GPUs Are Supercharging Algorithmic Finance

Six Ways GPUs Are Being Used in Financial Services

GPUs in the financial sector are rapidly becoming common as providers integrate them into their AI solutions. Here are some popular use cases.

1. GPUs Are Fueling AI-Powered Trades 

Capital markets firms are taking advantage of the efficiency gains derived from GPU-accelerated AI, optimizing high-frequency trading, risk modeling and portfolio management. GPU-powered AI is helping them gather key insights from large and varied data sets with unprecedented speed, bolstering their market analysis and their ultimate decision-making.

JPMorganChase, for instance, has used NVIDIA T4 GPUs to the tune of a 40x increase in the end-to-end speed of its risk calculations, while reducing cost of ownership by 75%. Risk calculations now run in minutes instead of hours. And by integrating GPUs into the global computing infrastructure, the company experienced roughly 70% GPU utilization rates every hour of the day.

2. GPUs Are Helping Large-Scale Fraud Detection

Sophisticated fraud schemes are difficult for financial institutions to defend against. Nearly 80% of credit union and community bank leaders, for instance, said their institutions reported fraud losses of over $500,000 in 2023. In response, banks and payment processors are using GPUs to power AI-driven fraud detection systems, such as NVIDIA’s RAPIDS and RAPIDS Accelerator for Apache Spark.

NVIDIA’s RAPIDS seamlessly integrates with frameworks to support deep learning algorithms such as graph neural networks, which can reduce false positives in transaction fraud detection, enhance identity verification accuracy and make anti-money laundering efforts more effective. The RAPIDS Accelerator enables the processing of enormous volumes of sensitive data at unprecedented speed, all while maintaining commitment to regulatory compliance and data security. Together, they prove instrumental in establishing comprehensive fraud detection systems.

3. GPUs Are Improving RegTech Compliance

Recent regulation technology (RegTech) research published in the Journal of Financial Economics concludes that “technological advances will strengthen the linkages between compliance and operating functions, especially as financial institutions increasingly rely upon RegTech solutions for compliance and more customer information is digitized.” GPUs are helping drive such advances.

Banks are deploying GPU-powered AI to streamline compliance processes, monitor transactions for regulatory violations and automate reporting. This, in turn, makes governance more achievable across the board.

A joint report by Dell Technologies and NVIDIA notes: “Meaningful gains can be achieved through AI as even minor improvements in detection accuracy significantly reduce costs and improve regulatory compliance.” 

 Click the banner below for financial expertise on navigating regulatory compliance.

4. GPUs Help With Statistical Modeling

 GPUs are helping financial firms develop statistical models that incorporate millions of variables and their relationships to predict the changing contours of markets. The processing speeds of GPUs also enable financial institutions to reorient the models and algorithms as needed — in real time. 

5. GPUs Are Helping Banks Refine Investment Portfolios

Often in tandem with statistical modelling, GPUs are helping banks maximize returns and minimize risks in their current portfolios. Munich RE Markets, for instance, has developed an AI-based analytics and quality assurance tool for diversified portfolio reconstruction. When working with NVIDIA to rebuild its models, the company experienced a 50x performance boost.

GPU-powered generative AI also allows access to rich statistical data in easily comprehensible visual forms. That makes it simpler for financial institutions to understand the direction of their investments and pivot when necessary.

“That’s really what AI is all about: enabling every business to create and deliver intelligence to its business to better inform decisions,” Paikeday said. 

6. GPUs Are Behind RPA

Enhanced by GPUs to accelerate its processes, robotic process automation is an emerging form of intelligent automation technology in which robots are used to perform repetitive tasks, such as extracting data, filling in forms and moving files. This helps improve the efficiency of finance processes, simultaneously freeing up human personnel to tackle more complex jobs. It’s a key reason that roughly 80% of finance leaders have implemented or are planning to implement RPA, according to Gartner.

With all such GPU use cases, everything boils down to addressing one core need: improving business intelligence. Paikeday put it this way: “In this next industrial revolution, we’re really talking about enabling every enterprise to have their own ability to deliver massive intelligence at scale.”