Skip to content

Pricing & Billing

Carolina Cloud bills per hour of usage based on the resources you provision. There are no egress fees, no bandwidth charges, and no minimum commitments.

ResourceHourly rate
vCPU (general purpose)$0.005 / vCPU / hr
vCPU (high performance)$0.01 / vCPU / hr
RAM$0.005 / GiB / hr
Disk$0.0001 / GiB / hr
GPUVaries by model

Your hourly cost is the sum of all provisioned resources:

Hourly cost = ($0.005 × vCPUs) + ($0.005 × RAM GiB) + ($0.0001 × Disk GiB)

Example: A running instance with 8 vCPUs, 16 GiB RAM, and 200 GiB disk costs:

($0.005 × 8) + ($0.005 × 16) + ($0.0001 × 200) = $0.14/hr ≈ $3.36/day ≈ $102.2/mo
  • Running: Full cost (CPU + RAM + Disk + GPU).
  • Stopped: Storage only. Stop an instance and you immediately stop paying for compute. You keep your disk and data. Containers can do this automatically — see Auto-Stop on Idle.
  • Deleted: Nothing. Billing stops completely.

Billing is metered per second — the hourly rates above are just the unit we quote. There is no minimum billing increment and nothing is ever rounded up to the next hour. When you stop, resize, or delete an instance, you pay for exactly the seconds it ran up to that moment.

Carolina Cloud bundles the Sentieon Genomics Suite for high-performance variant calling. Starting July 1, 2026, you pay $0.06 / vCPU / hr — but only while Sentieon is actually running. It’s free for everyone until then.

This is metered by actual use, not a flat add-on for having the toolkit:

  • You pay only for the time a Sentieon tool is actually computing. While your pipeline downloads FASTQs, builds indexes, or runs non-Sentieon steps (GATK, BWA, samtools, your own code), there is no Sentieon charge — even on a genomics instance.
  • It scales with the instance’s vCPUs: a 16-vCPU box costs an extra $0.96 / hr during a Sentieon step, and $0 the rest of the time.
  • It applies to any instance that runs Sentieon — genomics containers and VMs provisioned with the suite.
  • Only the Sentieon compute tools are metered; the bundled license server doesn’t count.

The instance card shows a live “Sentieon running” indicator (with the current surcharge) whenever Sentieon is active, so you always know when the meter is on. The indicator may lag a start or stop by a few minutes.

Don’t want to pay for Sentieon? Just don’t run it. The rest of the genomics toolkit is billed at normal compute rates. See Genomics Container for details.

Warm S3 storage (Wasabi) is billed at $9/TB/month based on stored bytes, prorated hourly. Sizes follow Wasabi’s convention where 1 GB = 1024³ bytes and 1 TB = 1024⁴ bytes.

There is a separate, conditional charge for files deleted before they reach Wasabi’s 90-day minimum retention age. Most accounts never see it — see Storage Deletion Policy for the full rules.

Cold S3 storage is billed at $1.55/TB/month, prorated hourly, with a 108 TB minimum per user once cold storage is enabled. The minimum applies whether or not your cold bucket actually holds that much data.

Cold storage is opt-in — see Cold Storage for how to turn it on.

Every new account receives $5.00 in trial credits — no credit card required. Trial credits are consumed before any payment method is charged.

When trial credits are exhausted:

  • If you have a payment method on file, billing switches to Stripe automatically.
  • If you do not have a payment method, all running instances are destroyed to prevent unexpected charges.

Organizations that pre-purchase credits earn interest on the purchased balance. We pay a SOFR-referenced rate, compounded daily — by default the full SOFR rate, which typically beats a bank deposit. Interest applies only to an org’s purchased prepaid credits, never to free trial/promo credits, and is credited automatically with no opt-in or minimum term.

See Interest on Prepaid Credits for the rate, the day-count math, and where to see your accrued interest.

We use Stripe for payment processing. You can add a credit card or debit card from the billing section of the dashboard. Invoices are generated automatically.

Two CPU tiers are available:

  • General Purpose — Standard performance, suitable for most workloads.
  • High Performance — Higher clock speed CPUs (5+ GHz) for compute-intensive tasks.