Insurers' Mainframes Are Now the Thing Holding Them Back
A new whitepaper from NTT DATA and AWS argues the systems that once guaranteed reliability are now blocking speed — and lays out how to fix them without ripping everything up.
The Brief
NTT DATA and AWS make the case that insurers can modernise legacy mainframes gradually rather than through a risky, all-at-once replacement. The playbook: stabilise the systems that must stay, decouple what slows the business down, and move the right workloads to cloud where data access, flexibility and AI-ready design pay off. Joint projects, they say, have cut mainframe running costs by 50–80% and sped modernisation by 30–50% — with a large European insurer expected to trim annual opex by more than a fifth by 2028.
For decades the mainframe was the insurer's badge of dependability — the machine that never blinked. Now, with costs climbing, data trapped in silos and AI-native rivals moving fast, that same bedrock has quietly become the anchor slowing everyone down.
That tension sits at the heart of a new whitepaper from NTT DATA, produced with AWS, which argues that insurers can no longer treat their core platforms as untouchable — but also don't have to gamble the business to change them.
The bedrock that turned into ballast
The report's starting point is that many insurers are still pouring money into simply keeping legacy systems alive, even as customer demands, regulation and cyber risk all intensify.
Making matters worse, tightly interlocked architectures, dependencies nobody documented and a shrinking pool of mainframe specialists can turn even a minor change into a long, costly project. The upshot is that the very systems built for stability now make routine progress slow and expensive — the opposite of what a fast-moving market rewards.
A technology problem that's really a business one
The whitepaper deliberately reframes modernisation as a commercial issue rather than an IT chore.
Faster product launches, sharper underwriting and more responsive claims all depend on architectures that can handle real-time data and AI-enabled workflows — something a rigid legacy core struggles to deliver. Framed that way, standing still isn't a neutral choice; it's a competitive cost that compounds while rivals build on more flexible foundations.
Insurers don't have to trade the reliability of the mainframe for the flexibility of the cloud — they can modernise at their own pace and keep both. — AWS, on de-risking the transition
The pragmatic, phased route
Instead of a "big bang" rip-and-replace, NTT DATA and AWS push a staged approach built on assessment, prioritisation and coexistence.
In practice that can mean leaving stable workloads on the mainframe while exposing their data and functions through APIs, shifting change-heavy applications to cloud, and leaning on automation and AI to take the risk and grind out of refactoring. The authors note that cloud and AI have reshaped the economics of this work — dependency mapping, code analysis and transformation are far more manageable than they used to be. Even so, they're clear that tooling alone won't do it: deep insurance domain knowledge and disciplined governance are still needed to protect policy data, business rules and regulatory standing.
What NTT DATA and AWS put on the table
The recurring theme is that insurers shouldn't have to pick between stability and agility — and the paper backs that with numbers.
NTT DATA says its AWS work helps clients modernise mainframe environments while safeguarding continuity, lowering cost and laying the groundwork for AI and analytics. Joint projects, it claims, have reduced mainframe operating costs by 50% to 80% and accelerated modernisation by 30% to 50% using AI-supported delivery frameworks and accelerators.
The sharpest proof point is a customer story: a German multinational financial services and insurance group modernised across four markets, lifting large-scale workloads off legacy systems and into AWS without interrupting operations. That programme is projected to cut annual operating expenses by more than 20% by 2028 and bring new products and capabilities to market roughly 30% faster.
Key takeaways
- The mainframe flipped from asset to obstacle. Systems prized for reliability now make change slow and costly as data siloes and AI rivals reshape insurance.
- Modernisation isn't all-or-nothing. NTT DATA and AWS argue for a phased path — stabilise, decouple and migrate selectively rather than a risky wholesale rebuild.
- It's a business case, not an IT one. Faster launches, better underwriting and responsive claims all hinge on real-time, AI-ready architectures.
- Numbers back the pitch. Joint projects report 50–80% lower mainframe costs and 30–50% faster modernisation via AI-supported delivery.
- Governance still matters. Technology alone isn't enough — domain expertise and discipline are needed to protect policy data, rules and regulatory confidence.