Data Sovereignty and AI Compliance in Financial Services

The collision between AI's appetite for computing power and the financial sector's stringent regulatory requirements is reshaping data centre and digital infrastructure design. 

With AI’s burgeoning importance for the finance sector, its rapid innovation will be a key focus of our upcoming breakfast roundtable in New York, where industry leaders will examine what’s next for AI-driven transformation in the industry.

While Goldman Sachs Research forecasts AI workloads will account for around 27% of data centre power demand by 2027, financial institutions face a more complex challenge than simply securing rack space. 

These organisations must architect environments where every kilowatt and every byte remains compliant with expanding data sovereignty finance regulations.

Infrastructure operators face regulatory scrutiny

Data centre operators serving financial services clients must meet stringent financial regulator requirements. 

For instance, the Digital Operational Resilience Act (DORA), which took effect across the European Union in January 2025, fundamentally altered the relationship between financial institutions and their infrastructure providers. 

DORA does not merely set standards for banks – it grants EU financial supervisory authorities direct oversight over critical ICT providers, potentially including some cloud, colocation and other digital infrastructure providers.

Many financial institutions are increasing the frequency of on‑site reviews and ensuring they have contractual rights to conduct technical site visits and audits of critical IT suppliers – including data centre providers – to verify everything from backup power systems to physical security.

For operators, this means maintaining audit-ready documentation and implementing monitoring systems that satisfy regulatory requirements around uptime, data protection and incident response.

The shift towards digital sovereignty is driving new infrastructure requirements. Financial regulations increasingly demand that sensitive data remain within specific jurisdictions, creating demand for data centres with verifiable data residency rules and technical controls to prevent unintentional cross-border transfers.

Power density meets compliance constraints

AI infrastructure presents unique challenges when layered atop financial regulations. High-performance computing for AI training now requires 50-100kW per rack: 10 times the power density of traditional financial services workloads. 

This necessitates advanced cooling systems, from rear-door heat exchangers to direct-to-chip liquid cooling, all while maintaining continuous uptime and security that financial regulations mandate.

"From hardware servers to software platforms, our expertise has helped our customers, particularly in the public sector and financial services industry, explore and leverage the power of sovereign AI," observes Paul Ju, Senior Vice President of Global ISG at ASUS.

The challenge intensifies when financial institutions attempt to leverage external AI capabilities. 

Many AI services require transmitting data to large language models hosted by major technology providers. 

This has created demand for regulated cloud solutions that combine local data residency with access to advanced computing capabilities.

Architecting compliant AI environments

Data centre operators are developing specialised offerings for regulated workloads. 

These feature dedicated private environments with enhanced physical security, segregated network architectures and granular access controls that satisfy regulatory requirements while supporting high-density AI workloads.

"We're not anti-AI – we're pro-control," explains Jacob Mellor, CTO of Iron Software. "Organisations shouldn't have to choose between AI-powered productivity and data sovereignty. Our solutions enable the benefits of intelligent software while ensuring customer data never leaves their controlled environment."

But infrastructure requirements extend beyond computing and cooling. 

Financial institutions need data centres that provide detailed energy usage data for sustainability reporting while simultaneously housing high-performance computing equipment necessary for AI workloads. 

Colocation providers are adapting contract structures to include provisions for continuous AI compliance monitoring, regular technical audits and explicit commitments around data residency.

The infrastructure build-out required is substantial. An estimated 10GW of new data centre capacity will break ground globally in 2025, according to JLL, with financial services representing a significant driver. However, power availability remains the primary bottleneck – Deloitte research indicates 72% of data centre executives identify grid capacity as extremely challenging for infrastructure development.

With the World Economic Forum projecting AI spending in financial services will approach US$97bn by 2027, data centre operators that can deliver compliant, AI-ready infrastructure will find themselves in a commanding position. 

Success requires technical excellence in power delivery and cooling alongside deep expertise in navigating the complex regulatory landscape governing where data can be processed and how it must be protected.

This topic will be part of the wider discussion on how Generative AI is reshaping enterprise content, customer communications and compliant financial services operations.

Register your interest to join senior financial services and technology leaders at Breakfast at Tiffany's: Reimagining Financial Services Through the Power of AI on 29 January 2026.