Revolut moves to self-hosted GLM

This move gives Revolut greater control over its customer data and AI model performance.

Fintech company Revolut is transitioning to a self-hosted Generative Language Model (GLM), moving away from reliance on third-party providers for its AI services. This strategic shift aims to enhance control over its data, improve model performance, and potentially reduce operational costs. The company has been investing in its in-house AI capabilities, and this move represents a significant step in that direction. By managing its own GLM infrastructure, Revolut seeks to gain greater flexibility in model customization and deployment, tailoring AI functionalities to its specific financial products and customer needs.

Revolut's decision to self-host its GLM signals a broader trend among large enterprises to bring AI infrastructure in-house, driven by concerns over data privacy, security, and vendor lock-in. This approach allows for deeper customization and optimization of AI models for specific business needs, potentially leading to more efficient and tailored customer experiences in the competitive fintech landscape. As more companies explore this path, it could also spur innovation in open-source AI development and specialized hardware solutions.

The move away from third-party GLM providers by a major player like Revolut could impact the market for AI-as-a-service, potentially increasing pressure on existing providers to offer more competitive pricing and enhanced security features. It also highlights the growing maturity of in-house AI development teams within large corporations, enabling them to manage complex infrastructure and sophisticated model deployments. Companies that continue to rely on external vendors may need to reassess their strategies in light of this developing trend.