Today, Vantage is launching support for CoreWeave costs, giving AI and infrastructure teams visibility into their GPU cloud spend alongside the rest of their cloud and AI providers. Once connected, Vantage automatically imports daily CoreWeave usage by SKU, cluster, and region and prices it at CoreWeave's on-demand rates or your own contract rates, so teams can report on, budget for, and allocate CoreWeave spend like any other provider.

CoreWeave is an AI cloud purpose-built for training and serving models at scale. It offers NVIDIA GPU compute such as HGX H100, H200, and B200 and GB200 NVL72, along with CPU compute, AI Object Storage, networking, and managed CoreWeave Kubernetes Service (CKS) clusters, plus software for running AI workloads such as Mission Control and Weights & Biases. AI labs and enterprises use it for large training runs and production inference, and for teams that train or serve their own models, that GPU capacity is often the largest part of the AI budget. Previously, CoreWeave usage lived in the CoreWeave Cloud Console, separate from the rest of a team's cloud and AI spend. Teams that wanted it next to AWS or their model providers had to export usage, price it against their contract by hand, and upload the result to Vantage as a Custom Provider.

Now, with CoreWeave cost support in Vantage, customers connect a CoreWeave organization with an API access token, and Vantage imports your usage from your organization. The CoreWeave usage export is currently in Public Preview and does not contain costs, so Vantage provides customers a per-SKU price sheet, which defaulted to CoreWeave's public list prices and can be updated to your rates if you have negotiated enterprise discounts. This allows Vantage to calculate price estimates based on your usage to view CoreWeave costs alongside other Cloud and AI infrastructure, group into Virtual Tags for accurate chargeback and showback, or utilize Anomaly Detection or Budgets.

The CoreWeave integration is now available to all Vantage customers. To connect, go to the Integrations page in account settings and select CoreWeave. You can learn more via the CoreWeave section in Vantage's documentation.

Frequently Asked Questions

1. What is being launched today?

Vantage is launching native support for CoreWeave costs. Customers connect a CoreWeave organization with an API access token, and Vantage imports daily usage by SKU, cluster, and region. Vantage will calculate price estimates based on per-SKU rates managed in the integration’s settings, which are defaulted to public rates and can be customized based on any negotiated rates.

2. Who has access to this integration?

The CoreWeave integration is available to all Vantage customers across all subscription tiers.

3. How much does this integration cost?

There is no additional fee for using the integration. However, CoreWeave costs will be included in quota tier enforcement. If your CoreWeave costs push you over your current tier limit, you may be prompted to upgrade. See the Pricing page for details.

4. How do I set up the integration?

Create an API access token in CoreWeave, then add it in Vantage:

1. In the CoreWeave Cloud Console, open the Access Tokens page and create an API access token for a user or service account with the Observability Viewer and Billing Viewer roles. Copy the token secret.
2. In Vantage, go to the Integrations page in account settings, select CoreWeave, and click Add Account.
3. Paste the token into the API Key field, enter a Description to identify the organization, and click Connect Account.

Vantage checks that the token can read CoreWeave's billing data before it begins importing. To enter your own rates, open the connection from the CoreWeave integration page and use the Rates section (see question 5).

5. How are CoreWeave costs calculated?

CoreWeave's FOCUS Usage Export API is currently in public preview and reports raw usage, not dollars. In order to create price estimates, Vantage provides a per-SKU price sheet in the integration settings, which is defaulted to CoreWeave's on-demand list price for that SKU. Each cost in Vantage is the billable usage times the rate for that SKU. Enterprises with negotiated rates per SKU can update this to accurately reflect their costs, and once updated, Vantage will reprocess usage data based on inputted rates. Storage and public IP rates are entered per month, the way CoreWeave prices them, and Vantage converts them to hourly rates using 730 hours per month. CoreWeave's export doesn't denote reservations, spot usage, flex pricing, credits, or tax, so enter your effective rate for any SKU where your price differs from list, including a blended rate if you mix capacity types. Vantage reports each day in UTC, while CoreWeave's billing month ends at 5:00 PM ET on its last day, so totals near a month boundary can differ slightly from the invoice.

Vantage plans to support native prices directly from CoreWeave when available.

6. What permissions are required on the CoreWeave side?

The token needs the Observability Viewer and Billing Viewer roles, and your CoreWeave organization needs CoreWeave's FOCUS billing export enabled. If the connection fails with an authorization error after the roles are assigned, contact CoreWeave Support to request FOCUS export access for your organization.

7. What permissions are needed within Vantage?

You must have a Vantage Owner or Integration Owner role to add or remove the CoreWeave integration or change its rates.

8. What permissions does Vantage have in my CoreWeave account?

Vantage utilizes the Observability Viewer and Billing Viewer roles, which have read-only access to metrics and logs, and read-only access to billing data. Vantage will only ever read usage and billing data for this integration.

9. Which CoreWeave usage is included?

Vantage imports the usage in CoreWeave's FOCUS billing export, which currently covers CoreWeave Kubernetes Service (CKS): GPU and CPU instances, storage, and public IP addresses. Vantage will pick up additional CoreWeave products as CoreWeave adds them to the export.

10. Why does some CoreWeave usage show no cost?

CoreWeave's export reports measured usage and billable usage separately. Vantage shows the measured usage, and multiplies only the billable usage by the SKU's rate. When CoreWeave excludes usage from billing, for example during a proof of concept or under a credit, billable usage is zero and that usage appears with no cost. SKUs without a published list price, such as those that CoreWeave lists as Contact sales, also show usage with no cost until you enter a rate for them.

11. What dimensions can CoreWeave costs be filtered and grouped by?

CoreWeave costs support filtering and grouping on:

• Organization (the CoreWeave organization ID)
• Cluster (the CoreWeave cluster ID)
• Service (such as GPU Compute, CPU Compute, Distributed File System, and Network)
• SKU (such as gd-8xh100ib-i128)
• Product Family (GPU Compute, CPU Compute, Storage, or Network)
• Region (such as us-east-14)
• Charge Type (Usage)
• Tags, including coreweave:cluster_name

In the Group By menu of a Cost Report, Organization, Cluster, SKU, and Product Family appear as Billing Account, Account, Category, and Subcategory.

12. Can I view usage data for CoreWeave?

Yes. Vantage imports measured usage in the unit CoreWeave reports, such as instance hours for GPU and CPU instances, GPU hours for CoreWeave's single-GPU inference SKUs, GiB hours for storage, and IP hours for public IPs. GPU instances are measured per instance, so one hour of an 8-GPU instance is one instance hour.

13. Are there Active Resources for CoreWeave?

No. Active Resources are not available for CoreWeave.

14. How often does CoreWeave data refresh in the Vantage console?

CoreWeave data refreshes daily in the Vantage console. During the first five days of each month, Vantage also refreshes the previous month. CoreWeave's export runs a few hours behind, and its most recent hours can change slightly until they settle.

15. How far back will the data go?

On initial connection, Vantage imports CoreWeave usage back to January 2026, the earliest month CoreWeave's FOCUS export provides, within your account's data retention period.

16. Can I connect multiple CoreWeave organizations?

Yes. Each API access token belongs to one CoreWeave organization, so connect each organization with its own token and Description. Rates are set separately for each connection.

17. Can I deploy the Vantage Kubernetes agent on CoreWeave?

Yes. The Kubernetes agent doesn't pick up prices from the CoreWeave integration, so to get pod-level costs on CoreWeave clusters, add hourly rate annotations to each node (vcpu_hourly_rate, ram_gb_hourly_rate, and gpu_gb_hourly_rate for GPU allocation) and include those keys in the agent's allowed annotations. See Enable Neocloud GPU Support for setup. The agent's pod-level costs appear under the Kubernetes provider and cover the same nodes this integration imports, so avoid including both the CoreWeave and Kubernetes providers for the same clusters in one Cost Report: use the Kubernetes provider for namespace, pod, and label allocation, and the CoreWeave provider for total CoreWeave spend. See Avoid Double Counting Costs in Cost Reports for details.

18. What happens if I remove the CoreWeave integration?

If you remove your CoreWeave integration from Vantage, all associated costs will be removed from the Vantage console. Removing the integration does not revoke the API access token in CoreWeave, so delete it in the CoreWeave Cloud Console if it's no longer needed.

19. Does Vantage access my clusters, workloads, or data?

No. Vantage only reads billing usage: SKUs, quantities, clusters, and regions. It doesn't access your clusters, workloads, models, or anything stored in CoreWeave.