Financial institutions are shifting from building AI from scratch to fine-tuning open source models on their proprietary data, creating a new competitive advantage that traditional vendors can't replicate. According to the sixth annual NVIDIA State of AI in Financial Services report, based on a survey of more than 800 industry professionals, 84% of financial services organizations said open source models and software are important to their AI strategy, with 43% calling them very to extremely important. This represents a fundamental change in how banks approach artificial intelligence. Rather than waiting for vendors to build specialized solutions, institutions are now taking open source foundation models and customizing them with their own transaction data, customer interaction histories, and market intelligence. The result is AI capabilities that competitors cannot easily replicate, even if they use the same underlying open source tools. What's Driving Banks Toward Open Source AI? The shift reflects a practical reality: open source models offer flexibility and cost efficiency that proprietary solutions often cannot match. Banks can tailor development tools to their unique needs and make models more accurate by incorporating proprietary data that competitors don't have access to. This democratization of AI development has leveled the playing field between large institutions and smaller, more agile competitors. Helen Yu, CEO of Tigon Advisory Corp., explained the strategic advantage: "Open source models are fundamentally changing the competitive dynamics in financial AI. The real value capture happens when institutions fine-tune these models on their proprietary transaction data, customer interaction histories and market intelligence, creating AI capabilities that competitors cannot replicate". The financial impact has been substantial. Among survey respondents, 89% said AI is helping increase annual revenue and decrease annual costs. More specifically, 64% reported that AI helped increase annual revenue by more than 5%, with 29% seeing revenue increases exceeding 10%. On the cost side, 61% said AI decreased annual costs by more than 5%, and 25% reported cost reductions greater than 10%. How Are Banks Actually Using Open Source AI? The practical applications span multiple critical functions. Financial institutions are deploying open source-based AI for fraud detection, risk management, customer service, algorithmic trading, and document processing. A newer category, agentic AI systems, are now streamlining back-office operations and investment research. Agentic AI refers to advanced AI systems designed to autonomously reason, plan, and execute complex tasks based on high-level goals. The adoption rate has accelerated dramatically. In the latest survey, 65% of respondents said their company is actively using AI, up from 45% in the previous year. For generative AI specifically, 61% are using or assessing it, up 52% year over year. And 42% are using or assessing agentic AI, with 21% saying they've already deployed AI agents in production. Steps to Implement Open Source AI in Financial Services - Assess Your Data Assets: Identify proprietary transaction data, customer interaction histories, and market intelligence that can be used to fine-tune open source models and create competitive differentiation. - Choose the Right Model for the Right Problem: Evaluate whether open source or proprietary approaches work best for specific use cases, recognizing that leading banks need proficiency in both approaches. - Build Cross-Functional Teams: Establish teams with expertise in data science, risk management, compliance, and business operations to oversee model development and deployment. - Plan for Continuous Optimization: Allocate resources for ongoing model monitoring, retraining, and improvement as market conditions and customer behavior evolve. Alexandra Mousavizadeh, cofounder and co-CEO of Evident Insights, noted an important caveat: "Open source models can help banks close the gap with early movers, unlock cost efficiencies and safeguard against vendor lock-in, but they're not without their limitations. Proprietary approaches can unlock superior performance for domain-specific tasks. Leading banks need to demonstrate proficiency in both approaches, applying the right kind of model to the right problem, in the right context". What Does This Mean for Bank Budgets and Future Investment? The financial commitment to AI is intensifying. Nearly 100% of survey respondents said their AI budgets will increase or stay the same in the coming year. About 41% plan to invest in optimizing AI workflows and production, reinvesting in solutions that are already working. Another 34% are focused on identifying additional use cases for AI expansion, while 30% plan to build or provide more access to AI infrastructure, such as on-premises installations or cloud deployments. Investment in agentic AI is particularly notable. About 21% of respondents said AI agents have already been deployed, with another 22% planning deployment within the next year and beyond. These autonomous systems represent the next frontier in financial services automation, handling complex tasks that previously required human oversight. The strategic takeaway is clear: banks that treat proprietary data as a competitive asset and combine it with open source AI tools are positioning themselves to outpace competitors. As Yu emphasized, "The institutions winning in AI are treating their proprietary data as a strategic asset for building differentiated AI products". This approach allows banks to move faster than traditional vendors while maintaining control over their most valuable asset: their data.
