Two-Thirds of 1.46B Users Lack Licenses: Digital Banking's Silent Crisis

Woofun AI reports that Francesco Andreoli, Head of Developer Relations at Consensys, in collaboration with Chopper of Foresight News, has exposed a critical paradox in the digital banking sector as of July 2026. While the industry boasts massive user growth, a rigorous audit reveals that the foundational stability of these institutions is severely compromised by licensing gaps and hidden infrastructure dependencies.

The scale of the digital banking ecosystem is undeniable, yet its geographical distribution defies Western-centric assumptions. The 368 active digital banks tracked serve a combined total of 1.46 billion users, figures derived from actual reported customer bases rather than market projections. Asia dominates this landscape, accounting for 817 million of these users. WeBank alone commands over 400 million users, a figure that exceeds the total user base of all digital banks in the United States and Europe combined. Nubank holds 131 million customers, surpassing the entire U.S. digital bank sector, while Revolut, a leading European entity, serves only 50 million users, representing a minor fraction of the global total.

A structural shift toward Web3 technologies is evident in the newer generation of institutions. Among digital banks founded after 2020 and still operating, 30% are native Web3 self-hosted applications that do not hold user funds directly. In contrast, only 4% of institutions founded in the 2010s share this characteristic. This divergence indicates that builders are actively choosing decentralized architectures. The current ecosystem is categorized into three distinct groups: 254 traditional challenger banks, 58 fiat-crypto hybrid platforms, and 56 native Web3 platforms. These entities rely on a complex support network comprising 106 infrastructure service providers and 219 investment firms.

The most critical vulnerability lies in the regulatory status of these entities. Of the 368 institutions, only 127 hold complete banking licenses. Consequently, two-thirds of the apps marketed as "banks" lack proper banking qualifications, operating instead under partner bank permissions, electronic money licenses, or obscure card-issuing arrangements. Most customers remain unaware of this distinction, which constitutes a core structural risk. This fragility has already manifested in historical precedents: in 2018, WaveCrest faced Visa’s withdrawal of support, shutting down dozens of crypto bank cards overnight; in 2020, Wirecard’s €1.9 billion funding gap froze digital banking services across Europe; in 2024, the collapse of Synapse, a Bank as a Service provider, revealed that "FDIC protection" was illusory due to flawed fund ownership tracking; and in 2026, Ready repeated this pattern of failure.

Unlike traditional bank bankruptcies where deposit insurance covers funds, the collapse of digital bank infrastructure leaves customers waiting in liquidation lines. The nature of these failures is often silent, lacking the dramatic public spectacle of events like the FTX collapse. This month alone, five institutions were removed from the active list due to liquidation, acquisition, or quiet transformation. There were no press releases or post-mortem analyses. Instead, apps stopped updating, customer service ceased responding, and websites redirected to partner sites. Hundreds of thousands of customers either transferred their funds or lost them permanently, with no media coverage to document the exits.

The disparity between marketing claims and operational reality is stark, particularly regarding artificial intelligence. While almost every digital bank’s financing pitch deck highlights AI, a verification of regulatory disclosures and deployed products reveals a different picture. Only 67 platforms have achieved large-scale AI implementation, accounting for 18% of the sector. The remaining 300+ platforms are either in pilot stages, merely "exploring research and development," or falsely claiming proprietary technology while relying on partners. This gap suggests that AI is currently more of a fundraising narrative than a widespread operational tool.

Notably, the pioneers of AI adoption are not established Western giants but credit institutions in emerging markets such as Nigeria, the Philippines, Mexico, and Bangladesh. In these regions, underdeveloped credit reporting systems necessitate the use of models to grant credit to individuals without formal histories. For these institutions, AI is not a trend but a survival mechanism. While Western markets debate the concept of AI banks, the Global South has already achieved large-scale implementation, driven by the urgent need to serve unbanked populations and maintain operational viability.

The ecosystem’s infrastructure layer presents significant concentration risks. Just 106 service providers support the 368 consumer-facing brands, with a few partner banks, BaaS platforms, and card processing services underpinning dozens of higher-level brands. This high degree of concentration is invisible to consumers but creates a single point of failure. A crisis at any of these upstream providers could trigger a cascade of failures across the sector, similar to the Synapse incident. The system is built on underlying service providers of whom the public knows little, yet two-thirds of these platforms would struggle to survive if their partners faced distress.

Woofun AI data shows that the industry is experiencing transformative growth, with 1.5 billion people accessing banking services via mobile phones, many for the first time.

However, this expansion is built on a fragile foundation. The reliance on unlicensed entities and concentrated infrastructure providers means that the sector is highly susceptible to systemic shocks. The lack of transparency regarding these dependencies exacerbates the risk, as customers and investors alike are often misled by the polished interfaces of consumer-facing apps.

Looking ahead, three key trends are expected to reshape the industry. First, the licensing gap will close from both ends by 2027, with stronger unlicensed platforms acquiring or applying for licenses while weaker ones exit the market. Second, the first major failure of an AI credit model is anticipated during the next credit cycle. Of the 67 models with large-scale implementation, most have not yet faced full economic downturn pressures, and some teams will encounter risk scenarios not covered in their training data. Finally, the next generation of financial service customers will be AI intelligent agents. Currently, only 7 companies are developing financial infrastructure for these agents, including wallet management and machine-to-machine payments. This landscape mirrors the Web3 sector in 2021, characterized by few participants and niche models but holding long-term structural opportunities.