$185BB2B SaaS embedded-payments TAM·<20% captured·BCG + Adyen, Oct 2024
Embedded finance =
Every app is becoming a bank.
Shopee moves payments. Lazada finances its sellers. LINE Pay issues cards. Airtel holds balances. All of it powered by invisible banks behind a handful of APIs — except Grab, which got the licences: a major payment institution licence from MAS in Singapore and a digital bank licence in Malaysia. Once a BNPL-and-super-app story (Atome, Aspire, Money Forward), embedded finance is now the default playbook for vertical SaaS, marketplaces, B2B platforms, and even non-financial brands — anyone who can ship a payment, issue a card, write a loan, hold a balance, or read a bank account without sending users to a third party. The 18-month bank-charter project is now a two-week API integration.
How embedded finance in Asia-Pacific differs
There is no single Asia-Pacific stack. Licensing is per country — a payment institution licence from MAS in Singapore, digital bank licences in Malaysia, Hong Kong and the Philippines, UPI and a payments-bank regime in India — so the same super-app runs a different stack in every market it enters.
The super-apps are the embedded-finance players here. Grab, Shopee and LINE embed wallets, seller loans and cards inside marketplaces and messaging apps, and several of them now hold the bank licence themselves rather than renting one.
Cross-border is the layer that stands out: Airwallex, Aspire and Nium started as multi-currency infrastructure for regional businesses, and treasury and FX show up in stacks earlier than they do in the West.
Japan is its own pattern again — accounting-led, with Money Forward and freee embedding payments and financing into the ledger rather than the checkout.
Evidence from subprocessor lists, regulatory footers, trust pages, customer announcements, and ~594 press articles — produced by Apideck with content partners Open Banking Tracker and Embedded Finance Review.
Headline finding ·Stripe is named by 60 of 546 fintechs (11.0%) — the only vendor above 5% concentration in any infrastructure layer.
companies disclose their sponsor bank than name their KYC vendor (52 vs 21). The compliance layer goes dark by design.
FINDING 02
194
third-party providers serve the 546 fintechs in the dataset — defined as ≥2 disclosed buyers or recognized Open Banking Tracker vendor.
FINDING 03
38 non-bank fintechs
have crossed from vendor relationship to owned banking, clearing, or card-issuing infrastructure — Square Financial Services, Coinbase, LendingClub Bank and others.
FINDING 04
Zero overlap
between US, UK, and EU BaaS leaders. US: Sutton Bank / Evolve Bank & Trust / Cross River. UK: ClearBank / Modulr / Railsr. EU: Swan / Mangopay / Treezor.
FINDING 05
US discloses 2-5×
more than other regions. 25% of US fintechs name their sponsor bank vs. 5-13% elsewhere — the widest gap is US vs EU (25% vs 5%). Driven by US regulatory disclosure norms (FDIC member badges, BSA/AML attestations) that don't exist in UK/EU.
FINDING 06
Stripe #1
PSP by disclosed-buyer count (60 of 546, 11.0%). BaaS leader: Nium (7, 1.3%). Most-cited accounting connector: Xero (10.6%).
Section 02 · disclosure rates
Which layers are auditable. Which go dark.
For each layer of the embedded-finance stack: what share of 546 analyzed fintechs publicly named a vendor? The gap between the green and red rows is the dark half of the stack — the layers regulators, integrity researchers, and journalists can’t audit from outside.
Regulatorily disclosed (named on trust/regulatory surface) Commercially disclosed (named via case studies, partnerships) Trade-secret layer (rarely disclosed publicly)
What this meansThe payments side of the stack is auditable from outside. The risk side isn’t. Public datasets of vendor relationships — including this one — will always over-represent payments and under-represent KYC, fraud, and sanctions screening. Treat the red rows as a floor, not a ceiling. The “% masked” column is the second-order finding: even when a layer IS disclosed, a meaningful fraction of companies talk about their own-brand product (“Acme Pay”, “Acme Checkout”) without identifying the underlying provider. ~34% of fintechs that mention a PSP, and ~37% of fintechs that mention BaaS, hide the actual infrastructure behind a brand wrapper — a structural masking effect that makes the auditable-looking layers thinner than they appear.
Section 03 · dataset composition
What kind of fintech are these?
The 546 fintechs in this dataset split into 9 working categories. Infrastructure providers (BaaS, card issuing, PSP) sit alongside the platforms that consume them (vertical SaaS, marketplaces, B2B fintech) — the report is a snapshot of an ecosystem where buyer and seller often appear in the same chart. 849 entries were excluded as non-fintech (investor/advisory firms, non-financial orgs, press fragments) before this breakdown.
Fintech infrastructure
195·35.7%
Vertical SaaS
90·16.5%
Consumer neobank
66·12.1%
Incumbent FI
37·6.8%
B2B fintech
32·5.9%
Marketplace / platform
28·5.1%
Consumer brand (embedded)
27·4.9%
Non-fintech org
21·3.8%
SMB lender
19·3.5%
Crypto / Web3
15·2.7%
Other
10·1.8%
Insurtech
6·1.1%
Segments resolved via an enrichment cache + a local-profile fallback classifier sniffing each company’s public industry, tagline, and thesis. “Other” entries are kept in the dataset; we surface them as a visible gap rather than dropping them or guessing.
Section 04 · sponsor banks
Who backs the banking
Every non-bank fintech that holds customer balances or issues cards sits on top of a US regulated bank charter — the “sponsor bank.” This is the single most-disclosed layer in our dataset because it’s a regulatory requirement to name. No single sponsor bank dominates. The top 5 (Federal Bank, RBL Bank, Yes Bank, YES Bank, Column) together back 31% of disclosed relationships — the rest spreads across 48+ smaller banks.
53 distinct sponsor banks across 52 disclosed relationships. The primary percentage is share of all 546 analyzed fintechs — the one canonical denominator used across this report; the parenthetical “of disclosed” is share of the 52 fintechs that name a sponsor bank. A sponsor bank is the chartered institution that holds the regulatory charter — distinct from a BaaS provider (the API/platform layer), so this ranking differs from the BaaS leaderboard even where the same names appear.ALL 53 →
What this meansFor regulators: there’s no concentration risk; oversight has to scale to every regional bank doing BaaS. For VCs: fintechs have real sponsor-bank optionality, and the post-Synapse fragmentation is structural, not a transition.
Section 05 · category leaders
Who's most-disclosed in each layer · APAC
A handful of providers dominate every category — and the share of the leader tells you whether the layer is a winner-take-most market or still fragmented. Filtered to APAC (546 companies). Clear the filter to see global leaders.
Combines each provider's disclosures across all 12 stack layers into one ranking — the question a single category leaderboard can't answer on its own: who's the most-used provider in the whole dataset, full stop.
Traditional payment service providers — sold to merchants to take their own payments. Many category leaders (Stripe, Square, Adyen) self-process and don't appear as customers of any third-party PSP.
Open-banking aggregators (Plaid, TrueLayer, Tink, Yapily) plus unified-API providers over accounting / payroll-HRIS / FX (Apideck, Codat, Rutter, Merge, Finch, Pinwheel, Argyle). The most-consolidated infrastructure category — GoCardless holds 1% of disclosures alone — and the one most likely to reshuffle as open-finance regulation lands: CFPB §1033, UK CMA, and Brazil Phase 4 in 2026-27, with the EU's FIDA framework (still stalled in trilogue, realistic application closer to 2029-30) following later. DISCLOSURE: Apideck (the report publisher) appears in this leaderboard. Apideck's customer page was a seed source for the dataset, inflating its absolute count relative to vendors whose customer walls weren't scraped. The relative ranking of every other vendor is unaffected.
Platform-payments specialists — sold to vertical SaaS, marketplaces, and platforms so they can offer payments to their own customers. Distinct from the PSP card above; same buyer disclosure may appear on both when relevant. Curated allowlist (Finix, Moov, JustiFi, Tilled, Rainforest, Worldpay for Platforms, Modulr, Mangopay, Banking Circle, Swan).
CheckHQ has aggressively owned vertical SaaS payroll — most B2B platforms that ship payroll embed it. Includes EWA (W-2 wage advances: DailyPay, Payactiv, Wagestream) and gig-worker instant pay (Uber, Lyft, DoorDash) — the latter is technically faster-settlement of completed earnings, not a wage advance, but the user experience is identical and most analyst writeups collapse both.
Consumer (KYC) and business (KYB) identity verification, sanctions screening, onboarding decisioning. Tracked together because most vendors sell both (Alloy, Persona, Sumsub, Stripe Identity) — KYB specialists like Middesk and Baselayer sit alongside. Structurally the most under-disclosed layer; the leaderboard captures the publicly-disclosed subset (mostly via vendor case studies, not buyer-side surfaces). Real adoption is materially higher.
Fraud orchestration, transaction monitoring, AML decisioning (Sardine, Sift, Kount, Signifyd, Forter, ComplyAdvantage). Even more under-disclosed than KYC — fraud stacks rarely surface in public materials at all. The disclosed names here are mostly back-attributed from vendor case studies.
Treasury-operations APIs, sweep / cash-management products, yield partners, multi-bank cash visibility. Modern Treasury leads the ops side, Stripe Treasury / Mercury Treasury / Brex Treasury the sweep+yield side. Mercury, Brex, and Stripe themselves are tracked separately from their Treasury products — a buyer banking with Mercury doesn't auto-route to Mercury Treasury without explicit yield-product disclosure.
Programmable core-ledger infrastructure — the system of record that tracks every debit, credit, and reconciliation. Distinct from BaaS (which provides the bank rails) and Treasury (which manages corporate cash). Providers: Formance, TigerBeetle, Fragment, Twisp, Mambu / Temenos / Thought Machine (the core-banking platforms include a ledger layer).
Card / bank-transaction data cleaners — turn raw merchant strings into structured merchant names, categories, and MCCs. Used by neobanks, expense-management platforms, and accounting tools. Emerging category with a long tail; the top three (Dateio, Astrada, Spade) account for most disclosed adoption.
Sales-tax, VAT and indirect-tax engines that platforms embed to calculate, file and remit tax for their customers. A vendor is counted only when a buyer discloses it; tax vendors are not counted as buyers of their own competitors.
Invoicing, collections, and receivables-reconciliation platforms embedded white-label into other platforms' own products. Distinct from Lending & Liquidity's receivables FINANCING (invoice factoring, receivables-backed credit) — this is the invoicing/collections software itself, not capital advanced against it.
Corporate-card issuance and bill-pay / AP automation embedded into other platforms' own products. Distinct from Accounts Receivable (the other side of the ledger) and from Card Issuing (the BIN/processor layer the cards themselves ride on).
What success looks like
Three platforms whose 10-K filings show embedded finance pays
Public-company evidence that platforms which embed financial products end up out-earning their software business. From the comprehensive PDF report.
Toast
82% of revenue is fintech
FinTech Solutions revenue reached $4.05B of $4.96B total in FY2024 — ~82% of revenue. The fintech mix is what makes the unit economics work; software alone wouldn't.
Source · Toast 10-K · FY2024
Shopify
$4.2B in Shopify Capital originations
Shopify Capital originated $4.2B in 2025, ~40% year-over-year growth. Embedded lending now sits alongside payments as a primary monetization layer — the loan book scales with merchant GMV, not just SaaS seats.
Source · Shopify Capital · 2025 disclosure
ServiceTitan
95%+ gross dollar retention
Embeds payments, payroll, and financing across its contractor base. 95%+ gross dollar retention across FY2023-FY2025 — among the highest reported for vertical SaaS at scale.
BaaS, open finance, and sponsor banking look almost completely different depending on where the company is incorporated. Disclosure norms vary even more — US fintechs name their stack far more often than the rest of the world (see the sponsor-bank gap below), an artifact of regulatory disclosure rules (FDIC member badges, BSA/AML attestations) that don’t exist elsewhere.
Where the index’s companies are
6,593 analyzed companies across 106 countries, shaded by concentration — heaviest in 🇺🇸 United States (2,444). HQ country, not where a company’s customers or operations are.
FewerMoreNo disclosed HQ
🇺🇸 United States · 2444🇬🇧 United Kingdom · 655🇧🇷 Brazil · 340🇳🇬 Nigeria · 253🇩🇪 Germany · 216🇨🇦 Canada · 199🇫🇷 France · 186🇮🇳 India · 161+98 more
BaaS by region
No overlap between US (Sutton Bank, Evolve Bank & Trust, Cross River), UK (ClearBank dominant, then Modulr/Railsr), and EU (Swan, Mangopay, Treezor). Each region has its own BaaS oligopoly.
Open finance
Plaid leads US disclosures by a wide margin — more than every other open-banking aggregator combined. The category is genuinely competitive only in UK/EU, where TrueLayer, Yapily, Tink, and Salt Edge each hold real share.
Disclosure asymmetry
25% of US fintechs name their sponsor bank vs. 5-13% elsewhere — a 2-5× gap. Outside the US there often isn't a sponsor bank to name — UK/EU fintechs typically hold their own EMI license.
Region inferred from company_profile.headquarters. “Top” = the most-disclosed vendor within each region (not necessarily the highest revenue / market share). “Other” = HQ not inferable from the dataset.
Section 07 · self-operated infrastructure
The self-banked club
Most fintech companies rent their banking, card-issuing, and payment infrastructure from a vendor. But a growing tier of category leaders has crossed the threshold and now operates that infrastructure themselves — through an owned bank subsidiary, an in-house clearing entity, or a self-issued payment product.
When a company graduates from vendor relationship to owned stack, the economics flip: margin retained, latency controlled, but compliance, capital, and operational burden internalized. 38 of 546 companies in our dataset have crossed that line for at least one layer.
Payment processors and BaaS providers are routinely named on trust pages and subprocessor lists. The compliance layer — KYC, fraud, sponsor banking — is far more guarded. Comparing public disclosure rates across the dataset is itself the finding.
Why the asymmetry: sponsor-bank disclosure is regulator-driven (FDIC Section 18 “Member FDIC” signage rule, OCC third-party-risk-management OCC 2023-17); KYC-vendor identity is treated as competitive information under the GLBA Safeguards Rule (no disclosure obligation). One layer is forced into the open; the other is allowed to stay private. The data reflects exactly that.
Sponsor bank named
52
/ 546 · 10%
Disclosed in regulatory footers, terms of service, and licensing references.
KYC vendor named
21
/ 546 · 4%
Identity-verification providers (Persona, Onfido, Jumio) are almost never publicly attributed by the customer.
Fraud tool named
3
/ 546 · 1%
Fraud-scoring and decisioning vendors (Sardine, Sift, Forter) sit even deeper in the stack.
Known failures · 2024-25
Sourced from public regulatory orders
Why public-disclosure matters: the 2024-25 enforcement cycle
Six named failures since 2023. Each one took partner fintechs offline, locked out customers, or revealed pooled-account architectures with no independent source of truth. The compliance layer goes dark in normal times — and it’s the layer that fails first.
SynapseUS · 2024
Chapter 11. Serving ~100 fintechs and four partner banks. Over 100,000 customers locked out. Trustee identified $65-95M shortfall.
Root cause · Pooled FBO accounts with no independent source of truth.
OCC consent order (original 2022, amended Jan 2024). Grew to ~70 fintech partnerships at peak without scaling compliance. Required to reduce partnerships and raise capital.
Federal Reserve cease-and-desist (Jun 2024). Partner to Stripe, Mercury, Affirm, Airwallex, Dave. Created regulatory contagion across the BaaS ecosystem.
The infra layer either breaks in public or ships something concrete in public — same layer, same stakes, opposite outcome. Same bar as the failures above: every result here is a number or a before/after, not a marketing adjective.
Fundbox × LithicCard Issuing · 2026
Fundbox and Lithic built a loan-to-card program that collapses approval and spend into one session: borrowers apply, get approved, and spend from a virtual card immediately, instead of waiting days for an ACH transfer to land.
Outcome · Shipped in 7 weeks. Approval-to-spend: days (ACH) → one session.
2.5× more companies disclose a sponsor bank than name a KYC vendor.
Sponsor-bank relationships surface because they’re a regulatory requirement to disclose. Identity-verification providers are treated as trade secrets — and the rest of the compliance stack (fraud orchestration, sanctions screening, transaction monitoring) is darker still. The vendor catalog you can build from public evidence is therefore systematically biased toward the payments side and away from the risk side.
53 distinct sponsor banks appear across 52 disclosures. Percentages are share of all 546 analyzed fintechs (different cut from Section 04, which shows share of the 52 disclosed-only subset).
Charter queue · 2025-26
Sourced from OCC decision letters and filing announcements
Who is applying to become a bank
38 applications tracked since the OCC reopened de novo chartering: 8 application filed, 26 conditional approval, 2 approved, 2 denied. The ten most recent are below; the full queue has every row with its filing and decision record. Most are national trust banks — custody and stablecoin settlement under one federal supervisor instead of 50 state licences — not deposit-taking banks. Each one is a sponsor-bank relationship that ends when the charter opens.
Primary source or two independent outlets, per row
Who is building shared rails together
14 industry alliances tracked — 8 operational, 6 forming. Distinct from a company’s own bilateral partnership (that lives on its dossier): this is several organizations jointly building a shared standard or a new, jointly capitalized entity. The full page has every member and source.
Open Standard (OUSD stablecoin)#Operational · launched 2026-09-30
A USD-denominated stablecoin (OUSD) built as shared, co-owned payments infrastructure rather than a single company's product — mintable and redeemable at par with no fees or volume caps, live on Ethereum, Solana, Base and Tempo. Founding partners share governance and reserve income rather than a single issuer capturing it, positioned for banking, cross-border payments, settlement, institutional trading and lending.
Canada's Big Six tokenized-deposit network#Forming · announced 2026-09-22 · targeting no date set; phase 1 limited to transfers between the 6 participating banks
A shared permissioned ledger letting Canada's six largest banks move tokenized Canadian-dollar deposits between one another instantly, 24/7 — phase 1 covers interbank transfers only (not consumer wallets), targeting faster back-end settlement and programmable payments later, while keeping the same legal/regulatory treatment as ordinary bank deposits. Framed explicitly as competitive defense against USD stablecoins, informed by the Bank of Canada's Project Agorá tokenization research.
Global bank USD stablecoin consortium#Forming · announced 2026-09-01 · targeting H1 2027 (USD token); additional G7 currencies, EUR prioritized, to follow
21 banks and asset managers across North America, Europe, East Asia, the Middle East and Africa committing to jointly form and capitalize a company that will issue a USD-denominated stablecoin on public blockchains (not a closed private ledger) for wholesale, institutional and retail use, including cross-border payments and digital-asset settlement.
Because fintechs rarely publish their identity-verification vendor, we inverted the lens: pulled the public customer pages of the five major KYC vendors themselves, then back-attributed each named customer to its vendor. 115 disclosed integrations across 8 vendors.
Customers compiled from each vendor’s public case-study, customer-grid, or success-story page, plus first-party integrations observed directly in buyer onboarding flows. Includes companies outside the embedded-finance dataset; counts reflect disclosed integrations, not market share.
Looking forward
Four things to watch in 2027
The patterns above set up several testable predictions. The dataset refreshes quarterly, so deltas surface as new disclosures land — not only at the annual edition.
PREDICTION 01
KYC disclosure stays under 10%
Sponsor banks disclose because regulators force them to. KYC vendors don't have that pressure, and our prediction is the gap widens — not narrows — as more fintechs view their fraud and identity stack as a competitive moat. Watch for 1-2 high-profile vendor disclosures from regulated neobanks; everyone else will stay quiet.
PREDICTION 02
US BaaS oligopoly consolidates further
Cross River, Bancorp, Pathward, and Lead Bank already concentrate disclosed US sponsor-bank relationships. Synapse's collapse pushed a wave of mid-tier fintechs onto a smaller bench. We expect the top-4 to absorb ≥75% of new US BaaS launches by end of 2027, with Column and Lead Bank making the biggest gains.
PREDICTION 03
Embedded crypto crosses the chasm
Stablecoin and on/off-ramp providers (Brale, BVNK, Bastion, Crossmint, Fireblocks) appeared in the 2026 dataset for the first time. By 2027 we expect ≥10% of analyzed fintechs to disclose at least one stablecoin or embedded-crypto vendor — up from <2% today — driven by cross-border payouts and B2B settlement.
PREDICTION 04
Open finance keeps eating open banking
Open banking unlocked the current account; open finance unlocks pensions, investments, insurance, and tax. With US CFPB §1033, EU FIDA, and Brazil Open Finance Phase 4 all landing in 2026-2027, expect Financial Data Connectivity to be the most-reshuffled leaderboard — and a wave of consent-management entrants with no category last year.
with Open Banking Tracker · Embedded Finance Review · last updated 2026-10-15
01 · COLLECT
Evidence across nine signal categories
For each target we run nine parallel Exa web-search queries plus a direct fetch of the company’s trust / legal / footer pages, collecting subprocessor lists, customer announcements, regulatory disclosures, and script signatures.
02 · GRADE
Tier 1 / Tier 2 / Tier 3 evidence
Subprocessor and regulatory disclosures (Tier 1) outrank job posts and pricing inferences (Tier 2), which outrank structural integrations (Tier 3). The model is required to cite the source URL for every named vendor.
03 · SYNTHESIZE
Claude Opus 4.7 + canonical backstop
Opus 4.7 maps evidence onto a 12-layer embedded finance schema. A canonical vendor-to-slot mapping ensures known providers (Marqeta → card issuing, Check → payroll, Cover Genius → insurance) always land in the correct block.
Conflicts & curation
What a number does — and doesn’t — mean
Publisher in the ranking. Apideck publishes this index and ranks in it; the author is Apideck’s founder. We disclose rather than exclude it — weight Apideck’s own placement accordingly.
Two kinds of counts, blended. Most rows come from organic buyer-side disclosure. A few vendors that publish customer lists (Apideck, American Express Sync) were populated from those lists, which raises their absolute counts relative to vendors we only see from the buyer side. These rows carry a vendor-disclosed tag.
Editorial seed. A vendor counts as a “Provider” with ≥2 disclosed buyers or membership in a hand-curated canonical list. That seed is a judgment call, not a measurement.
“Provider” ≠ verified. Nothing here is independently audited. A higher number means more disclosed in our sources — not “bigger.”
Sourcing strategy
Where the evidence comes from — and where it doesn’t
Every datapoint in this report traces back to a public web URL. We don’t hold provider APIs, NDAs, or paid intelligence licences — the dataset is reconstructable by anyone with an Exa key and the source list below. The geographic shortfall in the Limitations block is a direct consequence of this sourcing strategy.
Covered surfaces
Trust / security / privacy pages. Subprocessor lists are the highest-fidelity source for KYC, fraud, and BaaS attribution.
Customer pages on vendor sites. Persona, Onfido, Jumio, Marqeta, Stripe, Adyen — every name lifted with a back-reference URL.
Regulatory footers + Member FDIC signage. Drives the sponsor-bank disclosure rate.
10-K / S-1 / 20-F filings. SEC + LSE + Euronext primary docs for public companies (Toast, Shopify, Affirm, ServiceTitan…).
Embedded Finance Review archive. 594 articles indexed by company & vendor; founder interviews and stack-story coverage feed Tier 2 evidence.
Open Banking Tracker registry. Provider taxonomy + regulated-entity records for the vendor side.
Press releases on PR Newswire / Business Wire / Globe Newswire. Quoted disclosures with date stamps.
Not covered (known gap)
Non-English press & trust pages. The Exa query layer runs English-only prompts; Spanish, Portuguese, Arabic, Mandarin, and Indian-language disclosures aren’t reached. Direct cause of the LATAM / MENA / APAC / Africa shortfall in the Limitations section.
Vendor-side APIs & private partner directories. No paid feeds.
NDAs, internal docs, leaked decks. Excluded by design.
Mobile-app SDK signatures. Not statically inspectable without binary analysis; out of scope for v1.
Card-network or scheme-side records. Visa / Mastercard partner disclosures aren’t public at the granularity needed.
How we’re closing the gap for 2027: a curated seed list of ~250 LATAM / APAC / MENA / Africa fintechs and embedded-finance vendors (drawn from Distrito, Finnovista, FinTech Saudi, ADGM RegLab, Disrupt Africa, Briter Bridges, and Wamda directories) is queued through the same analyzer that handles US/UK/EU companies. Localized-language sourcing (es-AR, pt-BR, ar-SA, zh-CN, hi-IN), mobile-binary stack inspection, and a regulator-side partner-attestation feed for non-US sponsor-bank coverage are on the roadmap. The 2027 edition will publish a per-region delta showing which gaps closed.
Open Banking Tracker maintains the most comprehensive public index of embedded finance, BaaS, open banking, and card-issuing providers worldwide. Each vendor surfaced in this report links to its full Open Banking Tracker profile — coverage areas, regulated entities, customer-disclosed integrations, and product taxonomy. openbankingtracker.com/embedded-finance.
Embedded Finance Review covers the embedded-finance ecosystem with original reporting, founder interviews, and weekly story round-ups. We indexed 594 Embedded Finance Review articles by company and vendor — every profile in this report surfaces relevant coverage in the “In the press” section where it exists. embeddedfinancereview.com.
Limitations
What this report is — and isn’t
Public web only. No NDAs, no provider APIs, no paid intelligence sources. If a vendor relationship isn’t disclosed publicly somewhere, it’s not in here.
English-language sources, US/UK/EU bias. LATAM, APAC, MENA, and Africa fintechs are underrepresented because their stack disclosures live in non-English press and trust pages we didn’t scrape.
Mentions ≠ market share. A leaderboard count is the number of analyzed companies that publicly disclose using a vendor. Stripe scores high in part because companies are willing to say they use Stripe. Adyen’s real share is likely higher than its count.
Point-in-time snapshot. Generated 2026-10-15. Vendor relationships churn — last quarter’s Synapse customer is this quarter’s Column customer.
Confidence varies per row. Subprocessor disclosures and regulatory footers (Tier 1) outrank press mentions (Tier 2) and structural integration signals (Tier 3). Per-profile confidence is captured in the underlying schema but not rolled up to the leaderboard view.
The compliance layer is genuinely opaque. See Section 02 — KYC, fraud, and sponsor-banking vendors disclose at fundamentally lower rates than payments and BaaS. Any insurance / KYC leaderboard built from public evidence will under-represent reality.
Dataset · 546 companies · 194 providers · 594 Embedded Finance Review articles indexed · generated 2026-10-15 “Providers” = vendor with ≥2 disclosed buyers in our dataset OR a recognized Open Banking Tracker provider (total mentions including singletons: 1038).
Section 09 · the landscape
300+ providers.16 categories. One map.
The Embedded Finance Landscape maps every major provider — BaaS, card issuing, payments, lending, payroll, compliance, and the data infrastructure underneath — into a single quadrant view. Published by the Open Banking Tracker team, refreshed quarterly. The card below shows 174 highlighted providers across the 12 categories most relevant to this report; view the full 300+ provider, 16-category map on Open Banking Tracker ↗.
PUBLISHED BY · OPEN BANKING TRACKER BY APIDECK
The Embedded Finance Landscape — 2026
300+ providers across 16 categories — BaaS, card issuing, payments, lending, compliance, and more
The first wave of embedded finance built APIs for humans operating software — accept payments, issue cards, originate loans from inside a SaaS. The next wave is the same layer being consumed by AI agents acting on a user’s behalf. The fintechs that ship agent-readable surfaces now — MCP servers, function-calling toolkits, structured product docs — set the defaults for who agents reach for when they need to move money, verify identity, or fetch a balance.
Stripe Agent Toolkit
function-calling SDK
Stripe
Toolkit + middleware for AI agents to call Stripe APIs safely (charges, customers, products) with built-in idempotency. Distributed via NPM and Python; integrated with the OpenAI SDK and Anthropic SDK.
Plaid MCP server
MCP server
Plaid
Plaid published an MCP server exposing account, transaction, and identity endpoints over the Model Context Protocol. Lets agents read user-permissioned bank data without learning Plaid-specific glue.
Adyen MCP server
MCP server
Adyen
Open-source MCP server (adyen-mcp) exposing the Checkout API (payment sessions, links, refunds, cancellations) and Management API (merchant accounts, terminals, webhooks, credentials) as callable functions — an agent can act on a natural-language instruction like "refund order #12345" without hand-written glue code.
Brex AI / Brex Assistant
in-product agent
Brex
Conversational agent layered on Brex's expense + treasury data. Pulls policy rules, transactions, and approvals so finance teams can resolve exceptions without leaving the dashboard.
Ramp Copilot
in-product agent
Ramp
Copilot surface over Ramp's spend data — drafts approvals, flags policy violations, summarises spend by vendor or department on request. The agent has read/write access to the user's workspace.
Mercury AI
in-product agent
Mercury
Agent layer over Mercury's banking, treasury, and bill-pay data for startup finance teams. Composes payments, surfaces anomalies, and answers natural-language queries against the ledger.
Agentic Business Accounts
in-product agent
Airwallex
AI agents operating inside Airwallex's own account infrastructure under human-set guardrails, framed around an "observe → decide → act → reconcile" loop. The expense policy agent enforces spend rules across entities, currencies, and languages; other agents handle multi-entity cash management and supplier invoices. Announced 2026-09-30 by co-founder/CEO Jack Zhang.
Klarna AI assistant
consumer agent
Klarna
OpenAI-powered customer-service and shopping assistant. Klarna disclosed in 2024 that the assistant resolved the workload of ~700 full-time agents; by Q3 2025 the company put the figure at more than 850 — one of the largest production deployments of agentic finance to date.
Backbase AI
bank-tech vendor platform
Backbase
Agentic AI layer for incumbent banks running on Backbase's Engagement Banking Platform. Modular agents embed into service and sales journeys, governed by AI-act-compliant controls. Distributed via ~120+ bank implementations (ABN AMRO, Citizens, RBC).
Dolfin AR/AP Agents
white-label agent
Dolfin
Autonomous accounts-receivable and accounts-payable agents — create invoices, chase payments, pay bills, reconcile, flag a cash gap or an overdue customer — white-labelled into other platforms' own products (vertical SaaS, fintechs, banks, PSPs) rather than run inside Dolfin's own app, so the platform's SMB end customers never leave it. Acts within guardrails each SMB sets.
Proof point
This report is agent-readable.
The full report ships as plain-text Markdown at /llms-full.txt and a short index at /llms.txt following the llmstxt.org convention. ChatGPT, Perplexity, and Claude can ingest the numbers without rendering JS or fetching dynamic endpoints. The denominators, citations, and methodology mirror the HTML report exactly.
Section 11 · ERP & accounting connectivity
Beyond accounts: open finance is open data
Open banking opened account data — Plaid, TrueLayer, Tink, Yapily. But money doesn’t move on bank data alone; it moves on accounting ledgers, payroll runs, ERP records, and tax-authority data. The unified-API tier — Apideck, Codat, Merge, Rutter, Finch — is the next leg of open finance, and it’s where the CFPB §1033 / EU FIDA / UK CMA regulatory wave actually lands.
Solved: open banking
GoCardless leads with 4 disclosed buyers (0.7% of the dataset).
1GoCardless
4·<1%
2Plaid
4·<1%
3Codat
3·<1%
4Rutter
2·<1%
5TrueLayer
2·<1%
6Finvu
2·<1%
Open: the rest of financial data
Accounting, ERP, payroll, FX — unified APIs are the connective tissue.
Xero leads accounting connectors (58 disclosed buyers), but the unified-API providers — Apideck, Codat, Merge, Rutter, Finch — are the abstraction layer above them. One API, multiple downstream systems (QuickBooks, Xero, NetSuite, Sage; ADP, Gusto, Workday; FX rates, payroll runs, KYC providers). The Plaid of accounting hasn’t emerged yet — the category is wide open.
The regulatory clockThree open-finance frameworks land in 2026-27: CFPB §1033 (US — extends data-access rights beyond bank accounts to accounting/payroll), UK CMA open-finance roadmap, and Brazil Open Finance Phase 4. A fourth, EU FIDA(Financial Data Access — full open-finance scope including investments, pensions, insurance), remains stalled in trilogue since June 2025, with realistic application closer to 2029-30. The category leaders for “open ERP / open accounting” will likely be decided before any of these rules fully bind — meaning the unified-API tier is racing to set the defaults regulators end up codifying.
Deep stacks
Fintechs that disclose the most
Twelve companies whose public surfaces name the most layers of their embedded-finance stack — out of 12 we track. These are the showcase reads: concrete examples journalists can point to when they need a named fintech instead of a leaderboard. Each card links to the full profile.
Each link opens a profile — founding story, funding, regulatory licenses, product timeline, recent developments, and the full stack analysis. Showing 462 of 546 analyzed fintechs (the subset with ≥1 publicly verifiable signal — disclosed vendor, Apollo record, or analyzer confidence).
Pre-baked one-liners for newsletter writers, analysts, and journalists. Each one already has its citation attached on copy. Free to use; attribution appreciated.
194 providers serve 546 fintech and embedded-finance companies — the full public stack analyzed in one place.
Stripe is named by 60 of 546 fintechs we analyzed — the most-disclosed PSP, 11.0% of the dataset.
More fintechs name their sponsor bank than name their KYC vendor: 52 vs 21. The compliance layer goes dark by design.
Financial Data Connectivity is the most-consolidated infrastructure layer: GoCardless leads with 4 disclosed buyers, Plaid is #2 with 4.
38 non-bank fintechs have crossed from a vendor relationship to owned banking, clearing, or card-issuing infrastructure — Square Financial Services, LendingClub Bank, Coinbase and dozens more.
Embedded finance is no longer a B2C phenomenon — roughly 16% of the companies we analyzed are vertical-SaaS platforms (restaurant, healthcare, salon, construction software) embedding payments, banking, or lending.
Outside the US, Adyen leads payments with 21 disclosed buyers — the most-named European PSP in the dataset.
First systematic source-cited census of embedded finance from the customer side, not the vendor side — every claim traceable to a subprocessor list, regulatory footer, or press disclosure.
Each copy includes the citation tail — Embedded Finance Index 2026, www.embeddedfinanceindex.com. Need a custom data cut, embargoed access, or interview? Email marketing@apideck.com.
For media
Press kit
Pre-cleared assets and quotes for journalists, analysts, and newsletter writers. Free to use with attribution.
QUOTE · OPEN FINANCE
“Open banking unlocked the current account. Open finance is unlocking everything else.”
Three regulatory clocks land in 2026-27: UK CMA’s open-finance roadmap, Brazil’s Open Finance Phase 4, and the US CFPB §1033 final rule. The EU’s FIDA framework remains stalled in trilogue, with realistic application closer to 2029-30. Expect the Financial Data Connectivity leaderboard to be the most reshuffled category between this report and next year’s.
QUOTE · DATA CONNECTIVITY
“Financial Data Connectivity is the fastest-moving infrastructure category in embedded finance — and the most consolidated.”
GoCardless takes 1% of all disclosures in the category by itself.The open-finance regulatory wave landing 2026-27 (CFPB §1033, UK CMA, Brazil Phase 4) — plus the EU’s FIDA framework, still stalled in trilogue with realistic application closer to 2029-30 — will shift the next leg: accounting, payroll-HRIS, FX, investments, toward unified-API providers (Apideck, Codat, Merge, Rutter, Finch). Expect this leaderboard to look unrecognizable in next year’s edition.
QUOTE · MARKET STRUCTURE
“Embedded finance has zero overlap between US, UK, and EU BaaS leaders. The stack is regional, not global.”
US: Sutton Bank, Evolve Bank & Trust, Cross River. UK: ClearBank, Modulr, Railsr. EU: Swan, Mangopay, Treezor. Three separate oligopolies that don’t compete with each other.
QUOTE · SCALE
“546 fintech stacks. 194providers.”
The first systematic, source-cited census of embedded finance from the customer side rather than the vendor side. Every claim is traceable to a subprocessor list, regulatory footer, or press disclosure.
CITATION
Apideck, Open Banking Tracker, Embedded Finance Review. Embedded Finance Index 2026. 2026-10-15. www.embeddedfinanceindex.com
TWEET / SHARE TEMPLATE
Embedded Finance Index 2026 — what 546 fintechs actually run on. 194 providers across BaaS, payments, lending, payroll, KYC, FX, data.
Journalists and analysts can request the full 546-row dataset (CSV + per-company JSON) for fact-checking and follow-up reporting. Embargoed and non-embargoed access available.
The Apideck research team is available for on-the-record commentary on embedded-finance market structure, BaaS consolidation, and compliance disclosure dynamics.
Data, analysis, and report design by Apideck, in partnership with Open Banking Tracker and Embedded Finance Review. Methodology + limitations documented above. Quarterly updates — the dataset refreshes throughout the year, not just at the annual edition.
We can comment on: embedded-finance market structure, vendor concentration, sponsor-bank disclosure rates, vertical-SaaS adoption, open-finance regulation (CFPB §1033, EU FIDA, UK CMA, Brazil), and the methodology behind the dataset. Custom data cuts and embargoed access available on request.
Want the deeper read?
The comprehensive report covers market sizing, use cases, vendor frameworks, ROI benchmarks, and a 50-operator practitioner survey.
