CoreWeave, Inc., widely recognized as The Essential Cloud for AI, has introduced CoreWeave Flexible Capacity Plans, a new offering designed to better align cloud infrastructure with the unpredictable nature of modern AI workloads. The announcement includes two new options Flex Reservations and Spot which aim to give organizations greater flexibility and cost control when running artificial intelligence applications at scale.

As AI adoption continues to grow across industries, companies increasingly face challenges in balancing infrastructure capacity with fluctuating workloads. Traditionally, teams had to choose between reserved capacity, which guarantees resources but can lead to over-provisioning, or on-demand capacity, which provides flexibility but may lack guaranteed availability. However, CoreWeave’s new Flexible Capacity Plans allow customers to move beyond this rigid choice and adopt a more dynamic approach.

AI Authority Trend: CoreWeave and Perplexity Partner to Scale AI Inference Workloads

In many cases, AI training workloads follow predictable cycles, but production-level inference traffic can be far less predictable. Sudden spikes in demand often force engineering teams to either allocate excess infrastructure in advance or risk slower response times during peak usage. To address this challenge, CoreWeave has developed new flexible options that allow businesses to handle fluctuating demand more efficiently while keeping costs under control.

“At production scale, infrastructure planning becomes as critical as deployment,” said Chen Goldberg, EVP of Product & Engineering at CoreWeave. “CoreWeave is setting the standard for the AI cloud by providing guaranteed capacity when it counts and flexible pricing when demand shifts. We’re bringing the original promise of the cloud – scale and efficiency – back to the AI pioneers pushing the limits of innovation.”

A Unified Capacity Framework for AI

Importantly, the new Flexible Capacity Plans expand on CoreWeave’s existing infrastructure options Reservations and On-Demand to better support the real-world behavior of modern AI workloads. By introducing Flex Reservations and Spot, CoreWeave now offers a unified capacity framework designed specifically for large-scale AI operations.

The framework includes four capacity models:

  • Reservations: Reliable, always-on capacity designed for stable and predictable workloads.
  • Flex Reservations: A first-of-its-kind approach that guarantees peak capacity while offering flexible economics for workloads that scale unevenly. Customers secure a capacity ceiling with a reduced 24/7 holding fee and pay full rates only when instances are actively used.
  • Spot: A lower-cost compute option ideal for interruption-tolerant tasks such as batch analytics or data backfills. Spot instances include explicit preemption signals, enabling engineers to checkpoint and recover workloads seamlessly.
  • On-Demand: Best-effort access for organizations that require immediate and incremental capacity.

AI Authority Trend: CoreWeave Launches ARENA to Help Businesses Test AI Workloads at Production Scale

Together, these options allow enterprises to reserve stable workloads, protect mission-critical capacity, and shift flexible or interruptible jobs to more economical resources. As a result, companies can better align their infrastructure spending with actual usage patterns.

Industry partners are already seeing benefits from this flexible model.

“At inference.net, our mission is to help teams get the most from AI models,” said Ibrahim Ahmed, CTO at inference.net. “ The most powerful AI models are not generic, off-the-shelf models, they’re specialized models trained on your data, for your problem. We built custom scheduling and orchestration software so that any team can train and deploy frontier-quality specialized LLMs on underutilized GPU capacity. Spot instances from CoreWeave are key to making this possible, giving our customers access to the compute they need to bring custom models to production at a fraction of the cost.”

Currently, Flex Reservations are available in preview through CoreWeave account teams in select regions and SKUs, while Spot instances are generally available worldwide.

CoreWeave’s AI cloud platform continues to stand out in the industry due to its end-to-end technology stack optimized for AI workloads. Furthermore, the company consistently delivers high performance benchmarks. Its infrastructure recently achieved industry-leading MLPerf benchmark results and remains the only AI cloud provider to earn the top Platinum ranking in both SemiAnalysis ClusterMAX 1.0 and 2.0, which evaluate performance, efficiency, and reliability across AI cloud platforms.

With the introduction of Flexible Capacity Plans, CoreWeave is reinforcing its commitment to helping organizations scale AI applications more efficiently while maintaining the flexibility required for next-generation AI innovation.

AI Authority Trend: NVIDIA Expands Partnership with CoreWeave to Power 5GW of AI Infrastructure by 2030

To share your insights, please write to us at info@intentamplify.com