Cloudflare R2 and MosaicML enable training LLMs on any compute, anywhere in the world, with zero switching costs

Training large language models (LLMs) and diffusion models requires massive infrastructure, including significant storage capacity for terabytes to petabytes of training datasets and model checkpoints. To manage storage costs and scalability, many machine learning teams have been moving to object storage providers like Cloudflare R2. However, these providers often charge high egress fees, making it difficult to leverage GPU capacity across multiple cloud providers or take advantage of lower pricing elsewhere. MosaicML's tools and Cloudflare R2 address these challenges by enabling efficient use of R2 as the durable storage backend for training LLMs on any compute provider with zero egress fees. This allows users to run training workloads on any compute provider, with total freedom and zero switching costs. The combination of MosaicML's platform and Cloudflare R2 provides maximum autonomy and control, allowing organizations to switch between cloud service providers as needed.