CoreWeave

What is CoreWeave?

CoreWeave is a cloud provider built for AI workloads. It rents NVIDIA GPU instances, from HGX H100 and H200 nodes to GB200 NVL72 and HGX B200 systems, with CPU compute, file, block and object storage, networking, managed services and bare metal servers around them. Teams use it for training large models, running inference, and VFX and rendering. Capacity comes in four plans: Reservations for steady baseline demand, Flex Reservations for guaranteed peaks, On-Demand for overflow top-up when capacity is available, and Spot for lower-cost work that can be interrupted with seven minutes of notice. Prices are published per hour for each GPU instance in North America and Europe, and some newer systems are listed as contact sales. Software includes a Mission Control layer for monitoring and the newer CoreWeave Forge platform for connecting the AI development loop. For example, an 8-GPU HGX H100 node is $49.24 per hour on demand and $19.71 per hour on spot.

Use cases

• Rent GPU nodes for model training
• Serve inference on dedicated GPUs
• Run interruptible batch jobs on spot capacity
• Reserve guaranteed GPU capacity
• Render VFX and 3D workloads
• Fine-tune models on short-term GPUs

Who it's for

• AI labs
• Machine learning engineers
• Platform teams
• Enterprises

Why use CoreWeave

• You can match each workload to a plan, with reservations for baseline load, flex for peaks, and spot for movable jobs.
• Per-hour GPU prices are published by model and region, so you can estimate costs before talking to sales.
• Spot instances give a lower-cost way to run batch work, backfills and experiments.
• The same provider covers GPUs, CPUs, storage and networking, so a training stack does not need several vendors.
• A range of GPU generations lets you choose between older cards for smaller jobs and Blackwell systems for large runs.

CoreWeave features

• NVIDIA GPU instances from A100 and H100 to GB200 NVL72 and HGX B200
• Four capacity plans: Reservations, Flex Reservations, On-Demand and Spot
• Published per-hour prices by GPU model and region
• Spot capacity with 7 minutes of interruption notice
• Inference single-GPU pricing for inference platform customers
• CPU compute on Intel Xeon and AMD EPYC processors
• File, block and object storage billed per GB per month
• Bare metal servers and managed services
• Mission Control monitoring layer
• CoreWeave Forge for the AI iteration loop

How to get started with CoreWeave

1. Review the pricing page to pick a GPU model and region, North America or Europe.
2. Choose a capacity plan: Reservations, Flex Reservations, On-Demand or Spot.
3. Contact CoreWeave for reserved or contract capacity, or for models marked Contact sales.
4. Deploy your training or inference workload on the instances with the storage and networking you need.
5. Monitor runs and adjust the mix of reserved, flex, on-demand and spot capacity as demand changes.

CoreWeave limitations

• Some newer systems such as GB300 NVL72 show Contact sales instead of a price
• Reservations and Flex Reservations pricing is not listed on the public pages
• Spot capacity can be interrupted, so jobs must be able to recover
• On-Demand depends on available capacity
• Single-GPU inference pricing applies only to customers of the CoreWeave inference platform

CoreWeave pricing

CoreWeave has no free plan.

CoreWeave plans

• On-Demand

  $49.24/hour (8x HGX H100)

  Pay as you run, no commitment

  • Capacity subject to availability
  • Not interrupted once running
  • Prices per GPU instance

• Spot

  $19.71/hour (8x HGX H100)

  No commitment

  • Lower cost
  • Interruptible with 7 minutes notice
  • Capacity subject to availability

• Reservations

  Price on vendor site

  Fixed term

  • Fully guaranteed capacity
  • For steady baseline workloads

• Flex Reservations

  Price on vendor site

  Fixed term, usage-based with a holding rate

  • Fully guaranteed capacity
  • For variable usage curves

Plans as listed on the official pricing page on 24 Oct 2026. Prices can change and may vary by country or exclude tax. Check the official pricing page.

CoreWeave alternatives

The main CoreWeave alternatives listed on Debuly are Lambda, Crusoe Cloud, Together AI, Runpod, Vast.ai and Novita AI. The table compares pricing and platforms; order is A–Z.

1. 

   Crusoe Cloud

   AI cloud with on-demand GPU instances, managed Kubernetes and Slurm, and managed inference and fine-tuning for open models.

   Developer platformsCoding
   PaidVisit Crusoe Cloud (official site)
2. 

   Lambda

   GPU cloud for AI with on-demand NVIDIA instances and reserved 1-Click Clusters for training, fine-tuning and inference.

   Developer platformsCoding
   PaidVisit Lambda (official site)
3. 

   Novita AI

   AI cloud with pay-per-token model APIs for 200+ models, dedicated GPU endpoints, agent sandboxes and GPU rentals.

   Developer platformsCodingAutomation and agents
   PaidVisit Novita AI (official site)
4. 

   Runpod

   Pay-per-second GPU cloud for AI: on-demand Pods, Serverless endpoints, multi-node Clusters and ready-made public model APIs.

   Developer platformsCoding
   PaidVisit Runpod (official site)
5. 

   Together AI

   AI cloud for running open-source models through serverless and batch inference, dedicated endpoints, GPU clusters and fine-tuning.

   Developer platformsCoding
   PaidVisit Together AI (official site)
6. 

   Vast.ai

   GPU cloud marketplace for renting on-demand, interruptible or reserved GPUs with per-second billing.

   Developer platformsAutomation and agents
   PaidVisit Vast.ai (official site)

CoreWeave FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from cdn.prod.website-files.com · fetched 23 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.