EN
Back to the archive

The archive · Money & Fintech · Financial decision · 2013–2026

Konfío's data-credit bet: 80K Mexican SMEs financed, now chasing a bank license

Mexico's SME lender bet machine learning on invoice data can underwrite businesses banks reject; ~MXN$100B disbursed since 2013.

Konfío

The betThat tax-invoice data plus machine learning can profitably underwrite Mexico's credit-invisible SMEs — and that a proven portfolio then justifies scaling into a bank.Scaling

What the business is

A Mexican fintech that lends working capital, business credit cards, and payment-terminal financing to small and medium enterprises that traditional banks turn away.

Starting capitalAbout US$235M in equity through the 2021 Series E, plus debt facilities; total funding of ~US$1.1B across 13 rounds

How it started

Founded in 2013 by David Arana, an ex-banker, and Francisco Padilla, an engineer, after they saw Mexico's structural gap: 99% of Mexican companies are SMEs and most cannot get formal credit. Their first loans capped around MXN$150,000 to micro-businesses that banks would not touch.

What happened

Konfío grew through Kaszek, QED, and other backers, and in September 2021 became Mexico's fourth unicorn with a US$1.3B valuation after a US$110M Series E extension that included SoftBank. Over a decade it disbursed nearly MXN$100B to ~80,000 companies, with more than 80% of clients getting their first business credit through the platform; management says the company crossed profitability in the last two years.

How it ended up

Still running and scaling: Konfío has applied for a bank license (CNBV, filed 2023) to cut funding costs and offer deposits, payments, and treasury products, and projects MXN$44B in new SME credit and 85,000 more companies between 2026 and 2028.

Background

Konfío was founded in 2013 by David Arana and Francisco Padilla on a simple observation: Mexican banks require roughly two years of operating history and collateral, so most of Mexico's small businesses — 99% of the country's companies — could not borrow at all. Their bet was that machine-learning models reading a business's SAT tax-invoice data could assess risk where banks saw none, and that this segment could be lent to profitably.

The strategy produced steady scale: nearly MXN$100 billion in cumulative disbursements to about 80,000 businesses over ten years, with MXN$17 billion+ disbursed in 2024–25 alone. More than 80% of Konfío's clients were receiving their first-ever business credit. In September 2021 the company became Mexico's fourth unicorn, valued at US$1.3 billion after a US$110M Series E extension with SoftBank participation.

The data thesis has now matured into a regulated-banking bet. Konfío applied for a CNBV banking license (filed 2023, in final review) to take deposits and cut funding costs, and projects MXN$44 billion in new SME credit and 85,000 additional companies between 2026 and 2028. Management says the company has been profitable for the last two years — the evidence it cites for why it deserves a bank charter.

What has to be true

  • Alternative data (SAT invoices) solved the information problem that made SMEs unbankable, without waiting for a formal credit bureau history to exist.
  • The wedge was narrow — working-capital loans to the smallest businesses — which banks ignored and large fintechs overlooked.
  • Each loan generated more behavioral data, improving underwriting and collection and creating a widening moat.
  • Profitability before the banking license gave the company leverage: the license is an expansion, not a rescue.
  • The model's risk is concentration in one regulatory and macroeconomic environment, which the bank-license move is designed to diversify.

What can be applied

When a segment is unbankable under traditional rules, underwrite against an alternative signal already produced — here, tax invoices. A performing portfolio then justifies cheaper funding.

Aftermath

As of September 2026 Konfío is live, profitable, and in the final regulatory stage for a Mexican banking license. It serves businesses in all 32 Mexican states without physical branches, has financed ~80,000 companies with nearly MXN$100B, and is targeting MXN$44B in additional credit plus 85,000 new borrowers by 2028, positioning itself as the eighth-largest SME lending portfolio in Mexico.

Sources

spotted an error? The archive wants to know.

Your turn

You just read one. Describe what you are building, and see who is betting on the same thing.

Free account · 3 free questions · no card

Related cases