Industry-leading environment for validating performance and cost at scale before full deployment
CoreWeave, Inc. (Nasdaq: CRWV), The Essential Cloud for AI™, today announced CoreWeave ARENA (AI-Ready Native Applications), designed to help teams test workloads on purpose-built AI infrastructure and software that mirrors how AI actually runs in high-demand production settings. Replacing traditional sandbox or demo setups, the industry-leading CoreWeave ARENA pairs production-scale compute with a standardized evaluation environment.
This press release features multimedia. View the full release here: https://www.businesswire.com/news/home/20260205676969/en/
Introducing CoreWeave ARENA, a new AI lab where customers can validate AI workload performance and cost readiness at scale.
As AI becomes part of day-to-day operations across industries, the integration from infrastructure to software determines ultimate success. Workloads are becoming more continuous and distributed, and performance or cost differences now matter more as everything from accelerators to agents evolve simultaneously. Organizations need earlier and more accurate insight into how their workloads will perform at scale, without the overhead of bespoke evaluation processes. CoreWeave ARENA provides a low-friction way to evaluate new workload performance through comprehensive, real-world benchmarking.
“Companies are now building products that have to perform in the real world, and quickly. But there’s been a gap between testing a model and actually running it at scale. CoreWeave ARENA closes that gap,” said Chen Goldberg, Senior Vice President, Engineering at CoreWeave. “We’re giving customers the tools and data they need to know exactly how their workloads will behave and what they will cost before they ever switch on production.”
For business and technology leaders across industries, CoreWeave ARENA offers a clearer way to understand whether an AI system is ready to run reliably at scale, before committing resources to teams, infrastructure, and operations.
“CoreWeave is providing an element of understanding and cost-certainty that was missing from the AI race, and is especially crucial for emerging companies,” said Dave McCarthy, research vice president, Cloud and Edge Services, Worldwide Infrastructure, IDC. “The inference stage, where models are actively processing live data and generating predictions, requires both compute power and intelligent system design for workloads to scale, and testing these ahead of production-altering decisions is key.”
Customers evaluating workloads through CoreWeave ARENA can trial the category-defining technologies that differentiate CoreWeave’s AI-native platform:
- CoreWeave Mission Control, CoreWeave’s AI operating standard, which provides baseline visibility into system behavior and optional use of the Mission Control Agent to pull key insights about metrics and workload behavior.
- AI-native infrastructure through SUNK (Slurm on Kubernetes) and CKS (CoreWeave Kubernetes Service), with optional integration into existing tools such as Weights & Biases for tracking and understanding model behavior.
- High-throughput data movement from object storage to AI infrastructure using LOTA (Local Object Transport Accelerator).
“Before committing to a full proof of concept, we wanted a clear view into how our workloads would actually perform, especially after seeing how much results can vary across providers,” said Xander Dunn, Member of Technical Staff, Periodic Labs. “Running our workloads on production-scale infrastructure gave us early, concrete insight into both performance and cost, which helped us evaluate next steps as we plan for scale, without slowing down execution.”
Customers are already starting to see the benefits:
- Speed to market: With earlier access to an NVIDIA GB300 NVL72 rack through CoreWeave ARENA, one customer quickly experienced >2× better performance compared to a previous generation.
- Cost efficiency: One customer lowered Total Cost of Ownership (TCO) by ~30% using CoreWeave ARENA.
- Architecture validation: One customer saw 10X improvement in training time compared to a competitive cloud running the same GPU.
Today’s announcement is part of CoreWeave’s broader platform strategy to unify the essential tools required to run AI at production scale on a single cloud platform, spanning high-performance compute, multi-cloud compatible data storage, and the software layer builders rely on to develop, test, and deploy AI systems. Recent platform innovations such as Serverless RL, the first publicly available, fully managed reinforcement learning capability, further extend this foundation. CoreWeave’s focus on performance and operational excellence is reflected in industry-leading MLPerf benchmark results and its distinction as the only AI cloud to earn top Platinum rankings in both SemiAnalysis ClusterMAX™ 1.0 and 2.0, reinforcing the company’s ability to deliver advanced AI infrastructure with reliability and efficiency at scale.
About CoreWeave
CoreWeave is The Essential Cloud for AI™. Built for pioneers by pioneers, CoreWeave delivers a platform of technology, tools, and teams that enables innovators to move at the pace of innovation, building and scaling AI with confidence. Trusted by leading AI labs, startups, and global enterprises, CoreWeave serves as a force multiplier by combining superior infrastructure performance with deep technical expertise to accelerate breakthroughs. Established in 2017, CoreWeave completed its public listing on Nasdaq (CRWV) in March 2025. Learn more at www.coreweave.com.
View source version on businesswire.com: https://www.businesswire.com/news/home/20260205676969/en/
