Skip to content

Data Warehouse Pricing Compared (2026)

TL;DR

A cloud data warehouse holding 1 TiB and scanning 10 TiB a month costs roughly $40 to $115 on usage pricing as of October 30, 2026, on our arithmetic. ClickHouse Cloud Basic comes to about $38.39, BigQuery on-demand $79.57, Firebolt $82.23 and Redshift Serverless $114.58. MotherDuck Business reaches $434.96 because of its $250 platform fee, while Snowflake and Databricks bill credits and DBUs you need their calculators to price.

Eight warehouses priced for one month of a small analytics workload, with scan-based, compute-hour and platform-fee billing side by side.

Updated
Publications that cover the tools tracked here
  • TechCrunch
  • Bloomberg
  • The Information
  • Fortune
  • The Verge
12,800+ tools 400+ categories
Toolradar comparison of the monthly cost of cloud data warehouses for 1 TB of data

A small data warehouse is cheap in 2026, but the vendors meter completely different things, so headline rates do not compare. BigQuery charges for every TiB your queries scan beyond a free first TiB each month, while Redshift Serverless, ClickHouse Cloud, Firebolt and MotherDuck charge for the hours compute is running. Storage is the small part of the bill everywhere; how often your queries run, and whether compute shuts off between them, is the big part.

This guide prices one workload on every warehouse: 1 TiB of data stored, 10 TiB scanned by queries in a month, and dashboards refreshed every hour. For compute-hour tools we assume each hourly refresh keeps the smallest engine busy for 5 minutes, which is 720 refreshes and 60 compute hours a month. Prices are list rates in US dollars in the vendor's US East or US Central region as of October 30, 2026, and every total is our arithmetic.

Toolradar data: our data warehouse category lists 13 tools. 23% of them offer a free or freemium plan and 77% are paid only; the category statistics show the full split.

Methodology: we applied the same workload to the published rates of each warehouse, ranked them by the monthly total, and noted where a different workload shape flips the order. For background on how warehouses differ beyond price, read our data warehouse explainer.

Top Picks

Selected by pricing one monthly workload (1 TiB stored, 10 TiB scanned, dashboards refreshed every hour) on each warehouse's published storage, scan and compute-hour rates. How we rate

Firebolt logo

Cloud warehouse for low-latency analytics with per-second engines that auto-stop.

Starting price

Small engine from $0.92/hr

4.5 (14 reviews)

Best for

Customer-facing analytics that need quick queries on modest data.

Data Warehouse Pricing Compared (2026): BigQuery vs Redshift vs ClickHouse vs MotherDuck compared: starting price, rating and best use, as of November 2026. Each rating is the average of the review sites listed on that tool's card, weighted by the logarithm of their review counts.
Tool Starting price Rating Best for
ClickHouse Cloud Visit ClickHouse Cloud website From $0.2181 per unit-hour (Basic) 4.5 23 reviews Best for: Fast dashboards and real-time analytics on a budget.
Google BigQuery Visit Google BigQuery website $6.25/TiB scanned; first TiB free 4.5 1,258 reviews Best for: Teams with light or irregular query volume who want no servers to size.
Firebolt Visit Firebolt website Small engine from $0.92/hr 4.5 14 reviews Best for: Customer-facing analytics that need quick queries on modest data.
Amazon Redshift Visit Amazon Redshift website Serverless from $1.50/hr 4.3 420 reviews Best for: AWS teams that keep data in S3 and want SQL close to it.
MotherDuck Visit MotherDuck website Lite $0; Business $250/org/mo + usage n/a Best for: Small data teams who like DuckDB and want it shared and hosted.
Snowflake Visit Snowflake website Check Snowflake's site 4.1 865 reviews Best for: Companies that need data sharing, governance and many teams on one platform.
Databricks Visit Databricks website Check Databricks' site 4.6 1,385 reviews Best for: Teams running data engineering, ML and SQL on one lakehouse.
DuckDB Visit DuckDB website Free and open source n/a Best for: Analysts whose data fits on one machine.
1
ClickHouse Cloud logo
Top Pick

ClickHouse Cloud

Managed ClickHouse with separate storage and compute that scales to zero when idle.

  • 4.5 on G2 (23 reviews)
Starting price
From $0.2181 per unit-hour (Basic)
Free plan
No
Rating
4.5 (23 reviews)

Best for: Fast dashboards and real-time analytics on a budget.

  • Pro: Compute scales down to zero after inactivity, so idle hours cost nothing.
  • Pro: Storage is billed on compressed size at $25.30 per TB a month.
  • Pro: New organizations get $300 of credit over 30 days.
  • Con: Basic is capped at 1 TB of storage and one availability zone.
  • Con: Scale raises the compute rate to $0.2985 per unit-hour.
Good value

ClickHouse Cloud is genuinely cheap for read-heavy analytical workloads thanks to aggressive columnar compression and per-minute compute billing.

Watch out

Cross-region data transfer is billed separately. Multi-region architectures with replication can double or triple effective storage costs.

Used ClickHouse Cloud? Rate it
2
Google BigQuery logo

Google BigQuery

Serverless warehouse billed per TiB scanned, or per slot-hour on its editions.

  • 4.5 on G2 (1,223 reviews)
  • 4.6 on Capterra (35 reviews)
Starting price
$6.25/TiB scanned; first TiB free
Free plan
No
Rating
4.5 (1,258 reviews)

Best for: Teams with light or irregular query volume who want no servers to size.

  • Pro: The first 1 TiB of queries and 10 GiB of storage are free every month.
  • Pro: On-demand queries share up to 2,000 slots per project with no sizing.
  • Pro: Editions let heavy users switch to slot-hours without migrating.
  • Con: A query that reads a full unpartitioned table bills every byte it scans.
  • Con: The old flat-rate model ended in July 2023, so steady users must size edition slots.
Good value

BigQuery's on-demand pricing ($6.25/TiB queried) with 1 TiB free per month is one of the most generous free tiers in cloud data warehousing.

Watch out

Minimum 10MB billed per table referenced: a query touching 50 small tables is billed for at least 500MB even if actual data scanned is 1MB

Used Google BigQuery? Rate it
3
Firebolt logo

Firebolt

Cloud warehouse for low-latency analytics with per-second engines that auto-stop.

  • 4.5 on G2 (14 reviews)
Starting price
Small engine from $0.92/hr
Free plan
No
Rating
4.5 (14 reviews)

Best for: Customer-facing analytics that need quick queries on modest data.

  • Pro: Compute is billed per second and engines stop when idle.
  • Pro: Storage is pass-through object storage at $0.0264 per GB a month.
  • Pro: Charged on compressed data size.
  • Con: An engine left running all month costs about $672 on the vendor's own estimate.
  • Con: The small engine rate quoted is for US East on Arm; other sizes and regions differ.
Good value

The Core tier is very generous for those comfortable with self-management.

Used Firebolt? Rate it
4
Amazon Redshift logo

Amazon Redshift

AWS warehouse sold as serverless RPU-hours or provisioned nodes.

  • 4.3 on G2 (404 reviews)
  • 4.4 on Capterra (16 reviews)
Starting price
Serverless from $1.50/hr
Free plan
No
Rating
4.3 (420 reviews)

Best for: AWS teams that keep data in S3 and want SQL close to it.

  • Pro: Serverless shuts down when idle and charges nothing for start-up time.
  • Pro: Base capacity adjusts from 4 to 1,024 RPUs.
  • Pro: First-time Serverless users get a $300 credit valid for 90 days.
  • Con: Each query is billed for at least 60 seconds.
  • Con: Provisioned clusters bill every hour, which makes idle time expensive.
Good value

Amazon Redshift's pricing is fair, especially with the free trial and reserved instance options.

Watch out

Storage costs beyond compute

Used Amazon Redshift? Rate it
5
MotherDuck logo

MotherDuck

Cloud data warehouse built on DuckDB, with a free Lite plan and a Business plan.

  • 4.5 on G2 (1 reviews)
Starting price
Lite $0; Business $250/org/mo + usage
Free plan
No

Best for: Small data teams who like DuckDB and want it shared and hosted.

  • Pro: Lite is free with 10 GB of storage and 10 hours of Pulse compute.
  • Pro: Pulse compute is metered per query at $0.60 per compute-unit hour.
  • Pro: Read-scaling replicas serve BI dashboards on Business.
  • Con: The Business platform fee dominates a small workload's bill.
  • Con: Business allows up to 10 internal active users before Enterprise pricing.
Good value

It's best for users prioritizing flexibility and a serverless approach to analytics.

Used MotherDuck? Rate it
6
Snowflake logo

Snowflake

Multi-cloud data platform billed in credits for compute plus a monthly storage fee.

Starting price
Check Snowflake's site
Free plan
No
Rating
4.1 (865 reviews)

Best for: Companies that need data sharing, governance and many teams on one platform.

  • Pro: Standard, Enterprise and Business Critical editions add features step by step.
  • Pro: Buy on demand or as pre-paid capacity at a discount.
  • Pro: Storage is billed on the compressed monthly average.
  • Con: The credit price depends on edition, cloud and region, so you need its calculator.
  • Con: Multi-cluster compute requires Enterprise or higher.
Good value

The model rewards disciplined warehouse management and punishes always-on clusters.

Watch out

Serverless features (Snowpipe, tasks, materialized views) consume credits automatically in the background. Cloud services usage is free up to 10% of your daily compute spend, but exceeding that threshold triggers additional charges that are hard to predict.

Used Snowflake? Rate it
7
Databricks logo

Databricks

Lakehouse platform whose serverless SQL warehouse is billed in DBUs.

  • 4.6 on G2 (1,362 reviews)
  • 4.5 on Capterra (23 reviews)
Starting price
Check Databricks' site
Free plan
No
Rating
4.6 (1,385 reviews)

Best for: Teams running data engineering, ML and SQL on one lakehouse.

  • Pro: One platform for pipelines, notebooks, ML and SQL.
  • Pro: Committed-use contracts can lower the DBU rate.
  • Pro: Storage has its own standard meter, the Databricks Storage Unit.
  • Con: DBU consumption varies by product and workload, which makes forecasting harder.
  • Con: Overkill for a team that only needs dashboards.
Fair value

It's best for enterprises with significant data processing needs and budgets to match.

Watch out

Separate cloud infrastructure bill on top of DBUs, varying with instance types and usage

Used Databricks? Rate it
8
DuckDB logo

DuckDB

Free, open-source analytical database that runs inside your process or laptop.

Starting price
Free and open source
Free plan
Free to use

Best for: Analysts whose data fits on one machine.

  • Pro: MIT-licensed with no usage fees.
  • Pro: Reads Parquet and CSV files directly, so small teams can skip loading data.
  • Pro: The same engine powers MotherDuck if you outgrow one machine.
  • Con: No shared server for a team of dashboard users.
  • Con: You manage backups, access control and scheduling yourself.
Great value

DuckDB is entirely free and open-source, making its pricing exceptionally fair and generous compared to any commercial offering.

Used DuckDB? Rate it

Other Data Warehouses tools worth considering

More published tools from our Data Warehouses category, ordered by our category ranking. They are not part of the editorial picks above.

What a cloud data warehouse bills for

A cloud data warehouse stores your business data in columnar tables and runs SQL over it for dashboards, reports and models. Every vendor here separates storage from compute, so you pay for each on its own meter.

Scan-based billing. BigQuery's on-demand model charges for the bytes each query reads. A badly written query that scans a whole table costs the same whether it takes two seconds or two minutes, so partitioning and selecting only the columns you need cut the bill directly.

Compute-hour billing. Redshift Serverless, ClickHouse Cloud, Firebolt and MotherDuck charge for the time an engine runs, by the second or minute. A fast query on a small engine is cheap, and an engine that shuts off when idle costs nothing between refreshes.

Credits and units. Snowflake sells credits consumed by virtual warehouses, and Databricks sells DBUs, a normalized unit of processing. Both add an edition or product choice on top, so the price per hour depends on more than size.

One month of the same workload on each warehouse

Every figure below is our arithmetic on list prices as of October 30, 2026: 1 TiB stored, 10 TiB scanned, 60 compute hours.

ClickHouse Cloud Basic: one compute unit (8 GiB of RAM, 2 vCPUs) at $0.2181 an hour for 60 hours is $13.09, plus $25.30 for 1 TB of storage: $38.39 a month. ClickHouse bills storage on the compressed size, so the real storage line is often far lower.

BigQuery on-demand: the first TiB scanned is free, so 9 TiB x $6.25 = $56.25. Active logical storage costs $23.55 per TiB a month in us-central1, about $23.32 after the free 10 GiB. Total: $79.57 a month.

Firebolt: a small compute-optimized engine at $0.92 an hour for 60 hours is $55.20, and 1,024 GB of storage at $0.0264 per GB is $27.03: $82.23 a month.

Redshift Serverless: the minimum base of 4 RPUs at $0.375 per RPU-hour in US East costs $1.50 an hour, so 60 hours is $90. Managed storage at $0.024 per GB-month adds $24.58: $114.58 a month. A provisioned cluster at the lowest listed rate, $0.543 an hour, costs $396.39 a month left running, before storage.

MotherDuck Business: the $250 platform fee, a Standard instance at $2.40 an hour for 60 hours ($144), and 1,024 GB of storage ($40.96): $434.96 a month.

BigQuery slots instead of on-demand: Standard edition costs $0.04 per slot-hour with no commitment. A 50-slot reservation busy for the same 60 hours is 3,000 slot-hours, or $120, so on-demand stays cheaper until roughly 20 TiB scanned a month on these assumptions.

Key Features to Look For

  • Billing unit (Essential)

    Bytes scanned (BigQuery on-demand), compute hours (Redshift Serverless, ClickHouse, Firebolt, MotherDuck) or credits and DBUs (Snowflake, Databricks). Pick the unit your workload uses least of.

  • Scale to zero (Essential)

    ClickHouse Cloud and Redshift Serverless stop charging for compute when idle, and Firebolt engines auto-stop. A provisioned Redshift cluster bills every hour it exists.

  • Minimum billing increment (Essential)

    Redshift Serverless charges at least 60 seconds per query; ClickHouse meters compute per minute; MotherDuck and Firebolt bill per second.

  • Storage measured compressed or logical (Important)

    ClickHouse and Firebolt bill compressed size; BigQuery bills logical bytes by default, with physical (compressed) storage as a separate rate.

  • Free tier and trial credit (Important)

    BigQuery gives 1 TiB of queries and 10 GiB of storage free each month; ClickHouse Cloud and Redshift Serverless each offer $300 of trial credit.

  • Platform fee (Important)

    MotherDuck Business charges a fixed monthly fee per organization before any usage; most others charge only for usage.

  • Concurrency for dashboards (Nice to have)

    MotherDuck sells read-scaling replicas for BI traffic, and BigQuery on-demand shares up to 2,000 slots per project.

Pick by workload shape

  1. Light, sporadic queries on a few hundred GB: BigQuery on-demand or MotherDuck Lite can cost close to nothing.

  2. Hourly dashboards on a small dataset: a scale-to-zero engine such as ClickHouse Cloud or Redshift Serverless keeps compute bills tied to refresh time.

  3. Heavy scanning, more than about 20 TiB a month: move BigQuery to edition slots or pick a compute-hour engine.

  4. Already on AWS with data in S3: Redshift Serverless avoids moving data between clouds.

  5. Data that fits on one machine and one analyst: DuckDB on a laptop costs nothing and needs no warehouse at all.

  6. Mixed SQL, data engineering and ML teams: Snowflake or Databricks, priced through their calculators for your region and edition.

Evaluation Checklist

  • Measure how many TiB your queries scanned last month, not only how much data you store.

  • Count dashboard refreshes per day and how long each one runs.

  • Check the idle timeout before compute scales to zero.

  • Compare logical and compressed storage sizes for your largest tables.

  • Price the same workload at monthly and committed rates.

  • Load a sample of real data during a trial and read the bill after a week.

Pricing Overview

Free and open source

DuckDB on your own machine, MotherDuck Lite with 10 GB of storage and 10 hours of Pulse compute, and BigQuery's free monthly 1 TiB of queries.

$0

Scan-based

BigQuery on-demand: no servers to size, cost follows the bytes each query reads.

Per TiB scanned

Compute-hour, scale to zero

ClickHouse Cloud, Firebolt, Redshift Serverless and MotherDuck instances, billed only while queries run.

$0.22-$2.40/hr

Provisioned clusters

Redshift provisioned nodes and BigQuery slot reservations for steady, predictable load.

From $0.543/hr

Credit and DBU platforms

Snowflake editions and Databricks SQL, priced per credit or DBU by region and edition.

Check vendor calculators

Pricing Comparison

Data Warehouse Pricing Compared (2026): BigQuery vs Redshift vs ClickHouse vs MotherDuck pricing comparison, as of November 2026
Warehouse and plan Billing used Workload, per month Main rates

Per minute, scale to zero

$38.39

$0.2181 per unit-hour, $25.30 per TB stored

BigQuery on-demand

Per TiB scanned

$79.57

$6.25 per TiB after 1 TiB free, $23.55 per TiB stored

Firebolt, small engine

Per second, auto-stop

$82.23

$0.92 per hour, $0.0264 per GB stored

Redshift Serverless (4 RPU)

Per second, 60-second minimum

$114.58

$0.375 per RPU-hour, $0.024 per GB stored

BigQuery Standard edition, 50 slots

Per slot-hour

$143.32

$0.04 per slot-hour plus storage

Redshift provisioned, lowest node

Hourly, always on

$396.39 + storage

From $0.543 per hour

MotherDuck Business

Monthly fee + usage

$434.96

$250 per org, $2.40 per Standard hour, $0.04 per GB

Snowflake

Credits

Check Snowflake's site

Credits by warehouse size and edition

Databricks SQL

DBUs

Check Databricks' site

DBUs by product and cloud

Our arithmetic for 1 TiB stored, 10 TiB scanned and 60 compute hours in one month, using list prices as of October 30, 2026 on ClickHouse Cloud pricing, BigQuery pricing, Firebolt pricing, Amazon Redshift pricing, MotherDuck pricing, Snowflake pricing and Databricks pricing. Storage is counted at 1 TiB uncompressed; compressed billing lowers it. Taxes, data transfer and trial credits are excluded.

Mistakes to Avoid

  • ×

    Comparing storage prices when compute hours or scanned bytes drive most of the bill.

  • ×

    Leaving a provisioned cluster running nights and weekends.

  • ×

    Running SELECT * on BigQuery on-demand, which bills every column read.

  • ×

    Setting a long idle timeout so a scale-to-zero engine never sleeps between hourly refreshes.

  • ×

    Buying a platform plan with a fixed fee for a workload that runs a few hours a month.

Expert Tips

  • →

    Partition and cluster large BigQuery tables by date so dashboards scan days, not years.

  • →

    Schedule dashboard refreshes on the same hour so one engine wake-up serves all of them.

  • →

    Set usage limits: Redshift Serverless Max RPU-hours and ClickHouse autoscaling caps stop runaway bills.

  • →

    Keep raw history in object storage and load only what dashboards query.

  • →

    Re-run this arithmetic every quarter with your real scan volume and compute hours.

Red Flags to Watch For

  • A compute engine with no auto-stop or idle timeout.

  • Prices that only appear in a calculator after you pick a region and edition.

  • Minimum charges per query that multiply small dashboard queries.

  • Platform fees billed whether or not you run a query.

  • Egress fees for exporting your own data.

The Bottom Line

For a small analytics workload, a usage-priced warehouse costs under $120 a month on list prices: ClickHouse Cloud is the cheapest here, BigQuery and Firebolt cost about the same, and Redshift Serverless is the pick for teams already on AWS. MotherDuck makes sense once a team uses the Business features its fee pays for, and Snowflake and Databricks belong on the list when governance, sharing or ML matter more than the smallest bill.

Compare the leaders head to head in BigQuery vs Snowflake, Amazon Redshift vs BigQuery and ClickHouse vs Snowflake, check Snowflake pricing and BigQuery pricing, weigh Amazon Redshift vs Snowflake, or browse Snowflake alternatives. If your app data lives in Postgres, our serverless Postgres pricing comparison covers the operational side, the business intelligence tools guide covers the dashboards on top, and ETL tool pricing compared prices the pipelines that load it.

Cite this: Toolradar, "Data Warehouse Pricing Compared (2026)", October 2026.

Frequently Asked Questions

How much does a data warehouse cost per month for 1 TB in 2026?

For 1 TiB stored, 10 TiB scanned and hourly dashboards, our arithmetic on October 2026 list prices gives about $38.39 on ClickHouse Cloud Basic, $79.57 on BigQuery on-demand, $82.23 on Firebolt and $114.58 on Redshift Serverless.

How much does BigQuery cost?

BigQuery on-demand charges $6.25 per TiB scanned after the first free TiB each month. Active logical storage costs about $23.55 per TiB a month in us-central1, with 10 GiB free. Standard edition slots cost $0.04 per slot-hour with no commitment.

When are BigQuery slots cheaper than on-demand?

When you scan a lot. In our example, 50 Standard edition slots busy 60 hours a month cost $120, which matches on-demand at about 20 TiB scanned. Below that, on-demand is cheaper; above it, slots win.

Is there a free data warehouse?

DuckDB is free and open source for data that fits on one machine. MotherDuck Lite is free with 10 GB of storage and 10 hours of compute a month, and BigQuery includes 1 TiB of queries and 10 GiB of storage free every month.

Is Redshift Serverless cheaper than provisioned Redshift?

For intermittent workloads, yes. Serverless at its 4 RPU minimum costs $1.50 an hour only while queries run, so 60 hours is $90. The cheapest provisioned node at $0.543 an hour costs $396.39 a month if left on.

How is Snowflake priced?

Snowflake bills compute in credits consumed by virtual warehouses and charges a monthly storage fee on the compressed average stored. The credit price depends on the edition (Standard, Enterprise, Business Critical), the cloud and the region, so check the current price on Snowflake's site.

Cite this page: Toolradar, "Data Warehouse Pricing Compared (2026)", updated November 2026, https://toolradar.com/guides/data-warehouse-pricing-compared

Spotted an outdated price or a wrong fact? Suggest a correction: we check it against the vendor's own page. Work for one of these vendors? Claim your listing to update it yourself.

Related Guides

Ready to Choose?

Compare features, read reviews, and find the right tool.

Some offers on this page may be paid placements or contain affiliate links.