Compute Is Banking’s New Major Efficiency Consideration

In the past, the cost-to-income ratio for banks was primarily driven by one factor: people. The conventional wisdom was to focus on optimising human resources, streamlining operations, and automating repetitive tasks to improve efficiency.

But as the financial services industry evolves and banks increasingly embrace AI-driven operations, a new driver of costs has emerged: compute.

In this article, we’ll explore the implications of this shift, why compute efficiency is critical, and how banks can position themselves to win in this new era.

From people to compute: A paradigm shift 

Historically, banks have relied on large workforces to deliver services, process transactions, and manage customer relationships. With the rise of digital transformation, much of this labour-intensive work has been automated, reducing reliance on human resources and driving costs down.

The closure of bank branches has been a defining trend in recent years. In the U.K. alone, 62% of branches have shut their doors, and branch banking in Europe is predicted to all but disappear by 2025. While this move has reduced costs, it has also created a gap in the personal touch customers once received. 

Banks now face a dual challenge: replicating that personalised, human-centric experience digitally while ensuring operational cost efficiency. AI offers a solution, but only when deployed effectively. 

But AI technologies, while transformative, come with significant compute requirements. Large language models, generative AI, and real-time data processing demand substantial computing power. As banks deploy these technologies, compute has become one of the largest operational costs. 

The environmental cost of AI 

It’s not just cost efficiency that’s under scrutiny—there’s also the carbon footprint of AI to consider. For example, generating a single ChatGPT response can have six times the carbon footprint of a Google search.  

 With compute-intensive AI models driving up energy usage, banks must factor sustainability into their AI strategies. Balancing cost and carbon optimisation is crucial for staying competitive and meeting growing regulatory and consumer expectations around environmental responsibility. 

Why compute costs matter 

Compute costs affect banks in several ways: 

• Direct impact on competitiveness. Compute costs are now a major determinant of a bank’s cost-to-income ratio. Banks that can optimise their compute usage will achieve a clear cost advantage over their competitors. 
• Scaling challenges. As banks scale their AI initiatives, compute costs can grow exponentially. Without careful management, these costs can erode profitability and hinder innovation. 
• Slowing revenue growth. In an environment where revenue growth is becoming harder to achieve, controlling operational expenses—including compute—is critical for maintaining margins. 
• Positive correlation between compute cost and carbon. As compute costs rise, in most cases so will carbon footprint. Banks must take this into consideration from an environmental, social, and governance (ESG) and sustainability perspective and limit the risk of any potential fallout with environmentally conscientious customers. 

Winning the compute efficiency race 

To thrive, banks must prioritise compute efficiency while ensuring they can scale AI applications effectively. Here’s how: 

1. Adopt efficient AI platforms. Look for AI platforms specifically designed to optimise compute usage. These platforms use advanced algorithms to reduce unnecessary processing and maximise the value of each computation. Carrying an entire toolshed everywhere you go is inefficient when all you need is a single screwdriver. The same principle applies to AI: relying on a large, general-purpose language model (LLM) can consume far more compute than necessary, while a specialized, efficient AI can achieve the same outcome at a fraction of the cost. 
2. Leverage cloud-based infrastructure. Cloud platforms allow banks to scale their compute resources dynamically, paying only for what they use. This approach can help manage costs while maintaining flexibility. 
3. Focus on real-time data orchestration. Signal-oriented banking relies on real-time data processing. By investing in advanced data orchestration tools, banks can ensure data flows efficiently to where it’s needed, minimising redundant processing and compute waste. 
4. Implement governance and monitoring. Establishing strong governance frameworks for AI and compute usage is essential. Regular monitoring can help identify inefficiencies and ensure compute resources are being used effectively. 
5. Innovate with edge computing. Edge computing, which processes data closer to its source, can reduce the load on centralised systems and cut compute costs. Banks should explore this approach for applications like fraud detection and real-time customer interactions. 

Balancing efficiency with effectiveness 

Compute efficiency is about more than just cost control—it’s about maintaining the ability to innovate and deliver exceptional customer experiences. The banks that can optimise their compute usage without compromising performance will emerge as leaders in the industry. 

By embracing a signal-oriented approach, banks can achieve this balance. AI-driven signals enable real-time decision-making, improved customer interactions, and fraud prevention, all while minimising inefficiencies in compute usage. 

The future of banking economics 

The shift from people-driven costs to compute-driven costs represents a fundamental change in the economics of banking. Banks must rethink their strategies, prioritise compute efficiency, and adopt technologies that enable them to scale sustainably. And they must do it before those compute costs get out of control. 

As the industry evolves, the question for banks isn’t whether to embrace AI—it’s how to do so in a way that optimises costs, minimises environmental impact, and delivers the personal touch customers crave in the digital age.