BNP Paribas has entered a five-year partnership with Google Cloud to gain access to its public cloud infrastructure and Gemini Enterprise models as the bank develops AI agents for internal use. According to the latest announcement, the agreement grants the French bank access to Google's public cloud infrastructure and its Gemini family of AI models. The partnership was announced on September 24, 2026, with a dual dateline from Paris and Sunnyvale, California. The agreement will begin with BNP's Corporate & Institutional Banking division, where purpose-built agents will be developed using Gemini Enterprise. As reported by PRNewswire, Thomas Kurian, CEO at Google Cloud, framed the collaboration as a milestone for financial services, stating that "Agentic AI marks a critical shift in how global financial institutions streamline operations, assess risk, and serve their clients."
The partnership will support targeted workflows across multiple business units, with Gemini models being integrated into LLM@CIB, the bank's generative AI assistant currently available to more than 65,000 employees. According to the latest announcement, the new capabilities will support AI agents for corporate credit memo preparation to assist teams, with further applications planned across sales, trading, research, and structuring activities. BNP Paribas will have access to Gemini Enterprise to build, evaluate and deploy purpose-built agents tailored to specific business needs. Google Cloud and Gemini technologies have already been deployed at Nickel, BNP's payment account business, where Gemini serves as the foundation for Nickel Assist, helping roughly 200 customer advisers locate answers based on the unit's procedures and operating guidelines. As reported by PRNewswire, Charles Holive, Chief AI Officer at BNP Paribas CIB, emphasized that "Agentic AI can help our teams handle more complex workflows while maintaining the appropriate level of control."
BNP Paribas will retain full ownership of its data while customer data will not be used to train, tune or improve Google's foundational AI models. The bank's existing security and data governance framework dictates which categories of data and workloads are permitted to be processed in a public cloud setting. As reported by Reuters, BNP Group Chief Information Officer Marc Camus explained that "The fact that we signed this agreement with Google doesn't mean that we will get rid of our on-premise infrastructure." The bank will continue using open-source models and maintain its collaboration with French AI startup Mistral, using different providers for different applications depending on security, cost, and performance requirements. The announcement describes the deployment as operating within a controlled, enterprise-grade governance framework. Each agent will be authenticated and granted access only to the resources required for their assigned tasks, with the bank monitoring and controlling both the connection of AI agents to the group's information systems and the agents' interactions to ensure operational security.
The bank sees potential to apply Google's technology to less sensitive tasks, including invoice processing and data management. According to Reuters, BNP plans to keep using open-source models and maintain its partnership with French AI company Mistral, with the providers used for different applications. The bank's AI agents will be authenticated, restricted to the resources required for their tasks and subject to ongoing monitoring to ensure operational security. As reported by PRNewswire, Marc Camus emphasized that "Our role is to empower our businesses with the technology capabilities they need while maintaining a consistent framework for security and control. A multi-cloud and multi-model approach gives us the flexibility to select, for each case, the right model, infrastructure and level of control to meet our requirements for quality and security while delivering the best balance between cost and value."
The approach highlights the challenge for European financial institutions of adopting generative and agentic AI while meeting privity, operational resilience and data sovereignty requirements. European regulators have been particularly attentive to how financial institutions manage data sovereignty and third-party technology risk. For BNP, which operates in 64 countries with more than 180,000 employees, the arrangement offers a way to experiment with advanced AI capabilities while preserving the control mechanisms that regulators and internal risk teams demand. The decision to keep certain data on-premise also reflects the reality that large financial institutions rarely adopt a single cloud vendor for all workloads. According to Unite.AI, the partnership reinforces BNP Paribas' multi-cloud and multi-model strategy, allowing the bank to select the right model, infrastructure and level of control for each case while meeting quality and security requirements.
