The financial services sector will be one of the leading verticals in terms of AI adoption, but that will also mean it could face key decisions on infrastructure investments ahead of other industries.
The financial services (FS) industry was undergoing a period of technological disruption well before the recent acceleration in Gen AI. Blockchain technology and cryptocurrencies represent a fundamental shift for FS. But disruption rarely happens in isolation; FS organizations are now contending with another wave of disruption but perhaps one that will offer more constructive and controllable change if embraced proactively.
AI data center use cases in financial services
The FS industry faces a lot of regulatory scrutiny but also adopts technologies that can improve processes and services because small gains in speed or efficiency can potentially yield large returns. According to IDC, FS is expected to spend the most on AI solutions up to 2028 and will account for more than 20% of all AI spending. According to research from NVIDIA, more than half of FS professionals are convinced that AI will play an important role in business success, and 91% of organizations are already assessing AI or deploying AI in production. Across the industry, Gen AI specifically is being evaluated or used in a variety of processes, from enhancing loan and credit risk assessment to managing regulatory compliance, detecting fraud, or enhancing customer service.
For example, the Visa Account Attack Intelligence (VAAI) Score uses Gen AI to evaluate more than 180 risk attributes in milliseconds and generate a score predicting the likelihood of a type of brute-force card fraud aided by bots. Visa also leverages Gen AI to combat card-testing fraud. The AI-powered VAAI Score has 6 times the fraud-detection features of previous models and has reduced the rate of false positives by 85%.
FS firms also see potential in Gen AI to enhance customer service and decision-making. Bank of America recently introduced an AI-powered virtual assistant, Erica, to provide customers with personalized financial guidance. Capital One is taking a similar approach with Eno, an AI-powered natural language SMS assistant.
Gen AI is also helping FS companies navigate a complex regulatory landscape. Compliance management software providers are embedding Gen AI and machine learning in their platforms to analyze regulatory rules, policies, and processes; and identify and assess compliance risks.
20%
of all AI spending up to 2028 will be led by financial services.
91%
of financial services organizations are already assessing or deploying AI.
85%
Visa’s AI-powered VAAI Score has reduced the rate of false positives in fraud detection.
Owned or outsourced data center infrastructure
As more and more FS organizations move from experimentation amid the hype phase of AI, to concrete use cases with tangible returns on investment, decisions will need to be taken on how much AI-supporting data center infrastructure needs to be owned vs outsourced to service providers.
According to another CIO survey from Gartner, on average, only 35% of AI capabilities will be built by internal IT teams. However, that may change over time as AI matures and organizations view specific AI use cases as business critical, necessitating more ‘owned’ and controlled investment.
There is a range of opportunities and challenges when it comes to investing in owned data center infrastructure/self-built AI vs outsourcing to a service provider. The positives include greater control over data – especially important when it comes to data sovereignty - and the ability to develop AI services closely aligned with specific business needs. Challenges include the considerable capital investment and scarcity of some accelerated compute infrastructure – especially GPUs – plus the risks of investing in bespoke systems in a fast-moving and nascent area.
But while the risks and costs of self-building AI services are tangible, it’s likely they will diminish over time. As more FS organizations see the value of investing in owned accelerated compute and AI capabilities, data center equipment suppliers, consultants, and the rest of the data center value chain will continue to improve the effectiveness of product development and support services.
AI imperatives in financial services
As FS organizations move from AI experimentation to practical use cases, they must decide whether to own or outsource AI-supporting data center infrastructure. This decision hinges on balancing control, cost, scalability, and compliance. To guide this decision, Vertiv highlights key imperatives for successful AI deployment in FS: transforming customer experiences, enhancing efficiency through automation and compliance, and building trust in AI applications.
| Imperative | Use cases | Data center requirement |
|---|---|---|
| Be transformative | Re-imagining the customer experience. | Cross-functional expertise. Holistic solution design. |
| Be efficient | Compliance management. Automated customer service. |
Extend value and function of existing systems. Close-coupled systems that minimize space and energy requirements. |
| Be first | Investment strategy decision support. | Designs that reduce time and risk. Pre-configured systems that reduce deployment times. |
| Be confident | Fraud detection. Risk Assessment. Enhanced forecasting |
Proven solutions supported by comprehensive services. |
| Be future-ready | Personalized financial services. Intelligent cashflow forecasting. AI-driven investment strategies. |
Interoperable, upgradable, and scalable solutions that can adapt to change. |
Key takeaways:
Transition to AI
data centers
FS organizations must decide between owned or outsourced AI data centers as they move from experimentation to tangible ROI.
Internal AI
development
Currently, only 35% of AI is built internally, but this may rise as AI becomes critical, necessitating more controlled investmen
Investment pros
and cons
Owning a data center offers control and tailored AI but requires high investment and scarce GPUs, but these challenges may lessen over time.