Why open source matters in fintech AI

Jonathan Bailey is chief technology officer of virtual card issuing platform and spend management software firm Extend.

In the first half of 2025, AI accounted for over half of global VC deals and nearly two-thirds in the U.S. But while funding for AI accelerates, investment in fintech has seen a marked slowdown. For fintech companies, navigating this moment isn't just a strategic priority, it’s a matter of relevance and survival.

Yet something crucial is missing from fintech's AI story: our voice in shaping it.

Fintech leaders need to start contributing to AI’s core infrastructure — tools, frameworks, datasets — or risk being shut out of how this technology evolves.

Across GitHub, Hugging Face, and Arxiv, fintech companies are heavy adopters of open-source AI tools but conspicuously light contributors. We use AI to automate underwriting, streamline payments, detect fraud, and power virtual assistants but we rarely give back. That lack of participation may be costing us more than we realize.

The rules are being written without us

While fintech sits on the sidelines, other industries are actively shaping AI's future. Health care companies have published biomedical models that establish safety standards for life-critical applications. Enterprise SaaS teams are releasing agent frameworks that define how businesses interact with AI. Even defense contractors have contributed synthetic datasets that influence how models handle sensitive information.

These open-source efforts aren't altruistic gestures. They’re strategic moves to ensure AI evolves in ways that serve industry needs. Every framework, standard, and best practice being established today will determine what's both possible and permissible in AI applications tomorrow.

Meanwhile, financial services — an industry that moves $5 trillion daily and touches every economic transaction — is letting others write the playbook.

A false dichotomy

Much of the conversation around AI has been framed as a battle between open and closed, community-driven tools and protected IP. That framing has trapped many fintech leaders in unnecessary paralysis.

The reality is that the most forward-thinking companies embrace both approaches strategically. They contribute foundational tools to accelerate ecosystem-wide innovation while investing in or utilizing proprietary layers that deliver differentiated value. It's not just possible to contribute to the ecosystem and protect what makes your product unique, it’s becoming the most compelling strategy.

Consider what we're ceding by staying quiet:

• Standards and safety protocols are being established by industries with varying risk tolerances and regulatory requirements. The frameworks emerging from health care and enterprise software may not account for financial services' unique compliance needs or the real-time nature of payment systems.

• Talent and partnership opportunities flow toward companies actively participating in the developer ecosystem. The best AI engineers want to work on problems that push the field forward, not just implement existing solutions behind closed doors.

• Regulatory influence is increasingly occurring via technical standards rather than traditional lobbying. When regulators need to understand AI capabilities and limitations, they turn to the companies and research that are working visibly and publicly.

Why fintech lags — and why it matters

The reasons for fintech's hesitation are understandable. We operate in heavily regulated environments, handle sensitive data and maintain systems that must be reliable and auditable. Releasing code publicly feels risky when one bug could trigger regulatory scrutiny or compromise customer trust.

But this caution has become counterproductive. Many fintech companies are solving identical problems — cleaning messy transaction data, building compliance-aware agents, integrating with legacy core systems — yet each starts from scratch. We're duplicating effort while missing opportunities to establish industry-wide standards that actually serve our needs.

According to recent industry data, 84% of financial firms report that open-source AI delivers "substantial business value," and nearly half plan to expand their use of open tools this year. We're already betting our businesses on community-driven innovation but we’re not helping steer it.

A matter of strategy, not charity

This kind of effort is more driven by competitive positioning than corporate social responsibility. We already see in other realms that the companies that help establish AI's foundational layer have disproportionate influence over its evolution.

Smart contribution should focus on problems that are common across the industry, but not critical for competitive differentiation. Tools for data sanitization, compliance checking, or legacy system integration would benefit everyone without revealing proprietary algorithms or customer insights.

At Extend, we built and open-sourced a toolkit for building secure, compliant AI agents because our developer partners needed faster, more accessible paths to implementation. It helps streamline workflows across our ecosystem and supports Extend’s work in developing agentic AI systems for financial services.

Other fintechs are also open-sourcing tools to accelerate adoption while protecting what differentiates them. Stripe has long invested in open developer tooling while maintaining clear competitive moats through their core platform capabilities. Ramp recently open-sourced their MCP Server, allowing customers to interact with business data using natural language, showcasing their AI capabilities without revealing any core proprietary IP.

These aren't philanthropic gestures, they’re strategic moves by companies that understand how influence works in developer-driven markets.

The proprietary layer still matters

Proprietary technology still plays a critical role. The AI agents Extend is building are deeply integrated with our platform, informed by customer-specific data, and tailored to workflows our users depend on. These capabilities are hard-won competitive advantages that we need to protect.

But that protection doesn't require total secrecy. The most valuable contributions often come from companies with the deepest domain expertise. Who’s better positioned to shape financial reasoning frameworks than the firms processing millions of transactions? And who knows more about AI compliance than the teams building the systems?

Build, contribute, lead

The future of fintech AI won't be determined by who adopts the flashiest tools the fastest. It will be shaped by builders — companies that participate openly, contribute to shared infrastructure, and move quickly without compromising trust.

Open-source AI is becoming the foundation layer where design patterns are established, safety practices are tested, and industry standards emerge. Fintech companies that participate in creating this layer will have a voice in how our rules are written.

The window for influence is narrowing. While we've been pondering whether to participate, other industries have been defining what AI looks like in regulated, high-stakes environments. Every month we wait is another month of standards being set without our input.

You don't need to open-source your secret sauce, but if you want to stay relevant in an AI-driven financial services landscape, contributing something isn't optional anymore. It’s perfectly fine to start small: a tool, a framework, some test data. Every contribution counts. The companies shaping the future of fintech AI won't do it alone, and they won't do it behind closed doors. The question isn't whether to participate in the open ecosystem. It's whether you'll help lead it or let others define what's possible in your industry.