Generative AI is revolutionizing banking, finance, and insurance. And only those who modernize fast will lead the charge.
The pace of innovation in the financial services sector has been unrelenting for the last decade. Yet the rate of change is destined to accelerate sharply due to generative AI (GenAI), graphics processing units (GPUs), and the advent of the AI data center. Anyone that depends on IT infrastructure as a point of competitive differentiation in banking, credit unions, hedge funds, insurance firms, financial media, and other financial services had better strap in for a period of heavy disruption over the next year or two.
Some industry leaders are already pushing ahead. Wells Fargo’s AI assistant, for example, has delivered more than 245 million interactions – all without any human handoffs and with no exposure of sensitive information. Others such as JPMorganChase, Bank of America, and DBS are all forging ahead with bold GenAI strategies and gaining familiarity with the many large language models (LLMs) that are available.
Disruption is inevitable
As many financial services companies operate their own data centers, and those using colocation providers often build their own cages, a period of upgrades and modernization is inevitable. Banking, financial services, and insurance organizations that can harness AI faster and more effectively are likely to dominate in the coming years.
But AI that requires a level of rack and data center density that has never been seen before. GenAI and GPUs can’t run on the 5 kW racks that have been the norm in traditional data centers. Racks of 25 kW, 50 kW, and beyond 100 kW are starting to appear. Those in financial services must quickly gain familiarity with compute densities of this magnitude as they prepare for the AI era.
An AI optimized data center, also known as an AI Factory, necessitates a complete overhaul of power and cooling infrastructure. Once density rises beyond a particular point, air cooling alone will no longer be enough. Liquid cooling must be introduced. The supporting energy infrastructure both inside and outside of the data center must be able to supply AI servers with the power they need. This transformation must extend across the entire financial services network – head office, branches, ATMs, remote access, and mobile services.
All of this must be accomplished while holding firm to industry standards for compliance and security. Already, rampant and unchecked usage of public (and therefore potentially insecure) GenAI apps like ChatGPT are commonplace in the enterprise. A BigID study reveals that 94% of workplaces admit to having no AI security strategy, almost 70% see AI data leaks by employees as their top security concern, and nearly half have no AI security controls in place. In financial services, less than 40% have deployed AI data protection measures.
Partner for AI Success
IT innovation has never been so rapid. New server technology, GPUs, AI models and applications, highly dense racks, next-generation cybersecurity and compliance tools. Each of these elements is required by an AI data center in financial services – and all of it all must be supported by paradigm shifts in power and cooling infrastructure. No one company can do it alone. Partnerships are essential for success.
Those in banking, finance, and insurance are advised to find a partner that can help them in several key areas:
Power distribution to the rack
AI demands high-density equipment racks supported by modular three-phase UPS systems, floor- and rack-mount power distribution
Cooling
AI applications need the latest in air and liquid cooling. Direct-to-chip (DtC) cooling, rear door heat exchangers (RDHx), and next-generation liquid cooling systems are appearing on the market that can cope with the extreme heat generated by GPUs and AI servers.
Advanced monitoring
Whether it is inside the central data center, or extending to branches, ATMs, offices, and across mobile services, advanced monitoring systems are vital to the implementation of AI in finance. Those overseeing IT and cybersecurity operations must be able to see at a glance everything that is happening across their networks and systems down to the desktop, laptop, server and application level. They need remote access tools to troubleshoot issues, access controls to prevent breaches, and effective ways of monitoring the entire network as well as all power and cooling infrastructure.
Services
Not all banking and insurance firms possess large, skilled teams of IT specialists. But even those with excellent teams often delegate certain maintenance and service functions to external partners. Others may lack the resources to provide the level of responsiveness needed to fix issues in their offices, branches, or ATM networks. The number of locations to service can sometimes reach into the thousands. Yet each one requires preventive maintenance and remedial work accompanied by rapid onsite response.
The ideal partner for all of this would cover each of these areas. Instead of dealing with one or two aspects, the best approach is to find a partner that provides solutions across the entire data center infrastructure spectrum.