Standard Chartered is seeking to support markets growth without technology costs rising at the same pace. Common platforms, flexible computing capacity and AI-assisted development offer routes to greater efficiency, while local requirements and operational resilience add complexity.

"If you grow linearly with costs as you grow in revenues, then you're not really adding much value," said Gareth Hughes, Standard Chartered's chief information officer (CIO) for markets. Standard Chartered's approach centres on common global platforms, selective use of private and public cloud infrastructure and changes to how technology is developed and operated.

Hughes said approximately 90% of the bank's payments technology stack is used globally and that the vast majority of its markets business runs on global systems. Regulatory requirements, market connectivity and restrictions on data movement nevertheless continue to require local arrangements.

Global platforms accommodate local market requirements

Standard Chartered operates its order management, electronic trading, risk management and deal storage capabilities on global platforms, with local interfaces or separate deployments where necessary. "If we need specialisation, we specialise at the edge," Hughes said, describing an architecture in which common systems support the bank's operations while selected functions are adapted to individual markets.

In markets like India, the bank connects global applications to local application programming interfaces (APIs) or reporting endpoints to meet regulatory requirements. Data residency rules, which govern where information may be stored or processed, can require more substantial departures from the common architecture. In markets like Korea, Standard Chartered maintains an instance of one of its global trade booking and lifecycle platforms running locally to meet data sovereignty requirements, alongside local settlement platforms. These arrangements accommodate jurisdiction-specific requirements but entail separate infrastructure and maintenance obligations.

Hughes regards differences in market connectivity as a routine part of operating across jurisdictions. The bank can connect directly to trading venues or use third-party providers to support integration, depending on the market. Its emphasis remains on maintaining common core systems while limiting the number of separate applications and infrastructure instances needed to serve individual markets.

Matching computing capacity to demand

Standard Chartered uses its own private-cloud infrastructure alongside public-cloud providers, with decisions influenced by workload characteristics, data restrictions and computing costs. Hughes said the bank's private cloud has increased its available computing capacity tenfold compared with its previous infrastructure. The bank has also incorporated redundancy within its data centres and geographically separate recovery arrangements to maintain service availability when components fail.

"The bigger challenge with the cloud is often the data," Hughes said. Restrictions on where information can reside and the permissions required across jurisdictions can complicate public-cloud use. For some applications, Standard Chartered therefore favours infrastructure operated within its own data centres. Computing costs and the availability of capacity also influence that choice.

Public-cloud services are used for workloads requiring substantial processing power over limited periods. Hughes described the bank's overnight market and credit risk calculations, which involve running complex scenarios across its portfolio of trades. "We do a huge amount of overnight calculations, and then we give the compute back," he said. Standard Chartered uses multiple public-cloud providers for these tasks and has designed its risk infrastructure to support switching between them.

Resilience requirements can limit the scope for consolidating infrastructure and providers. Hughes said Standard Chartered seeks to identify single points of failure and incorporate redundancy into its technology architecture. Within its private cloud, equipment failures can trigger failover to separate infrastructure with independent power and cooling, while a second data centre provides geographically separate recovery capacity. Its use of multiple public-cloud providers similarly forms part of its approach to limiting concentration risk.

External market connectivity introduces dependencies the bank cannot manage entirely through its own infrastructure. Brokers, trading venues and connectivity providers may experience disruptions affecting several institutions simultaneously. Standard Chartered maintains connections through multiple brokers and multi-dealer platforms to preserve alternatives when individual services become unavailable, although the effectiveness of those arrangements depends on the nature and extent of a disruption.

UK financial regulators began overseeing four designated critical third-party providers in July 2026, including Amazon Web Services, Google Cloud, Microsoft and Oracle. The regime reflects the potential for disruption at a shared provider to affect multiple financial institutions.

AI changes engineering without removing oversight

Standard Chartered's AI Factory provides engineering teams with access to multiple artificial intelligence (AI) models through a common platform with centrally managed safeguards. Hughes said this reduces the need for individual teams to develop their own model infrastructure and controls. Moving applications into production nevertheless requires suitable data access and evidence that they can handle incomplete or poorly defined inputs. He identified evaluation, regression testing, which checks whether changes introduce new errors, and reproducibility as important requirements for assessing reliability beyond a limited prototype.

Hughes said Standard Chartered's current AI applications are concentrated largely in engineering and technology support. Teams use AI to identify operational issues and potential gaps in technology controls, including test coverage and documentation. He described how quality assurance staff have developed tools to check software submissions against the bank's control requirements, flagging potential violations during development. The bank is also moving towards specification-driven workflows, in which AI assists with defining requirements, preparing system designs, generating code and developing tests. According to Hughes, engineers review the outputs at each stage, while product teams participate in defining and validating requirements.

Faster code generation does not necessarily shorten the development process, as testing, approval and deployment remain subject to controls. Google's September 2025 State of AI-assisted Software Development report found that greater AI adoption was associated with higher software delivery throughput but lower delivery stability. These associations do not establish that AI directly causes either outcome.

Governance depends on how an application will be used. Hughes distinguished between AI summarising documents with citations and systems recommending business decisions, which require more extensive scrutiny. Hughes said Standard Chartered's approach to agentic AI, in which software carries out sequences of tasks with some autonomy, continues to involve human oversight. "It supports decision-making, it doesn't do decision-making," he said. Applications making decisions more independently would require further evaluation, governance and auditability.

Hughes also linked modernisation to cybersecurity risks, noting that the time between the emergence of software vulnerabilities and their potential exploitation has shortened.

Separately, Standard Chartered uses an observability platform, which monitors system performance and operational events, and is applying AI to identify possible problems earlier. Hughes described a move towards site reliability engineering (SRE), which uses monitoring, automation and engineering practices to maintain service availability. The bank is seeking to identify emerging problems before they result in failures while maintaining systems that support trading operations.

Hughes expects some of these changes to extend into client-facing workflows. One application under consideration involves converting requests for quotes (RFQs) received by email into structured documentation and pricing requests, reducing manual processing. He anticipates further changes in client responsiveness and personalisation over the following six to 18 months, although the proposed workflows remain subject to testing and human oversight before wider deployment.

Faster markets activity places greater demands on processing controls

Standard Chartered is preparing its markets technology for increasing transaction volumes, greater use of digital assets and the possible emergence of commerce conducted by autonomous artificial intelligence (AI) agents.Hughes expects these developments to place additional demands on computing capacity, processing reliability and transaction controls.

Hughes expects increasing digital-asset activity and more automated transactions to require greater computing capacity and tighter controls throughout Standard Chartered's markets technology. He identified the prospect of T+0 settlement, where a trade settles on the day it is executed, as one development that could reduce the time available to detect and correct processing errors. Ensuring that payment instructions are accurate becomes more consequential as transaction cycles shorten.

The constraints extend beyond a bank's internal systems. The Bank for International Settlements' June 2026 analysis of foreign exchange (FX) settlement risk found that some existing payment-versus-payment arrangements, which coordinate the exchange of two currencies to reduce settlement risk, cannot accommodate certain same-day FX trades because of their operating schedules and cut-off times. Faster processing within banks therefore does not necessarily remove the limitations of shared market infrastructure.

Hughes identified deployment frequency, change failure rates, operational incidents and the relationship between revenues and operating costs as measures of progress. Together, these indicators will show whether Standard Chartered can increase technology capacity and accelerate software delivery without weakening operational reliability. Their significance will depend on performance as transaction volumes and the complexity of markets activity increase.

The bank's ability to separate business growth from technology cost growth will also depend on factors outside its direct control, including local regulatory requirements, third-party services and market infrastructure. As transaction processing becomes more automated, the interaction between these dependencies and the bank's internal systems will become increasingly important in determining how efficiently and reliably its markets technology can scale.