Verda GPU pricing

DataCrunch is now Verda. Verda's own announcement says the name change began in November 2025 and did not immediately change services, workflows or billing, with invoices and documents moving to the Verda name over time. The current Verda AI Cloud covers GPU instances, Instant Clusters, bare metal, serverless containers and managed endpoints, with pay-as-you-go and long-term contract models.

Live Verda prices (2026-07-08)

Cheapest listed per-GPU-hour rates verified at the source. Before storage, egress and networking.

Pricing model

Verda documents GPU pricing through its current products and pricing page, and its docs describe both long-term and pay-as-you-go payment models. Long-term contracts for GPU Instances and Instant Clusters are paid up front. Pay-as-you-go resources are billed in prepaid 10-minute increments, with unused portions refunded within the next billing period if a resource is terminated before the billed period ends.

Capacity tiers

Verda's AI Cloud page lists GPU instances, Instant Clusters, bare metal, serverless containers and managed endpoints. It positions the platform for the full AI lifecycle, from single-GPU prototyping to multi-node training and production inference. The platform also includes block storage, shared filesystem, container registry and object storage marked as coming soon on the AI Cloud page.

GPU lineup

Verda's AI Cloud page lists GB300 NVL72, B300, B200, H200, H100, A100, RTX PRO 6000, NVLink and InfiniBand. The confidential-compute docs currently list RTX PRO 6000 single-GPU support for confidential computing, with additional Blackwell configurations planned. Availability and exact pricing should be checked on the live pricing page because Verda has been expanding and renaming the platform during the rebrand.

Storage, egress and hidden costs

Verda documents block storage, shared filesystem and container registry as platform services, with object storage marked as coming soon on the AI Cloud page. The billing docs warn that services are prepaid and limited by available balance. If balance reaches zero, instances are discontinued and volumes can be deleted, with a documented restore window. Public egress pricing was not found in the official public docs reviewed. Not publicly documented.

Reliability and support posture

Verda's AI Cloud page claims historical uptime over a published threshold, sensible SLAs, fair compensation for disruptions, proactive support from infrastructure and ML engineers, and security and compliance certifications including SOC 2 Type II and ISO 27001 family certifications. Buyers should verify the current Trust Center documents and contract terms, especially because the DataCrunch to Verda transition is still reflected across public materials.

Free credits and referral programs

Verda docs include a Free Compute Credits section and say credit coupons may be offered during promotions, events and hackathons. The pricing and billing docs link to earning and redeeming credits. A general public referral program and fixed always-on startup credit amount were not found in the public pages reviewed. Not publicly documented.

Who should pick Verda

Pick Verda if you previously evaluated DataCrunch, need a European AI cloud posture, and want a platform that spans single GPU instances, clusters, bare metal, serverless containers and managed endpoints. It is most relevant for teams that care about current NVIDIA datacenter hardware, EU-oriented operations, and hands-on support. Check current naming, balance behavior, egress terms and live capacity before migrating production workloads.

FAQ

Profile verified 2026-07-06. Sources: verda.com · verda.com · verda.com · docs.verda.com · docs.verda.com · docs.verda.com · docs.verda.com · trust.verda.com · status.verda.com