Discussion

Cloud-Native Basics: Part One

Paul Williams, Senior Solution Architect

How well can you run a 21st century financial institution on a 20th century mainframe?

For decades, the industry assumption was that a bank’s core banking platform, the critical command centre processing deposits, running the general ledger, and orchestrating daily transactions, was too complex and sensitive for the public cloud.

But conventional thinking is costing banks their competitive edge. Today, the question is no longer if banks can run in the cloud, but how fast they can get there.

This blog is the first in our Cloud-Native Basics series, designed to break down the fundamentals of modern cloud infrastructure. In this instalment, we’ll trace the journey from rigid mainframes to digital cores, dismantle cloud myths, and examine how Starling rewrote the rulebook with Engine.

The move from monoliths to digital cores

At a recent event, I heard a speaker make the bold statement that it is not possible for banks to run their core platform in the cloud. I found this strange, as at Starling we’ve been doing just that for the last decade - and doing it quite successfully!

For years, banking relied on monolithic mainframes. Lurking in an on-premises basement, running legacy code, these systems were secure but functionally stagnant, plagued by clunky overnight batch processing and high-risk, multi-month release cycles.

The turning point arrived with the rise of hyperscalers. By 2014, digital-first challengers like Starling emerged, building their cores from scratch. Starling became one of the UK’s first leading digital banks, running the production platform 100% on public cloud.

Today, we find that this approach is the new normal. Wide-scale migration is underway, and regulators are increasingly seeing cloud-native architecture as the gold standard for operational resilience.

Risk vs reality: debunking common cloud fears

Legacy institutions often hesitate due to three perceived risks: security and sovereignty, regulatory compliance, and vendor lock-in.

While valid challenges, these rarely turn out to be the actual bottlenecks. Public cloud security routinely outperforms on-premises infrastructure, regulators are increasingly comfortable with cloud deployment, and multi-cloud strategies solve lock-in.

The true barriers are often internal. The hardest part of a cloud transition isn't code. In fact, it’s shifting a siloed IT mindset into a fast-paced DevOps culture, and managing high exit costs from legacy vendor contracts.

Case study: Starling’s operational excellence

Starling is a success story for banking on the cloud.

As a fully digital, branchless bank, the cloud was a natural place for us to establish ourselves. We architected the platform from the ground up to take advantage of the unique agility inherent in the cloud, and that decision has paid off ever since.

It is these benefits that help us to deliver massive, balance sheet-verified operational advantages. Achieving 99.99% availability to provide far greater resilience than legacy setups, Starling is able to unlock continuous innovation by supporting dozens of software deployments every working day.

Crucially, because the system is designed specifically for the cloud rather than trapped in old batch-processing cycles, it completely eliminates the need for scheduled maintenance windows or weekend system shutdowns.

It was this cloud-driven agility that enabled us, when the pandemic arrived in 2020, to launch a range of features within a few days to help our customers, such as the Connected Card.

The Engine behind Starling’s success

To give the banking industry access to this capability, Starling created Engine: a complete, 100% cloud-native platform built by bankers, for bankers.

Engine directly addresses the legacy pain points that bleed IT budgets—like batch-based processes, fragmented vendor structures, and archaic change cycles that force teams to glue systems together rather than deploy actual value.

Engine delivers a full suite of retail and business banking products, driven by modern APIs running 24/7/365. Today, this architecture powers over six million global customers across five banks.

It began with Starling’s low cost-to-serve model, expanded to Romania’s greenfield Salt Bank, capturing 4% market share in year one, and Australia’s AMP Bank SME offering. Expansion is underway in Canada with Tangerine and New Zealand’s SBS Bank.

A final note

Leveraging the benefits of banking on the cloud is a core business strategy, enabling growth and efficiency at scale. The institutions embracing it today will define tomorrow's banking landscape.

Successful cloud modernisation relies on four strategic pillars:

Start small

The days of large-scale, risky migrations are over; begin with a discrete, deliverable product or greenfield brand and build outward from there. This will help you deliver early value and develop skills and knowledge.

Prioritise culture over code

Build a genuinely delivery-focused organisation where continuous change is embraced as an ongoing process and a competitive opportunity, not an isolated event or a threat.

Leverage proven infrastructure

Utilise pre-integrated platforms like Engine to manage day-to-day operations and core ledgers, freeing your internal talent to focus entirely on unique customer innovation and differentiators.

Scale incrementally

Use the cloud's elasticity to lower the cost of experimentation, allowing you to rapidly test new ideas and scale infrastructure smoothly alongside real business growth. This move from capital expenditure to operational expenditure is a key benefit of cloud migrations.

Contact us

Want to learn more?

Whether you’re launching a new digital bank or migrating from a legacy platform: let’s discuss how Engine can help. Share details about your requirements and we will be in touch. Are you looking for a Banking-as-a-Service solution in the UK? Contact Starling’s dedicated B2B team to find out more about the services provided.

For information about how we use your data please read our privacy notice.