The Keynotes: AI, Sovereignty, and the Power of Open Collaboration
The main stage at OSFF Toronto delivered a clear consensus: open source is no longer just a cost-saving measure; it is the critical infrastructure required to safely scale artificial intelligence and enterprise architecture.
Open Source as a Core Differentiator
Hitesh Kamdar, Head of Capital Markets Technology Architecture at RBC Capital Markets, set the tone by declaring that open source "is not anymore just you know something on the side it is a differentiator it is not simply about cost savings". He emphasized that "it is no longer [an experiment] we're all using these projects, many of us in production. So it starts to become part of the ecosystem."
The Sovereign AI Imperative
As institutions scramble to deploy generative AI, the risks of relying on opaque, proprietary models have become glaringly obvious. Vincent Caldeira, Chief Technologist for FSI at Red Hat, tackled this head-on, warning that without independent control over your infrastructure, data, and models, institutions are effectively just "renting opaque intelligence." His takeaway was optimistic, however: "The great news is that open source is the solution to this control problem that you have starting with the platform itself."
Building Trust on Open Foundations
Andres Rojas from the Vector Institute reinforced the need for open AI ecosystems, explaining that the core algorithms behind AI cannot truly be walled off as intellectual property. "Open models don't require this blind trust right with open models you can verify it," Rojas explained. "Build on open foundations, right? The algorithms are already shared. Invest your edge in data, invest in domain expertise, invest in good governance. Those are the things that are particularly critical for growing a good AI foundation rather than the algorithm itself."
Innovation Through Community
Mia Gougisha, Director of Technology Research & Innovation at DTCC, showcased how the industry is already putting these collaborative models to the test. Reflecting on a recent cross-firm AI hackathon, she noted: "This hackathon was fundamentally about bringing together people beyond organizational boundaries and geographic lines to solve shared industry challenges. This is what modern innovation should look like."
Toronto's AI Legacy
Mark Paulsen, Head of the Open Source Program Office at TD Bank, brought it all back to the host city's unique position in tech history. "Toronto... this is where AI deep learning, you know, kind of started," Paulsen reminded the crowd. "Open source, as we're all aware, is what shared this with the world. And right now in this room... we are the ones who get to decide what we do with this stuff."
Next-Gen Compute: Abstracting the Infrastructure
One of the most highly anticipated sessions was the keynote panel on High-Performance Computing (HPC), featuring experts from Citi, RBC Capital Markets, and Amazon Web Services.
The panel tackled the immense complexity of managing multi-cloud APIs, diverse purchasing options (like Spot versus On-Demand), and fluctuating real-time capacity. The consensus solution centered around abstracting the underlying compute pools so developers and quants can focus purely on executing their workloads rather than navigating infrastructure.
By leveraging open-source projects like the Open Resource Broker (ORB) and HTC-Grid, financial institutions can create a unified API layer that seamlessly handles placement, retries, and capacity provisioning at scale. The discussion highlighted how projects like Citi's OpenGRIS act as that higher-level language for developers. Furthermore, RBC formally announced the contribution of FiveSpot, a project designed to harvest and reclaim idle on-premise compute capacity (the "night shift"), integrating seamlessly with ORB to prioritize internal resources before bursting to the cloud.
This was the perfect build up to FINOS announcing its HPC initiative today, as these projects came together to build a harmonize ecosystem for grid computing in finance services. Check out hpc.finos.org to get involved.
Breakout Themes: Architecture, Automation, and Data
Beyond the main stage, the breakout sessions dove deep into the technical weeds of the FINOS ecosystem.
- CALM (Common Architecture Language Model): Morgan Stanley showcased CALM, a machine-readable JSON meta-schema that treats "architecture as code". By codifying systems, organizations can embed security controls directly into their designs and even use LLMs to automatically analyze the business impact of infrastructure outages.
- Fluxnova 2.0: This enterprise-grade execution engine for BPMN and DMN models officially launched its 2.0 version. Sessions demonstrated how it handles complex, multi-step workflows at massive scale (capable of over 100 billion jobs a year on standard hardware) while integrating seamlessly with AI models.
- Secure FDC3 Interoperability: The FDC3 standard continues to evolve, with new focus on cross-firm security. Demos from BlackRock and RBC highlighted how FDC3 now incorporates application identity verification and data integrity checks, ensuring that sensitive information like trade prices can be transmitted securely across the enterprise desktop.
- The Open Data Commons: Data silos remain a massive cost center for financial institutions. Projects like Overture Maps Foundation and OpenData.org highlighted the push toward shared entity graphs—proving that open collaboration on foundational reference data can dramatically reduce reconciliation errors and improve AI training sets.
Moving Forward: Get Involved
The conversations started in Toronto are just the beginning. We encourage you to catch up on the sessions you missed and get actively involved in the FINOS community.
Catch Up on OSFF Toronto:
Join Us in London:
We are keeping the momentum going across the pond. Join us at OSFF London on June 25, 2026, at the Queen Elizabeth II Centre.
Get Involved with FINOS Projects:
Open source relies on active participation. Whether you are contributing code, writing documentation, or providing use-case feedback, your voice matters. Explore the links below to get started:
The foundation’s mission now is to extend that value across every layer of financial technology – turning open collaboration into shared infrastructure, shared assurance, and shared intelligence.
FINOS – Building the future of open finance together.
Learn more about FINOS - subscribe to our updates.