AI-assisted coding may be making software easier to develop, but getting it into banks remains a complex and costly process. These challenges are also reshaping the economics of specialist fintech providers, many of which face difficulties securing investment and achieving the scale required to serve major banking clients.
For United Fintech, the financial technology group backed by several major global banks, this creates opportunities for a shared infrastructure model that gives institutions access to specialist technology without developing and maintaining every capability independently.
“Software is becoming almost free to produce,” notes Christian Frahm, United Fintech’s Founder and CEO, talking to TradingTech Insight. “You still need somebody to spec it out and somebody who understands the domain, but the actual writing of the code is increasingly being done by agentic workflows. But that’s only a third of delivering a solution within a bank. Two-thirds is actually connectivity, data and compliance. So only a third of it is actually getting solved right now by AI within the bank.”The Cost of Bringing Technology into Banks
Banks have traditionally relied on internally developed systems and established technology vendors to support their trading operations. This can limit their ability to adopt newer technologies, particularly those developed by smaller specialist firms.
“If banks want to work with smaller tech firms, typically with fewer than 100 people, they have to spend two to three years onboarding them,” observes Frahm. “These companies are often really good at what they do, but how do you match that with a procurement process that takes two years? We’re now hearing from banks that they’re charging their internal teams real money to take on a new software vendor. So banks are limited in what they can work with. They can either use the large providers or build it themselves.”
United Fintech addresses this problem by bringing specialist technology providers together under a common platform. The group has acquired businesses and is also open to partnerships, providing centralised sales, marketing, finance and operational support while allowing individual product teams to concentrate on their areas of expertise. Banks can access multiple solutions through a single master service agreement, with United Fintech responsible for ensuring that the businesses within its platform are appropriately capitalised and managed.
One example is Cobalt, the FX credit technology business founded by former Citi executive Andy Coyne and Darren Coote. Following its acquisition by United Fintech, Cobalt has expanded from serving a relatively small number of major banks to becoming, according to Frahm, a solution used by most of the world’s largest banking institutions.
The Case for Shared Infrastructure
While proprietary trading strategies, pricing models and other differentiating capabilities may justify internal development, much of the underlying infrastructure serves common requirements across institutions. Frahm sees opportunities for banks to share technology in areas such as connectivity and credit systems, reducing duplicated development and maintenance costs.
However, banks may be reluctant to rely on infrastructure controlled by a competitor. United Fintech seeks to address this through a neutral platform that allows institutions to access shared technology without ceding control to another bank. Although five major global banks are investors, they remain minority shareholders, sharing two rotating board seats, while Frahm and the company’s leadership retain three.
The group says it now provides technology to 11 of the world’s 12 largest banks, alongside approximately 100 other financial institutions, including smaller banks and exchanges. CME is among its publicly identified customers.
While the model offers banks a way to access common infrastructure while retaining control over their own technology strategies, governance and operational resilience remain important considerations as shared solutions become more widely adopted.
Measuring the Returns from AI
United Fintech has adopted what Frahm describes as an AI-first operating model, using agentic workflows across its business and increasingly applying AI to develop new products. For established financial institutions, however, the financial benefits of AI investment remain difficult to quantify.
“Everybody’s talking about more efficiency, four or five times greater productivity from developers, but it’s not like they’ve got rid of a lot of developers. It’s not like they’ve taken out headcount to get close to that,” says Frahm. “So we’re in an interesting period where banks are seeing perceived value and perceived efficiencies. They’re also seeing token costs go up quite dramatically, and they’re struggling to measure the effect it’s having.”
Frahm says United Fintech has been able to assess AI’s financial impact partly through its hiring requirements. As a growing business, it can compare planned recruitment against actual headcount, alongside revenue performance. He says the company is operating significantly below its budgeted staffing levels while exceeding revenue expectations.
Autonomous AI also presents operational challenges. Frahm says banks remain cautious about allowing agents to operate within production environments, where errors or unintended actions could affect critical financial workflows.
Consolidation and the Changing Vendor Landscape
The changing economics of software development are coinciding with a shift in investment priorities, adding to the commercial pressures facing specialist capital markets technology providers.
“The capital markets space is consolidating for sure,” says Frahm. “There’s so much money going into AI that investors are looking for extraordinary returns, and the funding environment for capital markets technology businesses has become much harder. Banks are still funding it, but the sector also needs private equity and venture capital, and right now the party for them is somewhere else. It’s difficult to achieve scale in capital markets software, and bank sales cycles are getting longer, with less appetite to onboard new vendors or spend money with software providers.”
These conditions could create opportunities for larger technology groups to acquire specialist businesses with established products but limited resources to expand their customer base. United Fintech intends to continue pursuing acquisitions alongside its own AI-driven product development, while working with banking clients to identify common technology requirements.
As development costs fall and funding pressures encourage consolidation, how much of their technology should banks continue to build and maintain themselves, and how much could be sourced more efficiently through shared infrastructure and specialist providers? The answer will depend on where institutions see genuine competitive advantage, and how they balance the benefits of collaboration against their requirements for control, resilience and flexibility.
Subscribe to our newsletter
