This article is based on reporting originally published by PYMNTS.com, drawing on their Digital Banks And The Power Of The Cloud Tracker® produced in collaboration with NuoDB.
The Scale of the Legacy Problem
Banking technology moves slowly — sometimes by design, often by inertia. According to PYMNTS reporting, a significant share of financial institutions were still only beginning to adopt cloud computing solutions as recently as 2020, and the majority of the world’s largest banks continued to rely on physical mainframes running architectures designed in the previous century.
The reasons are well understood inside the industry: regulatory obligations, data residency requirements, decades of accumulated technical debt, and a reasonable caution about moving mission-critical systems. But those justifications are increasingly difficult to defend as digital-native competitors demonstrate what cloud-first banking actually looks like at scale.
Why Challengers Have the Upper Hand
Challenger banks such as Monzo and Revolut were built on cloud infrastructure from day one. They have no mainframes to decommission, no batch-processing windows to work around, and no organisational inertia tied to legacy vendor contracts. According to the PYMNTS Tracker, many FinTechs and digital- or mobile-only banks are following the same model, using cloud platforms to deliver core banking services with a responsiveness that traditional institutions struggle to match.
The competitive urgency this creates for legacy banks is real. Customers increasingly expect instant decisions, 24/7 availability, and seamless digital experiences. Branches are secondary. Digital is the default. For a bank still hunting through distributed mainframe records to answer a customer query, that expectation gap is a material business risk.
What “The Data Hunt” Actually Costs
One of the more concrete operational problems the PYMNTS Tracker identifies is what practitioners call the “data hunt” — the process of locating and reconciling customer or transaction data spread across siloed mainframe systems. This is not a theoretical inconvenience. It directly affects the speed at which a bank can respond to customer requests, investigate disputes, or model risk.
Cloud-based data platforms consolidate that information into accessible, queryable stores. BNP Paribas, cited in the Tracker, identified this as a key motivation for building out its private cloud capability — not just to modernise for modernisation’s sake, but to remove the operational friction of chasing data across legacy infrastructure while also creating a foundation for online banking innovation.
Blackrock, the asset management firm, took a different route, partnering with Microsoft Azure to gain both scale and an innovation platform. The examples illustrate that there is no single cloud migration playbook — the right approach depends on the institution’s size, regulatory environment, and strategic priorities.
Technology Partnerships Expanding What’s Possible
One pattern emerging across the industry is the use of targeted technology partnerships to accelerate cloud adoption without requiring a full-scale rip-and-replace of existing systems. The Tracker points to a partnership between banking software company Temenos and distributed SQL database company NuoDB as an illustrative example.
The collaboration allowed challenger banks to access Temenos banking software delivered via cloud, including tooling designed to handle customer queries rapidly and reduce infrastructure overhead. For institutions with limited capital to allocate to infrastructure, reducing that cost base frees budget for product development and customer-facing innovation — which is precisely where challenger banks want to compete.
Agility as an Operational Requirement, Not a Buzzword
In financial services, the word “agility” is often used loosely. The PYMNTS Tracker gives it a more operational definition, particularly in the context of unpredictable demand events. Ariff Kassam, then Chief Technology Officer at NuoDB, articulated the requirement clearly in comments to PYMNTS:
“Business processes and technical infrastructure must be agile enough to ensure a quick response to changing requirements. An agile infrastructure allows companies to accommodate surges or reductions in demand. A firm’s connected ecosystem should enable this agility, which means choosing technology partners that support it.”
Ariff Kassam, CTO, NuoDB (via PYMNTS)
The context here is significant. The original reporting coincided with the early months of the COVID-19 pandemic, when financial institutions faced sudden, massive shifts in transaction volumes — particularly around contactless payments and digital banking adoption. Infrastructure that cannot scale dynamically becomes a direct liability under those conditions. Cloud-native elastic scaling is not a feature; in that environment, it is a continuity requirement.
Key Takeaways for DevOps and Platform Teams in Financial Services
- Legacy mainframes are an organisational risk as much as a technical one. The inability to scale quickly or access data efficiently has measurable business consequences in competitive markets.
- Cloud migration strategies vary. Private cloud (BNP Paribas), hyperscaler partnerships (Blackrock/Azure), and SaaS delivery via vendor partnerships (Temenos/NuoDB) are all valid paths depending on context.
- Elastic scalability is a concrete operational advantage, not a marketing claim. Demand spikes — whether from a pandemic, a market event, or a product launch — expose the limits of fixed-capacity infrastructure immediately.
- Data architecture is central to the cloud case. Consolidating fragmented mainframe data into accessible cloud stores directly improves customer service response times and analytical capability.
- Challenger banks set the competitive baseline. Institutions evaluating cloud adoption should benchmark their target state against digital-native competitors, not against their own previous performance.
The original PYMNTS Digital Banks And The Power Of The Cloud Tracker® was produced in collaboration with NuoDB. Readers seeking the full data and case studies should refer directly to the PYMNTS source publication.