November 13, 2025 — For teams scaling AI, data mobility has become one of the industry’s most expensive constraints. Transferring training data or model outputs between clouds can cost up to $0.12 per GB in egress fees, quickly turning a single petabyte move into a six-figure expense. Most “free data transfer” offers waive egress fees if customers agree to close their entire accounts as part of a one-time switch, an ultimatum that slows innovation and traps data behind commercial fine print. These costs and conditions create a kind of gravity that keeps even the most advanced AI teams tethered to suboptimal infrastructure.
Data should never hold back innovation. That’s why CoreWeave has launched the Zero Migration ([0]EM) program, a secure, white-glove service that enables customers to move large-scale datasets from major cloud providers directly to CoreWeave. And it only gets better from there: CoreWeave will cover the egress fees when leaving their existing cloud, saving customers up to $1 million for the typical transfer. Once data is in CoreWeave AI Object Storage, customers will not pay CoreWeave egress fees, no matter where the data is used.
By removing the barriers to large-scale data movement, the [0]EM program gives teams total freedom with their data—ready to power training, fine-tuning, and agentic applications with the speed and performance of CoreWeave’s AI-native infrastructure and our AI Object Storage. Unlike other migration programs, CoreWeave takes a fundamentally different approach by allowing teams to transfer data at petabyte scale, retain full access to their existing environments, and quickly start building on CoreWeave AI Object. Where others impose trade-offs, CoreWeave removes them, because innovation should move as freely as data itself.
Talk with our team today to learn how 0EM can get you started building on the Essential Cloud for AI.
How the [0]EM program works
[0]EM removes the manual effort and uncertainty that typically come with large-scale data transfers. Customers simply provide read-only credentials, and CoreWeave orchestrates the rest—executing end-to-end checksum validation to guarantee integrity and accuracy across every object. Migrations move at petabyte scale, typically exceeding 1 PB per day, with zero data loss and full fidelity verified at every stage.
A real-time migration dashboard gives teams complete visibility into progress and performance, ensuring confidence that data remains secure, intact, and accessible throughout the process. Backed by CoreWeave’s expert team available via a dedicated Slack channel, the entire workflow is engineered for transparency and reliability, allowing platform leaders to migrate massive datasets quickly and start powering AI workloads with minimal interruption.
The performance edge of CoreWeave AI Object Storage
Once data moves into CoreWeave, it doesn’t just arrive—it thrives. CoreWeave customers already rely on CoreWeave AI Object Storage to power their most demanding workloads. As Xander Dunn, Member of Technical Staff at Periodic Labs, shares:
"CoreWeave AI Object Storage delivers the throughput and reliability our research pipelines depend on, balancing speed and efficiency across active and inactive data. It’s allowed us to experiment faster while keeping costs under control."
AI Object Storage delivers up to 7 GB/s per GPU of throughput with 11 nines durability and 99.9% uptime, providing a high-performance foundation for AI at scale. This unified dataset architecture allows compute and storage to scale together, maintaining consistent throughput and efficiency even as workloads grow to hundreds of thousands of GPUs.
Storage shouldn’t slow down innovation. With CoreWeave, it never will.
[0]EM marks a real shift for AI teams that build across clouds. For years, data mobility has come with trade-offs: steep egress fees, rigid lock-ins, and downtime that slows progress. CoreWeave ends those limits. Unlocking the power of AI Object Storage, the program gives teams the freedom to move and scale data anywhere, with zero egress fees and full multi-cloud flexibility. It’s part of our broader mission to build the Essential Cloud for AI, one that removes friction between data, compute, and innovation and embraces a multi-cloud world.
Take the next step toward your AI breakthrough:
- Talk with our team to learn more about joining the [0]EM program.
- Explore the latest updates to CoreWeave AI Object Storage.
- Dive into performance benchmarks to see CoreWeave AI Object Storage in action.
.png)
