
High moat — multi-country real-time payment flows: Airwallex, Thunes. These companies see more cross-border transaction data than most banks. Routing intelligence, FX optimization, and compliance pattern recognition all improve as transaction volume scales. The moat compounds as the network grows.
Moderate moat — domestic market behavioral data: Toss, WeLab. Dominant in their home markets with strong behavioral data. The limitation is geographic concentration — the data moat is Korean or Hong Kong-centric, not pan-regional.
Specialized moat — fraud network graph data: Tongdun. Graph data on fraud rings is genuinely proprietary and highly defensive — fraud rings evolve, so detecting them requires constantly updated, real-world feedback loops. Companies that don’t operate in the fraud detection space cannot buy this data.
License and Regulatory Stack
For acquisition purposes, regulatory positioning is as important as technology.
Full digital banking licenses (highest regulatory value):
- WeLab Bank (Hong Kong virtual bank license + Bank Saqu, Indonesia)
- Toss Bank (South Korean banking license)
- Kredivo / Krom Bank (Indonesian commercial bank)
These licenses have real scarcity value. Change-of-control approval requirements in each jurisdiction add timeline and regulatory complexity to any acquisition — but they also mean the acquirer inherits a license that cannot be replicated by a greenfield entrant.
Payment institution licenses:
- Airwallex (licensed in 60+ markets)
- Thunes (payment licenses in 130+ countries — the broadest geographic licensing stack on this list)
Software-only / no financial licenses:
- Perfios (B2B software platform; the banks are the licensed entities)
- ADVANCE.AI (enterprise SaaS for banks; holds no consumer licenses)
- Tongdun (software vendor; sells to licensed institutions)
Software-only companies have faster change-of-control processes but do not carry license scarcity premium. The tradeoff: easier to acquire, less regulatory leverage in price negotiations.
Company Profiles
Airwallex
Airwallex crossed $1 billion in annualized revenue in October 2025 — 90% growth year-on-year — and closed a $330 million Series G at an $8 billion valuation in December 2025. It processes over $235 billion in annualized transaction volume, doubled from 2024. With 150,000+ business clients across 60+ markets and new licenses in Korea, Japan, Malaysia, Vietnam, Brazil, Mexico, and the UAE in 2025, Airwallex has effectively become the cross-border treasury operating system for mid-market and growth-stage businesses globally.
The AI layer: Airwallex has moved from payment rails toward agentic finance. Its Expense Submission Agent, Expense Policy Agent, and embedded AI Assistant automate the full AP/AR and expense lifecycle. The vision is a self-operating finance function — where AI handles compliance, routing, and reconciliation autonomously.
Exit pathway: IPO at $12–15 billion is the most likely outcome within 18–24 months. An outright acquisition at current scale would require a mega-institution — JPMorgan, HSBC, Stripe, or a global payment network. The $8 billion valuation is above typical APAC strategic acquirer appetite.
Toss (Viva Republica)
Toss reported its first annual net profit in 2024 and has sustained 35–41% revenue growth through 2025. Its December 2025 funding round at an implied $10–15 billion valuation included GIC, Baillie Gifford, Wellington, and WCM — the investor profile of a pre-IPO, not a venture-stage, company. The international HQ in Singapore signals intent for Southeast Asian expansion beyond Vietnam.
The AI layer: AI credit scoring for the Korean thin-file population, real-time fraud monitoring at 30+ million MAU scale, and personalized recommendation engines across insurance, investments, and banking. The models are trained on full financial histories — Toss users give the app access to all their bank, investment, and insurance accounts.
Exit pathway: US Nasdaq or NYSE IPO is the stated objective. At this valuation, Toss is more likely an acquirer of APAC assets than an acquisition target. It has the capital, the product depth, and the organizational scale to consolidate Southeast Asian digital banking — if management chooses that path over a pure IPO.
WeLab
WeLab’s $220 million Series D in January 2026 — with Prudential, Fubon Bank, HSBC, Allianz X, and the Hong Kong government’s investment arm as investors — was described as the largest digital banking capital raise in Asia in 2025. The investor lineup is unusually strategic: a major insurer, two banks, and a government fund all writing checks signals that WeLab is being groomed for a larger role in Hong Kong’s financial ecosystem.
The AI layer: Google partnership formalised in 2025. Gemini-powered investment research agents, AI FX optimization (Hong Kong’s first AI-driven FX service with a best-rate guarantee), AI-generated marketing (TV commercials made entirely by Gemini and Veo), and behavioral hyper-personalization across three markets.
Exit pathway: The investor composition suggests a structured liquidity event rather than an open-market IPO — possibly a strategic acquisition by HSBC (which already holds a stake via the Series D), Standard Chartered, or a mainland Chinese financial group expanding post-regulatory normalisation. Timing likely 2027–2028.
Advance Intelligence Group
Advance Intelligence Group operates three distinct businesses on a shared AI infrastructure: ADVANCE.AI (enterprise B2B), Atome Financial (BNPL across Southeast Asia), and Kredit Pintar (P2P lending in Indonesia). The SoftBank Vision Fund 2 and Warburg Pincus participation signals both capital depth and the expectation of a significant liquidity event.
The AI layer: ADVANCE.AI’s eKYC product uses computer vision and liveness detection to process identity documents and onboard customers in under 60 seconds for 500+ institutions. The credit scoring models use 3,000+ data points for the unbanked, with over a decade of behavioral data from Indonesian, Filipino, and Vietnamese consumers.
Exit pathway: A PE take-private (Warburg Pincus is already on the cap table) followed by a strategic sale or IPO is plausible. The enterprise B2B business (ADVANCE.AI) could also be carved out and sold independently to a global RegTech acquirer — Experian, LexisNexis Risk, or FICO have all made similar acquisitions in adjacent markets.
Kredivo Holdings
Kredivo is the furthest along in building a full financial services stack for Indonesia’s mass market. Since its $270 million Series D in 2023, it has converted Krom Bank to a full digital bank (enabling deposit-taking), acquired GajiGesa (earned wage access) in February 2025, and grown its loan book 28% to IDR 9.6 trillion. The combination of BNPL origination, a banking license, and payroll-linked EWA makes Kredivo the most complete consumer credit operating system in the Indonesian market.
The AI layer: 2,000+ data point credit decisioning for BNPL approval in under 60 seconds. Net charge-off rate of 3.8% against an Indonesian bank average of 5.6%. The model is trained on a decade of Indonesian consumer behavioral and e-commerce data — increasingly rare and defensible.
Exit pathway: Indonesia is the prize for pan-Asian financial groups. SCBX paid $860 million for Home Credit Vietnam; Kredivo is a comparable or larger prize. Korean financial groups (KB Kookmin, Shinhan) are active in Indonesia. Japanese megabanks are looking at the same market. GoTo Financial partnership dynamics add further strategic optionality.
Thunes
Thunes achieved unicorn status in early 2025 on a $150 million Series D led by Apis Partners and Vitruvian Partners. It connects banks, payment service providers, merchants, and mobile wallet operators across 130+ countries — the broadest geographic payment network on this list. Its World Economic Forum unicorn community membership signals active preparation for either a strategic exit or a large pre-IPO round.
The AI layer: ML-driven compliance screening across 130 regulatory environments makes real-time decisions on payment routing, AML screening, and sanctions checking at global scale. The value is less in the AI itself and more in the data from processing hundreds of millions of cross-border transactions — routing intelligence and compliance pattern recognition that cannot be built from scratch.
Exit pathway: Global payment networks are the natural acquirers: Mastercard, Visa (via B2B Connect), Western Union, or Nuvei. E-commerce platforms with cross-border settlement needs — Shopee, Lazada, Alibaba’s international businesses — are also strategic fits. An IPO on the Singapore Exchange is possible given the APAC focus and regulatory familiarity.
Perfios
Perfios is the least well-known company on this list outside India — and arguably the most strategically important infrastructure company in APAC credit AI. It has achieved profitability ($12.5 million net profit on $80 million revenue in FY2025), is actively acquiring (CreditNirvana, Clari5, IHX in Q1 2025 alone), and has launched CAM AI — a credit assessment automation product that replaces multi-day manual credit analysis with autonomous document extraction and reconciliation.
The AI layer: Perfios aggregates financial data in real time from 7,500+ sources, processes 5 million+ transactions per month, and has 18 country deployments. CAM AI is deployed by 50+ institutions. The company is the infrastructure layer powering credit decisions for a significant fraction of APAC lending — and it improves with every transaction it processes.
Exit pathway: IPO on Indian markets (BSE/NSE) is the most likely near-term path given Indian investor appetite for profitable fintech infrastructure. Strategic acquisition by a global credit bureau (Experian, TransUnion, Equifax) attempting to build AI-native underwriting capability in Asia is the alternative. The active acquisition spree suggests management is building for scale before a formal exit process.
Tongdun / TrustDecision
Tongdun is China’s oldest pure-play AI risk company. Its international brand, TrustDecision, has established a meaningful presence in Southeast Asia, South Asia, and Latin America, serving banks and fintechs that need fraud detection without building it in-house. The Temasek investment is an important signal of international credibility given Temasek’s careful vetting of Chinese technology companies for APAC expansion.
The AI layer: Graph neural networks for fraud ring detection (not just individual fraudulent transactions), behavioral biometrics (device fingerprinting, typing pattern analysis), real-time risk scoring with explainable AI output. The explainability layer is meaningful: Southeast Asian banking regulators increasingly require AI decisions to be auditable, and Tongdun’s architecture was designed for this.
Exit pathway: A strategic acquisition by a global risk intelligence firm (LexisNexis Risk Solutions, FICO, Experian’s APAC division) or by a Southeast Asian bank building proprietary fraud infrastructure in-house. The China headquarters creates some cross-border M&A complexity; the Singapore entity structure is the likely vehicle for any international transaction.
M&A Transaction Log: What Has Already Happened
The following deals from 2024–2026 are the most instructive precedents for pricing and deal structure in APAC AI fintech.
| Acquirer | Target | Value | Date | Rationale |
|---|---|---|---|---|
| Mizuho Financial Group | Upsider (AI corporate credit) | ~$313M (70% stake) | July 2025 | Japanese megabank acquiring AI credit model for SMB; 80,000+ corporate customers; defensive move against digital-native credit competitors |
| SCBX (Thailand) | Home Credit Vietnam | $860M | Feb 2024 | Largest SCBX acquisition; buying consumer credit infrastructure to establish Vietnam presence |
| Aurionpro Solutions | Arya.ai (deep learning for banks) | ~$16.5M (67%) | April 2024 | Fintech software vendor acquiring AI-native bank platform for credit analysis, fraud, and KYC automation |
| NTT DATA | GHL Systems (Malaysia payments) | ~$165M | May 2024 | Japanese tech giant acquiring SEA payment processing rails |
| Kredivo Group | GajiGesa (earned wage access) | Undisclosed | Feb 2025 | Vertical integration of credit stack into payroll-linked lending |
| 360 One WAM | ET Money (wealth/investment app) | ~$44M | June 2024 | Wealth management consolidation via AI-powered retail investment distribution |
| Investcorp | NSEIT (NSE AI/cloud/cybersecurity spinout) | ~$120M | April 2024 | PE acquiring AI and digital transformation company spun out of India’s National Stock Exchange |
| Perfios (as acquirer) | CreditNirvana, Clari5, IHX | Undisclosed | Q1 2025 | AI credit lifecycle rollup: adding collections, fraud, and health insurance data in one quarter |
Key pattern: Japanese and Korean financial groups are paying strategic premiums for AI capability — the Mizuho/Upsider deal implied a capability premium rather than a pure revenue multiple. Southeast Asian financial groups (SCBX) are paying for credit infrastructure and geographic expansion simultaneously. Indian credit infrastructure companies are beginning to consolidate horizontally.
Valuation Benchmarks
Current revenue multiples by sub-vertical, as of Q4 2025–Q1 2026:
| Sub-vertical | Revenue Multiple | Notes |
|---|---|---|
| AI credit infrastructure (B2B SaaS) | 8–15x | Highest premiums; proprietary data moats |
| RegTech / Compliance AI | 6–12x | Regulatory tailwinds; high switching costs |
| AI fraud / risk detection | 6–10x | Pure-play AI vendors command strong premium |
| Wealth management / robo-advisory | 5–10x | AUM scale determines upper range |
| AI payments / cross-border | 4–6x | Commoditization pressure; network scale matters |
| AI lending platforms | 2–4x revenue; 6–10x EBITDA | Banking license adds premium; blended model |
Reference data points:
- Global AI acquisitions commanded an average revenue multiple of 24x versus 12x for traditional software in 2025 (Capstone Partners data)
- General APAC fintech M&A averaged 4.2–4.7x revenue in Q4 2025, recovering from 2.6–2.8x in Q1 2025
- AI-focused fintech pulled $7.2 billion in H1 2025 globally, a pace 40% ahead of full-year 2024
- AI-related deals represented 17% of all fintech transactions in Q3 2025
The Mizuho/Upsider deal is the most relevant APAC AI fintech precedent: an implied enterprise value of approximately $450 million for a company with a relatively small disclosed revenue base signals that Japanese acquirers are willing to pay technology and capability premiums that pure revenue multiples do not explain.
Exit Pathway Assessment
Sorted by likelihood of M&A exit versus IPO in the next 24–36 months:
Most likely M&A exits:
- Perfios — Profitable, acquisitive, building infrastructure that global credit bureaus need. IPO possible but strategic acquisition by Experian or TransUnion offers a clean exit.
- ADVANCE.AI — SoftBank portfolio monetisation pressure and the separability of the enterprise B2B business create favorable conditions for a carve-out or full acquisition.
- Tongdun / TrustDecision — At $1B valuation with a Temasek investment, strategic sale to a global risk intelligence firm is the most logical path.
- WeLab — Strategic investor lineup (HSBC, Prudential, government funds) reads like preparation for a structured exit rather than open-market IPO.
Most likely IPO candidates:
- Toss — US listing explicitly targeted; investor base already includes institutional pre-IPO holders.
- Airwallex — Revenue scale ($1B annualized), profitability trajectory, and global footprint point to $12B+ IPO.
- Kredivo — Indonesian market depth and banking license make this a compelling standalone IPO on Jakarta Stock Exchange or a dual listing.
Wild card — possible acquirer: Thunes — While it is positioned as an M&A target, it could itself be the acquirer of smaller cross-border payment networks in Africa and Southeast Asia, building toward a scale that makes an IPO or a very large strategic sale the only viable exit.
How Amafi Advises in This Market
The APAC AI fintech sector requires advisors who understand the regulatory nuances of each market, the acquirer dynamics of Japanese and Korean financial groups, and the technical due diligence that AI-native companies face that traditional fintech companies do not.
Typical mandates we support in this vertical:
- Sell-side processes for AI fintech companies approaching Japanese megabanks, Korean financial conglomerates, and global payment networks
- Carve-out advisory for AI spinouts from larger fintech groups (the NSEIT / Perfios / ADVANCE.AI model)
- Strategic positioning for AI fintech founders preparing for a 12–18 month exit process — including technical IP documentation, data room structuring, and acquirer outreach
- Competitive intelligence on who is buying in this space and at what multiples
If your company operates in AI fintech in APAC and you are considering a strategic process, contact us for a confidential conversation.