In an era of geopolitical fragmentation and AI exuberance, the unglamorous can still matter. John Sharratt, Standard Chartered‘s global head of technology & infrastructure, has articulated a vision where resilience is built not through heroics, but through the deliberate pursuit of standardisation, commodity hardware, and the careful decoupling of software from ageing physical assets.
This narrative explores the key industry trends underpinning his strategy, drawing on referenced data to show why “boring” is the most exciting prize in banking technology today.
The search for a cure to legacy technical debt
The burden of technical debt in financial services is no longer just an IT concern; it has become a material financial exposure. Industry estimates suggest that financial institutions spend as much as 70 to 80% of their IT budgets maintaining legacy systems, leaving limited capacity for innovation.
This structural drag directly impacts cost-to-income ratios and limits revenue potential, as delays in launching new products or adjusting pricing in response to market shifts translate into missed opportunities and margin compression.
Sharratt’s approach at Standard Chartered tackles this head-on by distinguishing between hardware decay and software stability. He notes that software generally becomes more stable as it ages, while hardware failure rates increase.
The bank’s strategy, therefore, is to continuously modernise software while refreshing hardware according to its lifecycle, decoupling the two disciplines through virtualisation.
This pragmatic approach avoids the trap of treating technical debt as a monolithic problem to be solved by a single, disruptive overhaul. Instead, it is managed with the same discipline as other risks, focusing on targeted investments that deliver faster returns.
Why the cost of waiting has become unsustainable
Sharratt’s distinction between hardware decay and software stability is more urgent in 2026 than ever. Industry data shows that legacy maintenance costs are rising by 18–25% year-on-year, driven by three structural forces that have permanently altered the cost curve.
First, the developer exodus has reached a critical threshold: the average age of a COBOL developer is now 62, and specialised contractors command $180–$250 per hour, up from $120 in 2022.
Second, cybersecurity liabilities have monetised age—modern ransomware groups specifically target unpatched legacy modules, and a single breach of a legacy customer database in 2026 triggers mandatory breach notification costs averaging $4.8 million plus regulatory fines.
Third, compliance standards have shifted: the 2025 update to SOC 2 and ISO 27001 introduced “architectural obsolescence” as a reportable finding, forcing organisations to purchase custom support agreements from original vendors that rose 30% in 2026.
The average annual cost to maintain a single legacy enterprise application in 2026 now ranges from $850,000 to $3.2 million per system, depending on language stack and transaction volume. For large financial institutions with 20 or more legacy systems, total maintenance budgets average $22 million annually, up 40% since 2023. When indirect costs—talent premiums, productivity loss, compliance remediation, and integration workarounds—are included, the true cost is typically two to three times the visible line item.
Financial analysts have begun using “legacy tax” as a KPI. Publicly traded companies with a legacy maintenance ratio above 65% see a 12–15% discount on their forward P/E multiples, because investors recognise that every innovation dollar is being cannibalised by technical debt.
The implication is stark: technical debt is no longer an IT line item—it is a balance-sheet risk that directly affects shareholder value.
The legacy modernization market itself reflects this urgency. Valued at $25 billion to $30 billion in 2025, it is projected to reach $66 billion to $90 billion by 2031–2034, growing at 14.9% to 17.6% CAGR—faster than the overall software market. Mainframe modernization alone, a distinct subcategory at $9.01 billion, is forecast to reach $25.94 billion by 2035.
What was previously a multi-year commitment with uncertain ROI now has clearer timelines and measurable conversion accuracy, as organisations treat modernization as a prerequisite for cloud and AI adoption rather than a deferred initiative.
The geopolitical imperative: Resilience and sovereignty
Geopolitical tensions are fundamentally reshaping digital supply chain strategies for global banks. Regulators are increasingly concerned about the systemic risks arising from financial institutions’ reliance on a limited number of non-European IT service providers.
The Dutch central bank and financial regulator recently warned that this dependency amplifies concentration and systemic disruption risk, where a failure at a single provider could affect large segments of the sector. This concern is heightened by the risk that state actors could exploit these digital dependencies for political leverage.
Standard Chartered’s investments in true geo-resilience directly respond to this volatile landscape. By moving to a software-defined infrastructure and eliminating physical server and firewall appliances, the bank has enabled its operations to span geographic regions.
Sharratt notes that the bank is well placed to navigate these conditions, addressing risks many banks cannot meet. This is further reinforced by the bank’s long-term strategic commitment with Broadcom, using VMware Cloud Foundation (VCF) to establish a secure, resilient private cloud foundation across its 54 global markets. The goal is to ensure uninterrupted availability and operational consistency, staying ahead of evolving regulatory and security requirements.
The rise of “sovereign AI” and sovereign models extends this trend. As banks face stricter local laws and cybersecurity requirements, they are turning to sovereign clouds and onshoring digital infrastructure to ensure data stays within national borders and mitigate retaliatory trade measures. Standard Chartered, having already mastered sovereign hosting as a commodity capability, is positioning itself to apply the same principles to AI deployment.
Questioning “What does commodity AI look like?”
While the rest of the tech world chases the most advanced and expensive proprietary hardware, Sharratt’s team at Standard Chartered is asking a fundamentally different question: “What does commodity AI look like?”
The answer, he suggests, is quite boring. In many ways, it looks like a commodity server, simply with one or more GPUs added. This approach is not just about cost-cutting; it is a strategic bet on scalability and flexibility.
This line of thinking is surprisingly rare. Industry data shows that while interest in generative AI is universal, significant barriers to adoption remain, with GPU cost and availability as a leading concern.
The broader market is also moving towards standardisation, with data centres increasingly embracing the 21-inch rack format popularised by the Open Compute Project. Analysts forecast that this larger rack format, designed for better airflow and higher power densities needed for AI, will make up over 70% of kit shipped by 2030.
At Standard Chartered, the unit of scale for computing is the rack itself, pre-assembled and tested as a complete unit, arriving ready to integrate into their computing estate. This approach recognises that supply chains benefit most from predictability and standardisation, an insight that is becoming critical as AI workloads push single rack densities to 50kW and beyond.
For APAC financial services, this commodity approach to AI may prove even more vital, allowing institutions to scale without being locked into expensive, proprietary ecosystems.
