Maximum Launches with $30M to Disrupt the Core Banking Oligopoly Using AI

For decades, a tight oligopoly of legacy software providers has dominated the banking industry, leaving financial institutions tethered to aging infrastructure. Now, serial fintech entrepreneur Randy Fernando is launching Maximum, a new startup armed with a $30 million seed round led by CRV. The company aims to replace these legacy systems with an AI-native operating system built from the ground up, rather than simply layering artificial intelligence on top of outdated architecture.

The challenge facing Maximum is monumental, given the deeply entrenched nature of existing core providers. According to a recent American Bankers Association survey of 679 banks, 76% rely on one of the "Big Three" legacy vendors: Fiserv, Jack Henry, and FIS. Fiserv alone holds core contracts with 42% of the respondents. Furthermore, bank loyalty - or inertia - is exceptionally high, with 68% of surveyed banks reporting they have used their current core provider for 10 years or more, and 69% stating they would likely stay even if they were unsatisfied.

While major vendors like Fiserv, FIS, and Jack Henry (which recently partnered with Google AI for cybersecurity) are rushing to integrate agentic AI into their existing frameworks, Maximum is taking a foundational approach. Fernando envisions a platform where banks can internally build custom AI agents for highly specific, unique tasks. "We call it an operating system versus a core because we do more than what a typical core banking service would provide," Fernando explained, noting that the system allows for dynamic, personalized customer products.

One practical application Maximum is developing is a suspicious activity report (SAR) agent. This tool scans the Department of the Treasury's OFAC list nightly, cross-referencing it with all daily bank transactions. If a match is found, the agent instantly alerts a compliance officer and recommends a course of action. This reduces a manual process that typically takes hours or days down to mere seconds, while keeping a "human in the loop" for final decisions.

Maximum is not alone in the race to automate compliance and banking operations. Other software providers like Backbase, nCino, and Mambu are incorporating agentic AI into their offerings. Meanwhile, a crowded field of tech providers - including Oracle, Nasdaq Verafin, ComplyAdvantage, WorkFusion, Castellum AI, Fenergo, and Finzly - already offer AI agents specifically for AML compliance and OFAC screening.

Maximum marks Fernando's third major venture in the banking technology space. His first startup, an automatic retirement investing platform called Vault, was acquired by the investment app Acorns in 2017 for an undisclosed sum. He subsequently founded Power Finance, a cloud-based card issuer that achieved a massive exit in 2023 when it was purchased by Marqueta for $275 million in cash and $25 million in equity.

Following the acquisition, Fernando served as Marqueta's vice president of credit products for two years before stepping down to build Maximum. He acknowledges that displacing legacy cores will be a decade-long battle, but notes that modern banking companies have spent the last five to six years successfully chipping away at parts of the core banking stack.

Value is accruing 'up the stack.' Incumbents have historically provided utility to customers as the underlying system of record. Customers are demanding that software goes beyond existing as the system of record and instead 'do the work' as the system of action.

- Tom Seo, Founder, Inverted Capital

The true barrier to Maximum's success will not be its technological capability, but rather the strict regulatory environment governing financial ledgers. Current core banking systems operate on deterministic logic - if X happens, Y is the guaranteed result. This predictability is exactly what regulators require to audit transactions and ensure compliance. AI-based reasoning engines, however, are inherently probabilistic, raising severe red flags when placed too close to a bank's balance sheet.

If an AI-native core executes a transaction, the bank must be able to explain exactly why it happened. As JT Thykattil from Forrester Research pointed out, a truly AI-native system learns continuously from outcomes, such as a declined loan or a false fraud flag. While this creates operational efficiency in areas like credit underwriting for thin-file consumers or loan document processing, it makes the core system a moving target for auditors.

For Maximum to capture meaningful market share from the Big Three, it must prove that its "system of action" can operate with the absolute, unyielding certainty of a traditional ledger. Until regulators are comfortable with probabilistic logic handling core deposits, Maximum's most viable path forward may be serving as a highly advanced orchestration layer for mid-sized banks, rather than a total replacement for the foundational ledger.