Firebolt Pricing, Plans, and Availability

#7 in Data Warehouse

by Firebolt · firebolt.io

Cloud data warehouse built for low-latency analytics on large datasets.

Visit website

Pricing at a glance

◇ Pricing model

Usage-based managed pricing with per-second compute billing and pass-through storage pricing; free self-hosted Firebolt Core is also publicly described.

↻ Billing notes

Firebolt publicly describes managed-service compute as billed per second, with scale-to-zero behavior and auto start/stop. Storage is billed separately on a pass-through object-storage basis. The official pricing page also mentions $200 free credits, and the comparison pages describe Firebolt Core as forever free and self-hosted. The supplied documents do not disclose a public monthly seat minimum, commitment discount, or renewal term, so those details are not publicly disclosed in the available sources.

Plans and pricing

10 tiers

Firebolt Core

Not publicly disclosed
Not publicly disclosed
  • Self-hosted edition
  • Full query engine capabilities
  • Same performance and features as managed service
  • Deploy anywhere: local laptop, cloud, datacenter, Kubernetes

Public materials describe it as forever free and without usage restrictions except building competing SaaS, but no seat-based or consumption price is disclosed.

Managed service — compute

Per second
Not publicly disclosed
  • Engine-based compute billing
  • Scale to zero
  • Auto start and auto stop
  • Auto scaling
  • Workload isolation
  • Online version upgrades

Engines can scale up from S to XL and out across 1 to 256 nodes; concurrency can use up to 10 parallel clusters. Exact dollar amounts vary by configuration and region and are shown in the calculator rather than as a single universal list price.

Managed service — storage

Monthly
$0.0264 / gb · month
  • Object storage pass-through at cost

Pricing is shown for object storage and the page notes it is pass-through at cost. No additional bundled storage allowance is publicly disclosed in the supplied documents.

Storage-optimized engine S

Hourly
$1.84 / hour
  • 1 node
  • 1 vCPU
  • 8 GB memory
  • 64 GB disk

US East (N. Virginia) example pricing shown on the official site; other regions are listed but not priced in the supplied text.

Storage-optimized engine M

Hourly
$3.68 / hour
  • 1 node
  • 16 vCPU
  • 128 GB memory
  • 3,750 GB disk

US East (N. Virginia) example pricing shown on the official site.

Storage-optimized engine L

Hourly
$7.36 / hour
  • 1 node
  • 32 vCPU
  • 256 GB memory
  • 7,500 GB disk

US East (N. Virginia) example pricing shown on the official site.

Storage-optimized engine XL

Hourly
$14.72 / hour
  • 1 node
  • 64 vCPU
  • 512 GB memory
  • 15,000 GB disk

US East (N. Virginia) example pricing shown on the official site.

Storage-optimized engine 2L

Hourly
$14.72 / hour
  • 2 nodes
  • 64 vCPU
  • 512 GB memory
  • 15,000 GB disk

US East (N. Virginia) example pricing shown on the official site.

Storage-optimized engine 2XL

Hourly
$29.44 / hour
  • 2 nodes
  • 128 vCPU
  • 1,024 GB memory
  • 30,000 GB disk

US East (N. Virginia) example pricing shown on the official site.

Storage-optimized engine 4XL

Hourly
$58.88 / hour
  • 4 nodes
  • 256 vCPU
  • 2,048 GB memory
  • 60,000 GB disk

US East (N. Virginia) example pricing shown on the official site.

Add-ons and extras

Free credits

$200 free credits

The official pricing page says new users get $200 free credits. No duration, expiration, or eligibility rules are disclosed in the supplied text.

Hidden and indirect costs

Worth budgeting for
⚠  The official materials indicate that storage is charged separately as pass-through object storage, which means the storage bill can vary with the cloud provider and actual consumed storage. That can make total cost higher or lower than engine-only pricing suggests, depending on data volume and retention. The pricing page also emphasizes per-second billing and autoscaling, so idle usage may be cheaper than always-on clusters, but customers still need to watch engine configuration, node count, and region selection.
⚠  The managed-service pages show regional pricing and an engine calculator rather than a single universal list price, so buyers should expect configuration-dependent costs. Additional operational costs may arise from choosing larger nodes, more nodes, more clusters for concurrency, or storage-optimized versus compute-optimized hardware. The supplied documents do not disclose all region-specific rates, so a full quote may be needed for precise budgeting.

What buyers actually pay

A team wants a small managed analytics engine for light production workloads.

Expected costNot publicly disclosed; the official site exposes configuration-based hourly engine pricing rather than a single flat subscription.
Engine sizeNumber of nodesStorage-optimized versus compute-optimized hardwareRegionData volume

A team stores a large compressed dataset and expects storage charges to be a meaningful part of the bill.

Expected costStorage is priced at $0.0264 / gb · month, plus compute charges if engines are running.
Compressed storage volumeRetention durationCloud object-storage pass-through ratesCompute uptime

A startup wants to test Firebolt without committing to paid infrastructure.

Expected cost$200 free credits and a forever-free self-hosted Firebolt Core option are publicly described.
Whether the team uses managed service or self-hosted CoreCompute and storage usage if using managed service

Pricing FAQ

No. The supplied official pricing page shows configuration-based engine pricing, per-second compute billing, and separate storage pricing instead of a single universal subscription fee. That means buyers should use the engine calculator or request a quote for a precise number. The public documents also indicate that Firebolt Core is forever free and self-hosted.

Yes. The official materials mention $200 free credits on the pricing page, and the comparison pages describe Firebolt Core as a forever free, self-hosted edition. The supplied documents do not provide an expiration period or eligibility details for the credits. They also do not describe a free trial length beyond the credit offer.

Compute is billed per second according to the official pricing page. The page also says engines can scale to zero and support auto start and auto stop, which suggests customers only pay for active compute time. Exact total spend still depends on the chosen engine size, node count, and region.

Storage is billed separately as pass-through object storage at $0.0264 / gb · month in the supplied official text. The pricing page frames this as compressed-size billing, so the amount of data stored is what matters, not the raw uncompressed footprint. Because the storage is pass-through at cost, underlying cloud storage charges can affect the final bill.

Next: Alternatives