Years ago, Capital One made a momentous announcement: it had migrated all its on-premises data centers to the cloud via Amazon Web Services (AWS). Since then, Capital One has been deploying and scaling applications in the cloud while introducing a range of technological solutions—many powered by artificial intelligence—across the business.
This practice is known as FinOps, or cloud financial management, also referred to as cloud cost management, cloud optimization, or cloud financial optimization. The Growing Importance of FinOps at Financial Institutions, a report from Matthew Gaughan, Payments Analyst at Javelin Strategy & Research, examines how this corner of the banking industry has become a critical component for many financial institutions across all areas of their operations.
The Growing Cloud
The cloud is the defining component of FinOps, but the role itself is becoming more complex. A well-structured FinOps team has to understand both the FI’s business model and its underlying technology stack. Positioned at the intersection of multiple functions, the team has to ensure that the FI’s goals, costs, and decision-making processes are aligned. As emerging technologies, such as generative AI, become more commonplace, the challenges affecting various parts of the FI will only grow.
“More banks are starting to utilize the cloud beyond just as a means for storage,” said Gaughan. “Applications and services are going on in the background—things that most people might not even think of as banking functions.”
FinOps creates a common language among diverse teams working on the products and services that ultimately run in the cloud. These teams may include engineering, finance, and even business-focused groups. For instance, the team overseeing the loan process could benefit from more comprehensive credit scoring capabilities. It’s at this critical intersection of business, finance, and technology that FinOps delivers its true value.
Managing AI Resources
One key area that has gained importance is the management of computing resources. A 2024 presentation from Capital One employees at AWS’ re:Invent conference highlighted how they optimized the inputs and teams responsible for handling software projects. Their approach spanned the entire full-stack architecture, from the underlying infrastructure (computing, storage, networking), to the platform layer (operating systems, middleware, and runtime), and finally the application layer.
That breadth presents a challenge for many enterprises. It’s not easy to find people—or even teams—that can coordinate all of these functions.
“Making the actual on-the-ground decisions is an internal function,” Gaughan noted. “But it requires both an understanding of the third-party vendors that they’re utilizing to power these services, as well as the ability to work with those teams to make sure they’re applying them in the correct way.”
Across all of these technological frameworks, the Capital One team was able to reduce costs and even the energy load by optimizing chip usage, programming languages, and libraries. The streamlined environment allowed Capital One to attract top tech talent and ultimately become one of the most innovative FIs in the industry.
The forward-thinking approach has also positioned Capital One well to develop and deploy emerging AI solutions. “The promises of AI have made this a more compelling proposition for many banks,” said Gaughan. “We’ve seen banks adding more services that are underpinned by AI as well as vast amounts of consumer data, like anti money laundering checks or know your customer checks.”
The Role of the Hardware
From a hardware standpoint, central processing units (CPUs) and graphics processing units (GPUs) are critical components of cloud architecture. A bank’s need for each depends on its specific technology requirements. Traditional computing, which relies on CPUs and standard computer chips, is more common, making it easier for businesses to scale their computing power up or down. On the other hand, GPUs support more computing-intensive solutions but are scarcer and more expensive.
FinOps helps banks reduce these costs and energy consumption by taking a more proactive approach to resource management.
“The need to have insight into managing those CPUs and GPUs will become even more important, as data becomes a bigger component of the financial products and services that banks roll out,” said Gaughan.
These hardware needs are no longer limited to a bank’s back-office functions. Many FinOps-driven solutions directly impact the customer experience, playing a crucial role in driving business growth.
As banks introduce new products and features powered by AI and vast amounts of consumer data, managing these solutions becomes even more complex. Because of this complexity, FinOps is more of a concern for larger financial institutions.
“The largest banks are definitely using it more,” Gaughan said. “As you go further down the curve, the products and services offered by smaller banks become less varied. They may utilize the cloud in some form, but not to the extent or level of complexity of a larger bank that has multiple financial products that they’re rolling out across both consumer-facing and business-facing areas.”
“They have additional layers of complexity, such as servicing international customers with more complicated needs, that may or may not exist at a regional or smaller bank,” he said.
A Key Tool for Financial Institutions
It’s important to remember that FinOps wasn’t conceived by bankers—it has existed for as long as cloud computing itself. Enterprises of all type have turned to FinOps to oversee their underlying infrastructure.
However, financial institutions, with their wealth of data that can be leveraged for a wide range of purposes, may be uniquely positioned to benefit from FinOps. Ultimately, FinOps is as much a cultural initiative as it is a technological one. Fostering an internal dynamic that encourages cross-functional communication and provides teams with a shared understanding of the project at hand may be the key to achieving cost-effective, growth-focused outcomes.
