Snowflake Pricing Examples: Three Worked Monthly Bills (2026)

A Snowflake bill is credits times a price per credit, plus storage by the terabyte, data transfer and, for AI, a separate AI Credit. These three illustrative Snowflake pricing examples build one month line by line, without data transfer or unlisted lines: a small team at $750.70, a mid-size platform at $7,452.70 and a larger account with Cortex AI at $33,201.36, all on Enterprise Edition, on demand in AWS US East (N. Virginia), at $3.00 per credit. Every fact and figure in this guide was checked against Snowflake’s documentation.

This article is part of our guide to Snowflake pricing explained.

Which prices do these Snowflake pricing examples use?

All three illustrative bills use Enterprise Edition, on demand, in AWS US East (N. Virginia). Prices come from Snowflake’s Service Consumption Table, effective 29 October 2026: $3.00 per Platform Credit, $23.00 per TB of storage per month, and $2.00 per AI Credit with global routing (ANY_REGION). Edition, region and contract set the credit price; region and contract set storage.

The AI price carries a condition. Bill 3 (Enterprise Edition, on demand, AWS US East) assumes the account sets the CORTEX_ENABLED_CROSS_REGION parameter to ANY_REGION, which Snowflake prices at $2.00 per AI Credit. Snowflake lists claude-sonnet-5 under cross-region inference, not as native to AWS US East. The default depends on your account, and an explicit setting overrides it, so check the value in yours. Accounts in organizations created after 9 March 2026 default to ANY_REGION. Other AWS accounts in the United States or Canada default to AWS_US, a regional setting at $2.20 per AI Credit. At $2.20, Bill 3’s AI line is $1,980.00 and its total is $33,381.36. Snowflake sets the AI Credit price by routing, not by edition or region. Capacity contracts price them by contract size, from $2.00 down to $1.88 with global routing, not at a flat $2.00, so check your order form.

Other rates and assumptions

The other rates come from the same table:

• Gen2 warehouses: Gen2 is Snowflake’s newer standard warehouse. Since Snowflake’s 2026_03 behavior change, new standard warehouses default to Gen2 wherever it is available, while existing Gen1 warehouses stay Gen1 until you change them. The three bills assume Gen2 warehouses, so check the resource_constraint column in SHOW WAREHOUSES for your own. On AWS, Gen2 uses 1.35 times the credits of Gen1: X-Small 1.35, Small 2.7, Medium 5.4, Large 10.8 and X-Large 21.6 per hour.
• Snowpipe: 0.0037 credits per GB loaded.
• Optima Clustering: 0.007 credits per uncompressed GB ingested, times an overlap factor from 0 to 1.
• Query Acceleration: 1 credit per compute-hour.
• AI_COMPLETE with claude-sonnet-5: 1.20 AI Credits per million input tokens and 6.00 per million output tokens.

Every hour, gigabyte and token count below is an assumption, not customer data. The bills leave out data transfer and any other line not listed in a table. A month has 30 days, and 22 of them are working days.

How does Snowflake bill warehouse time in these examples?

Snowflake bills a running warehouse per second, with a 60-second minimum each time it starts or resumes. Idle time before auto-suspend counts as running time. So a cycle that runs 210 seconds bills 210 seconds, and a warehouse that suspends within its first minute still bills 60 seconds.

Every warehouse line uses one formula: hours per day x days x credits per hour = credits, then credits x $3.00 (Enterprise Edition, on demand, AWS US East) = dollars. A warehouse uses credits while it runs, even when no query is running, so the hours include idle tails.

For example, take Bill 1’s loading warehouse. It resumes 12 times a day. Each resume runs 150 seconds of work plus a 60-second idle tail, so Snowflake bills 210 seconds, not four whole minutes. Then twelve resumes make 2,520 seconds, or 0.7 hours a day. Snowflake checks for suspension about every 30 seconds, so a real idle tail runs a little longer. Even a warehouse that runs for only 10 seconds pays the minimum, which is 0.0225 credits on a Gen2 X-Small, but a 10-second query followed by the 60-second idle tail bills 70 seconds.

What does a Snowflake pricing example look like for a small BI-and-loading team?

In this illustration, a small team with one BI warehouse and one loading warehouse pays $750.70 a month on Enterprise Edition, on demand in AWS US East. The warehouses cost $704.70, storage adds $46.00, and cloud services cost nothing because they stay under the 10% adjustment.

The BI warehouse runs 6 hours on each working day, idle tail included. Snowflake bills table data on its compressed size, and storage cost also covers Time Travel and Fail-safe history. Cloud services use is assumed at 5% of each day’s warehouse credits. Snowflake charges for cloud services only when a day’s use passes 10% of that day’s warehouse credits, so this line is zero.

Snowflake applies this rule day by day, so a month with uneven days can bill some cloud services even when the monthly share looks low. For example, take illustrative Bill 1 (Enterprise Edition, on demand, AWS US East). Cloud services use equal to 5% of the month’s warehouse credits, spread evenly over 30 days, would bill 1.62 credits, or $4.86, because the 8 non-working days use little warehouse time.

What does a mid-size platform with serverless features pay?

In this illustration, a mid-size platform with three Gen2 warehouses, Snowpipe loading, newly clustered tables and 35 TB of storage pays $7,452.70 a month on Enterprise Edition, on demand in AWS US East. Warehouses cost $6,528.60 of that, Snowpipe and Optima Clustering add $119.10, and storage adds $805.00.

Cloud services use is assumed at 8% of each day’s warehouse credits, so no day passes the 10% line. For Snowpipe, billed size means uncompressed size for text files and observed size for binary files such as Parquet. The 6,000 GB here is billed size. Since 1 September 2026, newly clustered tables use Optima Clustering, while older clustered tables stay on Clustering Classic, which bills by compute-hour. Here, an overlap factor of 1 is the worst case, where every ingested byte needs clustering. Our guide to Snowflake serverless costs lists every serverless rate.

Query Acceleration is on by default for new Gen2 and multi-cluster warehouses on Enterprise, and it bills only when a query uses it. Bills 1 and 2 assume none did.

What does a larger account with Cortex AI pay?

In this illustration, a larger account with a multi-cluster BI warehouse, Query Acceleration, Snowpipe, 180 TB and Cortex AI pays $33,201.36 a month on Enterprise Edition, on demand in AWS US East, with AI at the $2.00 global routing price. Warehouses cost $25,434.00 of that, cloud services add $1,017.36 and AI adds $1,800.00.

A multi-cluster warehouse bills every running cluster at the full rate for its size. Here one cluster runs 12 hours and a second runs 3 hours each working day, so the BI line counts 15 cluster-hours. Cloud services use is assumed at 14% of each day’s warehouse credits. Snowflake lists patterns such as high-frequency SHOW commands from applications and third-party tools. Snowflake bills the part above 10%, which is 339.12 of 1,186.92 credits used.

AI line assumptions

AI functions bill input and output tokens. So, in this illustrative bill (Enterprise Edition, on demand, AWS US East), 450 million input tokens cost 540 AI Credits, and 60 million output tokens cost 360. That is 900 AI Credits at $2.00. The bill assumes no cached input tokens. This bill assumes the account sets CORTEX_ENABLED_CROSS_REGION to ANY_REGION, which Snowflake prices at $2.00 per AI Credit. Accounts in organizations created after 9 March 2026 default to ANY_REGION. Other AWS accounts in the United States or Canada default to AWS_US, a regional setting at $2.20. At that price the same tokens cost $1,980.00, and Bill 3 costs $33,381.36. Snowflake lists claude-sonnet-5 under cross-region inference, not as native to AWS US East. The warehouse that runs the query bills separately, and Snowflake recommends no larger than Medium for AI functions.

How do the three bills compare?

In this illustration, the three bills total $750.70, $7,452.70 and $33,201.36 a month on Enterprise Edition, on demand in AWS US East. A BI warehouse is the largest line in each, and warehouses make up 93.9%, 87.6% and 76.6% of the total. That share falls here because the assumed storage, cloud services, serverless and AI lines grow faster.

All figures are illustrative: Enterprise Edition, on demand, AWS US East (N. Virginia), Gen2 warehouses, $3.00 per credit, and global routing for Bill 3’s AI line. They exclude data transfer. Bill 2 costs 9.9 times Bill 1, and Bill 3 costs 4.5 times Bill 2.

Which line item deserves a second look in each bill?

In these illustrative bills, the line to check first in Bill 1 is the idle tail on the loading warehouse. In Bill 2 it is the serverless lines, which resource monitors cannot cap. In Bill 3 it is cloud services, which are free below the 10% line, and AI, which has its own price per credit.

Bill 1: the idle tail

In illustrative Bill 1 (Enterprise Edition, on demand, AWS US East), 60 of the 210 billed seconds in each loading cycle are idle, which is 29%. The default AUTO_SUSPEND is 600 seconds, so a new warehouse left on the default would bill 750 seconds per cycle. As a result, the loading line rises from $170.10 to $607.50, an extra $437.40, and Bill 1 rises from $750.70 to $1,188.10.

Bill 2: the serverless lines

In illustrative Bill 2 (Enterprise Edition, on demand, AWS US East), Snowpipe and Optima Clustering total 39.7 credits, or $119.10. They scale with gigabytes, not hours, and they run on no warehouse you manage. Snowflake says resource monitors work for warehouses only, and that they cannot control Snowpipe. A budget can track these lines instead.

Bill 3: cloud services and AI

In illustrative Bill 3 (Enterprise Edition, on demand, AWS US East), cloud services bill 339.12 credits, or $1,017.36, because use passes the 10% line every day in this example. The rule is one sentence: each day, Snowflake adjusts away the lesser of cloud services used and 10% of that day’s warehouse credits. Our guide to Snowflake cloud services credits covers the causes. AI adds $1,800.00 at $2.00 per AI Credit with global routing, not $3.00, and the AI batch warehouse adds $1,458.00 on top.

What changes a Snowflake bill the most?

Usage sets how many credits you use, and price sets what each credit costs. In illustrative Bill 2 (Enterprise Edition, on demand, AWS US East), one more hour a day on the BI warehouse adds $356.40 a month. Moving to Business Critical adds $2,215.90, and AWS EU (Frankfurt) adds $2,046.81, because each reprices every credit and Frankfurt also reprices storage.

Usage levers

• Hours and idle time. One more hour each working day on Bill 2’s BI warehouse adds 118.8 credits, or $356.40 a month.
• Size. Each size up doubles credits per hour. Bill 2’s transformation job on Medium instead of Large, for the same hours, would use 405 credits ($1,215.00) instead of 810 ($2,430.00). However, each size up approximately doubles the computing power, per Snowflake, so a smaller size can take longer and the real saving depends on run time. Snowflake also notes that larger is not necessarily faster for small, basic queries, so test your own job.
• Generation. The same hours on Gen1 would use 1,612 credits instead of 2,176.2, so Gen2 adds 564.2 credits ($1,692.60). Snowflake says most queries should finish faster on Gen2, so test your own workload.

Price levers

• Edition. Bill 2’s 2,215.9 credits cost $4,431.80 on Standard, $6,647.70 on Enterprise and $8,863.60 on Business Critical, while storage stays $805.00. Multi-cluster warehouses and Query Acceleration need Enterprise or higher. See how much a Snowflake credit costs.
• Region. Bill 2 in AWS EU (Frankfurt) costs $9,499.51, which is $2,046.81 more, because a credit costs $3.90 and storage $24.50 per TB there.
• Capacity or on demand. A capacity contract applies the discount in your order form to the Platform Credit price, so $3.00 is only the on-demand list price. Each $0.10 off the credit price (an assumed discount) lowers Bill 2 by $221.59 and Bill 3 by $908.71 a month. The Platform Credit discount does not apply to AI Credits, which follow their own price by annual contract value, from $2.00 down to $1.88 with global routing. Check your order form for your own tier. Our guide to Snowflake on-demand vs capacity pricing covers the contract terms, and our 90-day renewal plan covers the timing.

How do you check these line items against your own account?

Query credits billed by service type from the METERING_DAILY_HISTORY view, then price each row yourself. Warehouse rows are WAREHOUSE_METERING, Snowpipe is PIPE, clustering is AUTO_CLUSTERING, and Query Acceleration is QUERY_ACCELERATION. AI usage is billed in AI Credits, so keep AI rows apart from warehouse rows when you price them.

-- Credits billed by service type, last 30 days (data can be up to 3 hours behind)
SELECT service_type,
       SUM(credits_used_compute) AS compute_credits,
       SUM(credits_used_cloud_services) AS cloud_services_credits,
       SUM(credits_adjustment_cloud_services) AS cloud_services_adjustment,
       SUM(credits_billed) AS total_credits_billed
FROM snowflake.account_usage.metering_daily_history
WHERE usage_date >= DATEADD(day, -30, CURRENT_DATE())
GROUP BY service_type
ORDER BY total_credits_billed DESC;

Run it as ACCOUNTADMIN, or as a role granted the USAGE_VIEWER database role. Storage bills per TB, not per credit, and the STORAGE_USAGE view shows average daily bytes.

How Seemore helps with Snowflake bill forecasting

Example bills show where credits go, but your own bill depends on your own hours, sizes and contract. Seemore’s Contract Utilization shows how much of your Snowflake contract you have used, where your current usage is projecting you to land, and whether that means overage, with the amount you would need to save to stay within it. Auto-Shutdown suspends idle warehouses, and SmartPulse works out the right size for each warehouse hour by hour from its real usage, so idle time and sizing do not depend on a monthly review. Cost Anomaly Detection compares spend across your accounts, warehouses and services, including AI services, against the last 14 days of your own history, and hands you the root cause when something drifts.