Flexify.IO brings data closer to GPUs with CoreWeave support

GPU capacity is only part of preparing an AI workload. If your training data lives in another cloud, you also need to get it to where your compute runs. Moving datasets into CAIOS ahead of a training run helps avoid spending allocated GPU time waiting for the initial transfer.

Flexify.IO now supports CoreWeave AI Object Storage (CAIOS), helping AI teams bring their data closer to the GPUs running their workloads.

Migrate existing datasets to CAIOS

For teams bringing an AI workload to CoreWeave, the first step is getting the existing dataset ready. Flexify.IO cloud-to-cloud migration transfers data from Amazon S3, Azure Blob Storage, Google Cloud Storage and other supported object storage providers into CAIOS at over 1 PB of data per day while avoiding source provider's data egress fees.

Keep fresh data flowing to your GPUs

The first migration gets existing data into CAIOS. Ongoing synchronisation keeps it up to date as new data arrive, labelled and curated datasets change for the next round of training, fine-tuning or evaluation.

Flexify.IO sync pipelines transfer new and updated objects to CAIOS without copying the entire dataset again. Synchronisation can be scheduled at regular intervals or use S3 notifications to bring changes across sooner. This makes bringing data close to your GPUs an ongoing process as datasets grow.

From the first transfer to the next training run

For example, an AI team might hold its training archive in Amazon S3 while running GPU workloads on CoreWeave. The team can migrate that archive to CAIOS, then configure a sync pipeline to bring across newly prepared training files as they become available.

Before each experiment, the team can prepare a versioned dataset and complete its transfer to CAIOS. Bulk migration handles the starting dataset; sync carries subsequent additions and updates to the cloud where the GPUs run.

Get started

Simply add your source storage and CoreWeave AI Object Storage to Flexify.IO account, then start a migration. Once the initial transfer is complete, set up synchronisation to keep the data your GPU workloads need up to date.

To plan your migration and sync to CAIOS, contact the Flexify.IO team.