Building the Enterprise AI Factory: Scaling AI Across Hybrid Cloud

As generative AI moves into enterprise-scale production, more GPUs alone won't solve it. The real challenge is building a platform that matches the right intelligence to the right infrastructure at the right cost, without giving up security or governance.

In this session, JP Morgan shares how they built an Enterprise AI Factory spanning on-premises infrastructure, multiple clouds, and specialized accelerators. The platform places workloads across GPUs and purpose-built inference accelerators, optimizing for performance, latency, economics, compliance, and resilience.

The discussion will cover how AI token economics is pushing enterprises toward open models and workload-aware infrastructure, matching models to use cases instead of defaulting to frontier models. It also looks at how partners like CoreWeave help extend AI capacity while preserving the security, governance, and operational controls required in highly regulated industries.