The rapid growth of the internet created a pressing demand for better data management and analysis technologies. In 2006, Amazon introduced S3 for cloud file storage, but it lacked standard data management capabilities. Apache’s Hadoop emerged as the de facto big data technology, but it was challenging to use and lacked a SQL interface until Hive’s release in 2010. However, Hive’s integration with Hadoop remained incomplete. In 2012, three experts (Benoit Dashville, Thierry Cruanes and Marcin Zukowski) developed Snowflake, a revolutionary cloud-based relational database management system (RDBMS), by rearchitecting the RDBMS for speed, scalability, ease of use and flexibility in data sharing. Distinct from traditional RDBMSs, on-premise MPP solutions and early cloud competitors such as Google BigQuery, AWS RedShift and Databricks, Snowflake transformed data management by separating storage and compute resources. This approach enables elastic scaling and pay-as-you-go pricing while overcoming performance limitations faced by conventional data warehouses with coupled storage and compute. Users must, however, actively manage Snowflake’s consumption-based model to maximize resource optimization and prevent excessive costs.
In this article, we’ll cover everything you need to know about Snowflake’s pricing strategy. From understanding the consumption-based usage model to breaking down the cost components, you’ll learn how Snowflake’s pay-per-usage strategy can help you save money and improve your cloud data warehouse spending.
How much does Snowflake cost?
Snowflake doesn’t charge a flat monthly fee. You pay for three separate meters:
Compute (virtual warehouses): Charged in Snowflake credits, billed per second with a 60-second minimum each time a warehouse starts or resumes.
Storage: Charged by the month per compressed terabyte. The rate depends on your cloud provider and region.
Cloud services layer: Metadata, security, and query planning services are included free up to 10% of your daily compute credits. Usage above that threshold is billed at the same credit rate as compute.
Data ingress (loading data into Snowflake) is free. Data egress (moving data out of Snowflake across regions or to a different cloud provider) is charged per gigabyte or terabyte, depending on the destination.
The dollar cost per credit depends on your edition, region, and cloud provider. We’ll break all of that down in the sections below. But first, let’s look at what editions are available and what each one includes.
You can pay for Snowflake usage month-to-month with Snowflake On-Demand, or commit to a pre-purchased capacity (for both compute credits and storage) in advance through Snowflake’s capacity contract model, which delivers meaningful discounts.
Snowflake Standard Edition
Standard offerings of Snowflake provide:
- Complete SQL data warehouse with support for large-scale storage, querying, and analysis
- Secure Data Sharing across regions and clouds, so you can share data with partners, customers, and other stakeholders
- Time Travel up to 1 day, covering basic data recovery and historical analysis
- Fail-safe data recovery for 7 days beyond the Time Travel window (for permanent tables)
- Premier Support, 24×365
- Access to the Snowflake Community
- Always-on enterprise-grade encryption in transit and at rest
- Customer-dedicated virtual warehouses
- Federated authentication (SSO and other encrypted authentications)
- SOC 2 Type II certification
- OAuth for account access authorization without sharing credentials
- Multi-factor authentication (MFA) support
- Object-level access control
- Database replication for high availability and disaster recovery across regions and clouds
- External functions for running custom code directly in the data warehouse
- Snowsight, Snowflake’s web-based data visualization and exploration interface
- Data Exchange for building a marketplace for data discovery and sharing
- Stored procedures with support for Java, JavaScript, Python, Scala, and SQL
- Snowflake External tables for referencing data in a cloud storage data lake
- Access to the Snowflake Marketplace, a curated data catalog from public and premium providers
- Streamlit in Snowflake for building, deploying, and sharing Streamlit apps
- Native JDBC and ODBC driver support
- Snowflake SQL API, a REST API for accessing and updating Snowflake data
Snowflake Enterprise Edition
Snowflake’s Enterprise offerings include all of the features and benefits of the standard offering, as well as the following advanced capabilities:
- Multi-cluster warehouses for improved concurrency and performance at scale
- Extended Time Travel up to 90 days for deeper data recovery and historical analysis
- Annual rekeying of all encrypted data
- Column-level security via masking policies on individual columns
- Row-level security via row access policies controlling which rows appear in query results
- Aggregation policies to enforce privacy by requiring data aggregation before results are returned
- Projection policies restricting SELECT statements on specific columns
- Object Tagging for tracking sensitive data and resource utilization
- User access history auditing via the Account Usage ACCESS_HISTORY view
- Materialized views for pre-computed query results, reducing repeated compute workload
- Search optimization service for efficient indexing and querying of large datasets
- Dynamic data masking for sensitive data protection
- External data tokenization for testing, reporting, and analytics without exposing source data
Snowflake Business Critical
Snowflake’s Business Critical offerings include all of the features and benefits of the Snowflake enterprise offering, as well as the following advanced capabilities:
- HIPAA compliance for protected health information (PHI)
- PCI DSS compliance for payment card data security
- Tri-Secret Secure, which uses customer-managed encryption keys for maximum data control
- Support for AWS Private Link, Azure Private Link, and Google Cloud Private Service Connect
- Amazon API Gateway private endpoints for external functions
- Database failover and failback for business continuity
- External functions with AWS API Gateway private endpoint support
- Support for U.S. Federal and state government requirements, including FedRAMP and ITAR
- Support for IRAP-Protected (P) data in specified Asia Pacific regions
Snowflake Virtual Private Snowflake (VPS)
Virtual Private Snowflake (VPS) offerings of Snowflake provide all of the features and benefits of the Business Critical offering, as well as the following advanced capabilities:
- Customer-dedicated virtual servers where the encryption key stays in memory only
- A customer-dedicated metadata store for strict compliance and data isolation
- 24-hour early access to weekly Snowflake releases for pre-production testing and validation
Before we dive into Snowflake pricing for each Snowflake services, let’s first understand what Snowflake credits are and the factors that affect them.
What are Snowflake credits?
A Snowflake credit is the unit of measure for compute consumption. Every time a virtual warehouse runs a query, every time a serverless feature executes, that activity is measured in credits. The dollar cost of a credit depends on your edition, region, and cloud provider.
Three factors determine your credit cost:
- Edition: Standard, Enterprise, Business Critical, or VPS. As a rough on-demand baseline for AWS US East: Standard runs approximately $2/credit, Enterprise ~$3/credit, and Business Critical ~$4/credit. VPS is custom-priced. Non-US regions typically carry a 10 – 50% premium over US baseline rates.
- Region: US regions represent the cheapest baseline. Non-US regions carry a premium that varies by location and cloud provider.
- Cloud provider: AWS, Azure, or Google Cloud. Rates are broadly comparable across providers within the same region, though minor differences exist.
Snowflake bills credits per second, with a 60-second minimum each time a warehouse starts or resumes. A 5-second query costs you 60 seconds of compute. If a warehouse resumes for a 3-second query, suspends, then resumes again 2 minutes later for another 3-second query, you’ve paid for 120 seconds total. That 60-second minimum is why auto-suspend configuration matters as much as warehouse sizing.
Snowflake pricing for each Snowflake service
1) Snowflake Storage cost
Storage is billed at a flat monthly rate per terabyte of compressed data. Snowflake automatically compresses data on load, so your actual billed volume is typically much smaller than raw file size. A terabyte of files might compress to 200-300 GB after loading. You pay for the daily average of compressed TB per month, and the rate depends on cloud provider and region.
Standard storage costs on AWS
| Region | On-Demand (TB/mo) | Tier 1 | Tier 2 | Tier 3 | Tier 4 | Tier 5 | Tier 6 | Tier 7 |
| US East (N. Virginia) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| US West (Oregon) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| EU Dublin | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| EU Frankfurt | $24.50 | $24.50 | $22.87 | $21.24 | $19.60 | $17.96 | $16.34 | $14.70 |
| AP Sydney | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| AP Singapore | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| Canada Central | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| US East 2 (Ohio) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| AP Northeast 1 (Tokyo) | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| AP Mumbai | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| US East 1 Commercial Gov | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| Europe (London) | $24.00 | $24.00 | $22.40 | $20.81 | $19.20 | $17.59 | $16.01 | $14.40 |
| Asia Pacific (Seoul) | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| US Gov West 1 | $39.00 | $39.00 | $36.40 | $33.81 | $31.20 | $28.59 | $26.01 | $23.40 |
| US Gov West 1 (FedRAMP High Plus) | $39.00 | $39.00 | $36.40 | $33.81 | $31.20 | $28.59 | $26.01 | $23.40 |
| Europe (Stockholm) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| Asia Pacific (Osaka) | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| South America East 1 (Sao Paulo) | $40.50 | $40.50 | $37.80 | $35.11 | $32.40 | $29.69 | $27.01 | $24.30 |
| EU (Paris) | $24.00 | $24.00 | $22.40 | $20.81 | $19.20 | $17.59 | $16.01 | $14.40 |
| Asia Pacific (Jakarta) | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| US Gov East 1 (FedRAMP High Plus) | $39.00 | $39.00 | $36.40 | $33.81 | $31.20 | $28.59 | $26.01 | $23.40 |
| EU (Zurich) | $26.95 | $26.95 | $25.15 | $23.37 | $21.56 | $19.75 | $17.98 | $16.17 |
| US Gov West 1 (DoD) | $39.00 | $39.00 | $36.40 | $33.81 | $31.20 | $28.59 | $26.01 | $23.40 |
| US West (Commercial Gov – Oregon) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| Africa (Cape Town) | $27.40 | $27.40 | $25.57 | $23.76 | $21.92 | $20.08 | $18.28 | $16.44 |
| Middle East (UAE) | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| Asia Pacific (Malaysia) | $22.50 | $22.50 | $21.00 | $19.51 | $18.00 | $16.49 | $15.01 | $13.50 |
| Asia Pacific (Thailand) | $22.50 | $22.50 | $21.00 | $19.51 | $18.00 | $16.49 | $15.01 | $13.50 |
| Asia Pacific (New Zealand) | $26.25 | $26.25 | $24.50 | $22.76 | $21.00 | $19.24 | $17.51 | $15.75 |
Capacity ACV tiers:
- Tier 1 = $0-$1,199,999
- Tier 2 = $1,200,000-$2,999,999
- Tier 3 = $3,000,000-$4,999,999
- Tier 4 = $5,000,000-$9,999,999
- Tier 5 = $10,000,000-$19,999,999
- Tier 6 = $20,000,000-$39,999,999
- Tier 7 = $40,000,000+
Standard storage costs on Azure
| Region | On-Demand (TB/mo) | Tier 1 | Tier 2 | Tier 3 | Tier 4 | Tier 5 | Tier 6 | Tier 7 |
| East US 2 (Virginia) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| West US 2 (Washington) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| West Europe (Netherlands) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| Australia East (New South Wales) | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| Canada Central | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| Southeast Asia (Singapore) | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| Switzerland North | $28.80 | $28.80 | $26.88 | $24.97 | $23.04 | $21.11 | $19.21 | $17.28 |
| US Gov Virginia | $39.00 | $39.00 | $36.40 | $33.81 | $31.20 | $28.59 | $26.01 | $23.40 |
| US Central (Iowa) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| North Europe (Ireland) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| Japan East (Tokyo) | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| UAE North (Dubai) | $25.40 | $25.40 | $23.71 | $22.02 | $20.32 | $18.62 | $16.94 | $15.24 |
| South Central US (Texas) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| Central India (Pune) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| UK South (London) | $24.00 | $24.00 | $22.40 | $20.81 | $19.20 | $17.59 | $16.01 | $14.40 |
| US Gov Virginia (FedRAMP High Plus) | $39.00 | $39.00 | $36.40 | $33.81 | $31.20 | $28.59 | $26.01 | $23.40 |
| Mexico Central | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| Korea Central | $25.00 | $25.00 | $23.33 | $21.68 | $20.00 | $18.33 | $16.68 | $15.00 |
| Sweden Central | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
| East US (Virginia) | $23.00 | $23.00 | $21.47 | $19.94 | $18.40 | $16.86 | $15.34 | $13.80 |
Standard storage costs on GCP
| Region | On-Demand (TB/mo) | Tier 1-7 |
| US Central 1 (Iowa) | $20.00 | $20.00 (flat) |
| US East 4 (N. Virginia) | $23.00 | $23.00 (flat) |
| Europe West 4 (Netherlands) | $20.00 | $20.00 (flat) |
| Europe West 2 (London) | $23.00 | $23.00 (flat) |
| Europe West 3 (Frankfurt) | $23.00 | $23.00 (flat) |
| Middle East Central 2 (Dammam) | $30.00 | $30.00 (flat) |
| Australia Southeast 2 (Melbourne) | $23.00 | $23.00 (flat) |
Note: GCP storage pricing is flat across all capacity ACV tiers — volume commitments don’t reduce storage rates on GCP.
Hybrid Tables storage costs on AWS
Hybrid Tables use a row-oriented storage format optimized for transactional workloads and are priced separately from standard columnar storage.
| Region | On-Demand & Capacity Hybrid Tables Storage (GB/mo) |
| US East (N. Virginia) | $0.34 |
| US West (Oregon) | $0.34 |
| EU Dublin | $0.34 |
| EU Frankfurt | $0.36 |
| AP Sydney | $0.37 |
| AP Singapore | $0.37 |
| Canada Central | $0.37 |
| US East 2 (Ohio) | $0.34 |
| AP Northeast 1 (Tokyo) | $0.37 |
| AP Mumbai | $0.34 |
| US East 1 Commercial Gov | $0.34 |
| Europe (London) | $0.35 |
| Asia Pacific (Seoul) | $0.37 |
| US Gov West 1 | $0.58 |
| US Gov West 1 (FedRAMP High Plus) | $0.58 |
| Europe (Stockholm) | $0.34 |
| Asia Pacific (Osaka) | $0.37 |
| South America East 1 (Sao Paulo) | $0.60 |
| EU (Paris) | $0.35 |
| Asia Pacific (Jakarta) | $0.37 |
| US Gov East 1 (FedRAMP High Plus) | $0.58 |
| EU (Zurich) | $0.40 |
| US Gov West 1 (DoD) | $0.58 |
| US West (Commercial Gov – Oregon) | $0.34 |
| Africa (Cape Town) | $0.40 |
| Middle East (UAE) | $0.37 |
| Asia Pacific (Malaysia) | $0.37 |
| Asia Pacific (Thailand) | $0.37 |
| Asia Pacific (New Zealand) | $0.43 |
Hybrid Tables storage costs on Azure
| Region | On-Demand & Capacity Hybrid Tables Storage (GB/mo) |
| East US 2 (Virginia) | $0.34 |
| West US 2 (Washington) | $0.34 |
| West Europe (Netherlands) | $0.34 |
| Australia East (New South Wales) | $0.37 |
| Canada Central (Toronto) | $0.37 |
| Southeast Asia (Singapore) | $0.37 |
| Switzerland North | $0.43 |
| Central US (Iowa) | $0.34 |
| North Europe (Ireland) | $0.34 |
| Japan East (Tokyo) | $0.37 |
| UAE North (Dubai) | $0.38 |
| South Central US (Texas) | $0.34 |
| Central India (Pune) | $0.37 |
| UK South (London) | $0.35 |
| Mexico Central | $0.34 |
| Korea Central | $0.37 |
| Sweden Central | $0.34 |
| East US (Virginia) | $0.34 |
SPCS block storage pricing on AWS
| Region | Instance Type | Block Storage Volume (TB/mo) | Block Storage IOPS (1,000 IOPS-mo) | Block Storage Throughput (GB/sec-mo) | Block Storage Snapshot (TB/mo) |
| US East (N. Virginia) | CPU/GPU | $81.92 | $5.00 | $40.96 | $51.20 |
| US West (Oregon) | CPU/GPU | $81.92 | $5.00 | $40.96 | $51.20 |
| EU Dublin | CPU/GPU | $90.12 | $5.50 | $45.06 | $51.20 |
| EU Frankfurt | CPU/GPU | $97.49 | $6.00 | $48.75 | $55.30 |
| AP Sydney | CPU/GPU | $98.31 | $6.00 | $49.16 | $56.30 |
| AP Singapore | CPU/GPU | $98.31 | $6.00 | $49.16 | $51.20 |
| Canada Central | CPU/GPU | $90.12 | $5.50 | $45.06 | $56.30 |
| US East 2 (Ohio) | CPU/GPU | $81.92 | $5.00 | $40.96 | $51.20 |
| AP Northeast 1 (Tokyo) | CPU/GPU | $98.31 | $6.00 | $49.16 | $51.20 |
| AP Mumbai | CPU/GPU | $81.92 | $5.00 | $40.96 | $51.20 |
| US East 1 Commercial Gov | CPU/GPU | $81.92 | $5.00 | $40.96 | $51.20 |
| Europe (London) | CPU/GPU | $95.03 | $5.80 | $47.52 | $54.30 |
| Asia Pacific (Seoul) | CPU/GPU | $93.39 | $5.70 | $46.70 | $51.20 |
| US Gov West 1 | CPU/GPU | $98.31 | $6.00 | $49.16 | $67.60 |
| US Gov West 1 (FedRAMP High Plus) | CPU/GPU | $98.31 | $6.00 | $49.16 | $67.60 |
| Europe (Stockholm) | CPU/GPU | $85.61 | $5.20 | $42.81 | $48.60 |
| Asia Pacific (Osaka) | CPU/GPU | $98.31 | $6.00 | $49.16 | $51.20 |
| South America East 1 (Sao Paulo) | CPU/GPU | $155.65 | $9.50 | $77.83 | $69.60 |
| EU (Paris) | CPU/GPU | $95.03 | $5.80 | $47.52 | $54.30 |
| Asia Pacific (Jakarta) | CPU/GPU | $98.31 | $6.00 | $49.16 | $51.20 |
| US Gov East 1 (FedRAMP High Plus) | CPU/GPU | $98.31 | $6.00 | $49.16 | $67.60 |
| EU (Zurich) | CPU/GPU | $116.95 | $7.00 | $58.48 | $60.40 |
| US Gov West 1 (DoD) | CPU/GPU | $98.31 | $6.00 | $49.16 | $67.60 |
| US West (Commercial Gov – Oregon) | CPU/GPU | $81.92 | $5.00 | $40.96 | $51.20 |
| Africa (Cape Town) | CPU/GPU | $107.21 | $6.50 | $53.25 | $60.93 |
| Middle East (UAE) | CPU/GPU | $99.12 | $6.10 | $49.15 | $56.32 |
| Asia Pacific (Malaysia) | CPU/GPU | $88.47 | $5.40 | $44.24 | $46.08 |
| Asia Pacific (Thailand) | CPU/GPU | $88.47 | $5.40 | $44.24 | $46.08 |
| Asia Pacific (New Zealand) | CPU/GPU | $103.22 | $6.30 | $51.61 | $59.08 |
SPCS block storage pricing on Azure
| Region | Instance Type | Block Storage Volume (TB/mo) | Block Storage IOPS (1,000 IOPS-mo) | Block Storage Throughput (GB/sec-mo) | Block Storage Snapshot (TB/mo) |
| East US 2 (Virginia) | CPU/GPU | $82.23 | $5.11 | $41.12 | $51.20 |
| West US 2 (Washington) | CPU/GPU | $82.20 | $5.11 | $41.12 | $51.20 |
| West Europe (Netherlands) | CPU/GPU | $97.18 | $5.84 | $48.59 | $51.20 |
| Australia East (New South Wales) | CPU/GPU | $82.23 | $5.11 | $41.12 | $56.30 |
| Canada Central (Toronto) | CPU/GPU | $90.40 | $5.84 | $44.86 | $56.30 |
| Southeast Asia (Singapore) | CPU/GPU | $98.68 | $5.84 | $49.34 | $51.20 |
| Switzerland North | CPU/GPU | $117.37 | $7.30 | $58.31 | $56.30 |
| US Gov Virginia | CPU/GPU | $98.68 | $5.84 | $49.34 | $107.00 |
| Central US (Iowa) | CPU/GPU | $92.70 | $5.84 | $46.35 | $51.20 |
| North Europe (Ireland) | CPU/GPU | $82.23 | $5.11 | $41.12 | $51.20 |
| Japan East (Tokyo) | CPU/GPU | $98.68 | $5.84 | $49.34 | $51.20 |
| UAE North (Dubai) | CPU/GPU | $99.42 | $5.84 | $49.34 | $61.40 |
| South Central US (Texas) | CPU/GPU | $82.23 | $5.11 | $41.12 | $51.20 |
| Central India (Pune) | CPU/GPU | $82.23 | $5.11 | $41.12 | $51.20 |
| UK South (London) | CPU/GPU | $94.94 | $5.84 | $47.10 | $54.30 |
| US Gov Virginia (FedRAMP High Plus) | CPU/GPU | $98.68 | $5.84 | $49.34 | $107.00 |
| Mexico Central | CPU/GPU | $90.11 | $7.30 | $45.06 | $56.32 |
| Korea Central | CPU/GPU | $93.39 | $5.70 | $47.10 | $51.20 |
| Sweden Central | CPU/GPU | $82.94 | $5.20 | $41.98 | $51.20 |
| East US (Virginia) | CPU/GPU | $82.94 | $5.20 | $41.98 | $51.20 |
SPCS block storage pricing on GCP
| Region | Instance Type | Block Storage Volume (TB/mo) | Block Storage IOPS (1,000 IOPS-mo) | Block Storage Throughput (GB/sec-mo) | Block Storage Snapshot (TB/mo) |
| US Central 1 (Iowa) | CPU | $81.92 | $5.00 | $40.96 | $51.20 |
| US Central 1 (Iowa) | GPU | $81.92 | — | $122.88 | $51.20 |
| US East 4 (N. Virginia) | CPU | $81.92 | $5.00 | $40.96 | $51.20 |
| US East 4 (N. Virginia) | GPU | $81.92 | — | $122.88 | $51.20 |
| Europe West 4 (Netherlands) | CPU | $86.02 | $5.00 | $43.01 | $54.27 |
| Europe West 4 (Netherlands) | GPU | $86.02 | — | $129.02 | $54.27 |
| Europe West 2 (London) | CPU | $93.18 | $6.00 | $47.10 | $58.37 |
| Europe West 2 (London) | GPU | $93.18 | — | $140.29 | $58.37 |
| Europe West 3 (Frankfurt) | CPU | $96.26 | $6.00 | $48.13 | $60.42 |
| Europe West 3 (Frankfurt) | GPU | $96.26 | — | $145.41 | $60.42 |
| Middle East Central 2 (Dammam) | CPU | $131.07 | $8.00 | $65.54 | $81.92 |
| Middle East Central 2 (Dammam) | GPU | $131.07 | — | $196.61 | $81.92 |
| Australia Southeast 2 (Melbourne) | CPU | $106.50 | $7.00 | $53.23 | $66.56 |
| Australia Southeast 2 (Melbourne) | GPU | $106.50 | — | $159.74 | $66.56 |
Archive storage and data retrieval pricing on AWS
| Region | Cool Tier Storage (TB/mo) | Cool Tier Retrieval (per TB) | Cold Tier Storage (TB/mo) | Cold Tier Retrieval (per TB) |
| US East (N. Virginia) | $4.00 | $30.00 | $1.00 | $2.50 |
| US West (Oregon) | $4.00 | $30.00 | $1.00 | $2.50 |
| EU Dublin | $4.00 | $30.00 | $1.00 | $3.00 |
| EU Frankfurt | $5.00 | $30.00 | $1.00 | $5.00 |
| AP Sydney | $5.00 | $30.00 | $1.00 | $5.00 |
| AP Singapore | $5.00 | $30.00 | $1.00 | $5.00 |
| Canada Central | $5.00 | $30.00 | $1.00 | $5.00 |
| US East 2 (Ohio) | $4.00 | $30.00 | $1.00 | $2.50 |
| AP Northeast 1 (Tokyo) | $5.00 | $30.00 | $1.00 | $5.00 |
| AP Mumbai | $5.00 | $30.00 | $1.00 | $5.00 |
| US East 1 Commercial Gov | $4.00 | $30.00 | $1.00 | $2.50 |
| Europe (London) | $5.00 | $30.00 | $1.00 | $5.00 |
| Asia Pacific (Seoul) | $5.00 | $30.00 | $1.00 | $5.00 |
| US Gov West 1 | $6.40 | $30.00 | $1.20 | $3.40 |
| US Gov West 1 (FedRAMP High Plus) | $6.40 | $30.00 | $1.20 | $3.40 |
| Europe (Stockholm) | $4.00 | $30.00 | $1.00 | $3.00 |
| Asia Pacific (Osaka) | $5.00 | $30.00 | $1.00 | $5.00 |
| South America East 1 (Sao Paulo) | $8.30 | $30.00 | $1.40 | $8.00 |
| EU (Paris) | $5.00 | $30.00 | $1.00 | $5.00 |
| Asia Pacific (Jakarta) | $5.00 | $30.00 | $1.00 | $5.00 |
| US Gov East 1 (FedRAMP High Plus) | $6.40 | $30.00 | $1.20 | $3.40 |
| EU (Zurich) | $5.50 | $30.00 | $1.00 | $5.00 |
| US Gov West 1 (DoD) | $6.40 | $30.00 | $1.20 | $3.40 |
| US West (Commercial Gov – Oregon) | $4.00 | $30.00 | $1.00 | $2.50 |
| Africa (Cape Town) | $5.00 | $30.00 | $1.00 | $5.00 |
| Middle East (UAE) | $5.00 | $30.00 | $1.80 | $3.30 |
| Asia Pacific (Malaysia) | $4.50 | $30.00 | $1.80 | $4.50 |
| Asia Pacific (Thailand) | $4.50 | $30.00 | $1.80 | $4.50 |
| Asia Pacific (New Zealand) | $5.25 | $30.00 | $2.10 | $5.25 |
Archive storage and data retrieval pricing on Azure
Azure currently only offers a Cool Tier for archive storage. Cold Tier is not available on Azure.
| Region | Cool Tier Storage (TB/mo) | Cool Tier Retrieval (per TB) |
| East US 2 (Virginia) | $4.00 | $30.00 |
| West US 2 (Washington) | $4.00 | $30.00 |
| West Europe (Netherlands) | $5.00 | $30.00 |
| Australia East (New South Wales) | $5.00 | $30.00 |
| Canada Central (Toronto) | $5.00 | $30.00 |
| Southeast Asia (Singapore) | $5.00 | $30.00 |
| Switzerland North | $5.71 | $42.90 |
| US Gov Virginia | $6.40 | $30.00 |
| US Central (Iowa) | $4.92 | $36.90 |
| North Europe (Ireland) | $4.00 | $30.00 |
| Japan East (Tokyo) | $5.00 | $30.00 |
| UAE North (Dubai) | $5.00 | $30.00 |
| South Central US (Texas) | $4.80 | $36.00 |
| Central India (Pune) | $5.00 | $30.00 |
| UK South (London) | $5.00 | $30.00 |
| US Gov Virginia (FedRAMP High Plus) | $6.40 | $30.00 |
| Mexico Central | $4.95 | $33.00 |
| Korea Central | $5.00 | $30.00 |
| Sweden Central | $4.00 | $30.00 |
| East US (Virginia) | $4.00 | $30.00 |
Archive storage and data retrieval pricing on GCP
| Region | Cool Tier Storage (TB/mo) | Cool Tier Retrieval (per TB) | Cold Tier Storage (TB/mo) | Cold Tier Retrieval (per TB) |
| US Central 1 (Iowa) | $4.00 | $20.00 | $1.20 | $50.00 |
| US East 4 (N. Virginia) | $6.00 | $20.00 | $2.50 | $50.00 |
| Europe West 2 (London) | $7.00 | $20.00 | $2.50 | $50.00 |
| Europe West 3 (Frankfurt) | $6.00 | $20.00 | $2.50 | $50.00 |
| Europe West 4 (Netherlands) | $4.00 | $20.00 | $1.20 | $50.00 |
| Middle East Central 2 (Dammam) | $6.00 | $20.00 | $2.70 | $50.00 |
| Australia Southeast 2 (Melbourne) | $6.00 | $20.00 | $2.50 | $50.00 |
Snowflake Postgres storage on AWS
| Region | Postgres Storage (TB/mo) | Postgres Storage – High Availability (TB/mo) |
| US East (N. Virginia) | $117.76 | $235.52 |
| US West (Oregon) | $117.76 | $235.52 |
| EU Dublin | $129.55 | $259.10 |
| EU Frankfurt | $140.15 | $280.30 |
| AP Sydney | $141.32 | $282.64 |
| AP Singapore | $141.32 | $282.64 |
| Canada Central | $129.55 | $259.10 |
| US East 2 (Ohio) | $117.76 | $235.52 |
| AP Northeast 1 (Tokyo) | $141.32 | $282.64 |
| AP Mumbai | $117.76 | $235.52 |
| US East 1 Commercial Gov | $117.76 | $235.52 |
| Europe (London) | $136.60 | $273.20 |
| Asia Pacific (Seoul) | $134.25 | $268.50 |
| US Gov West 1 | $141.32 | $282.64 |
| US Gov West 1 (FedRAMP High Plus) | $141.32 | $282.64 |
| Europe (Stockholm) | $123.06 | $246.12 |
| Asia Pacific (Osaka) | $141.32 | $282.64 |
| South America East 1 (Sao Paulo) | $223.74 | $447.48 |
| EU (Paris) | $136.60 | $273.20 |
| Asia Pacific (Jakarta) | $141.32 | $282.64 |
| US Gov East 1 (FedRAMP High Plus) | $141.32 | $282.64 |
| EU (Zurich) | $168.11 | $336.22 |
| US Gov West 1 (DoD) | $141.32 | $282.64 |
| US West (Commercial Gov – Oregon) | $117.76 | $235.52 |
| Africa (Cape Town) | $154.11 | $308.22 |
| Middle East (UAE) | $142.49 | $284.98 |
| Asia Pacific (Malaysia) | $127.18 | $254.36 |
| Asia Pacific (Thailand) | $127.18 | $254.36 |
| Asia Pacific (New Zealand) | $148.38 | $296.76 |
Snowflake Postgres storage on Azure
| Region | Postgres Storage (TB/mo) | Postgres Storage – High Availability (TB/mo) |
| East US 2 (Virginia) | $118.21 | $236.42 |
| West US 2 (Washington) | $118.16 | $236.32 |
| West Europe (Netherlands) | $139.70 | $279.40 |
| Australia East (New South Wales) | $118.21 | $236.42 |
| Canada Central (Toronto) | $129.95 | $259.90 |
| Southeast Asia (Singapore) | $141.85 | $283.70 |
| Switzerland North | $168.71 | $337.42 |
| US Gov Virginia | $141.85 | $283.70 |
| Central US (Iowa) | $133.26 | $266.52 |
| North Europe (Ireland) | $118.21 | $236.42 |
| Japan East (Tokyo) | $141.85 | $283.70 |
| UAE North (Dubai) | $142.91 | $285.82 |
| South Central US (Texas) | $118.21 | $236.42 |
| Central India (Pune) | $118.21 | $236.42 |
| UK South (London) | $136.47 | $272.94 |
| US Gov Virginia (FedRAMP High Plus) | $141.85 | $283.70 |
| Mexico Central | $129.54 | $259.08 |
| Sweden Central | $118.78 | $237.56 |
| Korea Central | $134.25 | $268.50 |
| East US (Virginia) | $119.23 | $238.46 |
Check out the article below to learn more in-depth on Snowflake’s storage costs and how they are calculated:
Snowflake Storage Costs 101: An In-Depth Guide
2) Snowflake Virtual Warehouse cost
Snowflake virtual warehouses range from X-Small to 6X-Large. Each step up doubles both the compute resources and the credit consumption rate. That makes it straightforward to estimate: a Medium warehouse (4 credits/hour) costs exactly twice as much per hour as a Small (2 credits/hour), and exactly half as much as a Large (8 credits/hour).
Standard warehouse credits per hour
| XS | S | M | L | XL | 2XL | 3XL | 4XL | 5XL | 6XL |
| 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128 | 256 | 512 |
A 6X-Large warehouse at 512 credits/hour is 512 times more expensive per hour than an X-Small at 1 credit/hour. Whether you get 512x the performance is a different question — for most analytics workloads, you won’t. Query performance doesn’t scale linearly with warehouse size because query complexity, data distribution, and I/O bottlenecks all introduce ceilings. Bigger isn’t automatically better. Test your workloads at smaller sizes first.
Credit consumption broken down by time unit:
- X-small:
- 1 hour = 1 Credit
- 1 minute = 1/60 Credits ≈ 0.0167 Credits
- 1 second = 0.0167/60 Credits ≈ 0.000278 Credits
- Small:
- 1 hour = 2 Credits
- 1 minute = 2/60 Credits ≈ 0.0333 Credits
- 1 second = 0.0333/60 Credits ≈ 0.000555 Credits
- Medium:
- 1 hour = 4 Credits
- 1 minute = 4/60 Credits ≈ 0.0667 Credits
- 1 second = 0.0667/60 Credits ≈ 0.00111 Credits
- Large:
- 1 hour = 8 Credits
- 1 minute = 8/60 Credits ≈ 0.1333 Credits
- 1 second = 0.1333/60 Credits ≈ 0.00222 Credits
- X-Large:
- 1 hour = 16 Credits
- 1 minute = 16/60 Credits ≈ 0.2667 Credits
- 1 second = 0.2667/60 Credits ≈ 0.004445 Credits
- 2X-Large:
- 1 hour = 32 Credits
- 1 minute = 32/60 Credits ≈ 0.5333 Credits
- 1 second = 0.5333/60 Credits ≈ 0.008889 Credits
- 3X-Large:
- 1 hour = 64 Credits
- 1 minute = 64/60 Credits ≈ 1.0667 Credits
- 1 second = 1.0667/60 Credits ≈ 0.017778 Credits
- 4X-Large:
- 1 hour = 128 Credits
- 1 minute = 128/60 Credits ≈ 2.1333 Credits
- 1 second = 2.1333/60 Credits ≈ 0.035555 Credits
- 5X-Large:
- 1 hour = 256 Credits
- 1 minute = 256/60 Credits ≈ 4.2667 Credits
- 1 second = 4.2667/60 Credits ≈ 0.071111 Credits
- 6X-Large:
- 1 hour = 512 Credits
- 1 minute = 512/60 Credits ≈ 8.5333 Credits
- 1 second = 8.5333/60 Credits ≈ 0.142222 Credits
Credits are billed per second, with a 60-second minimum each time a warehouse starts or resumes. If a warehouse resumes for a 5-second query, you’re billed for 60 seconds. Auto-suspend configuration determines how long the warehouse stays alive after that query finishes. A 5-minute auto-suspend window means you’re paying for idle time every time a short query fires up the warehouse. Near-instant auto-suspend and auto-resume means you can pair tight suspend windows with responsive user-facing dashboards without worrying about cold-start delays.
Snowflake eliminates the data-deletion-on-suspension problem that plagues other cloud platforms. Traditional cloud data platforms typically delete compute-layer data when instances terminate, making restarts time-consuming and expensive if data needs to be reloaded. Snowflake’s architecture keeps data in persistent storage separate from compute, so a paused warehouse resumes instantly for a fresh query without any reprovisioning.
a) Snowflake Generation 2 Standard Warehouses costs
Snowflake has introduced Generation 2 (Gen2) standard warehouses. Gen2 warehouses use faster hardware (specifically AWS Graviton 3 ARM-based instances on AWS) and include intelligent software optimizations for analytics and data engineering workloads. Snowflake benchmarked Gen2 at 2.1x faster performance for core analytics on Snowflake tables as of May 2025, with even larger gains for DML operations (DELETE, UPDATE, MERGE) at up to 4.4x.
Gen2 became generally available across AWS, Azure, and GCP in November 2025. It supports sizes X-Small through 4X-Large only. 5X-Large and 6X-Large are not available on Gen2.
Limitations: Gen2 costs roughly 1.35x the credits per hour of Gen1 on AWS and GCP, and 1.25x on Azure. Whether you save money depends on how much faster Gen2 finishes your specific queries. If a query runs 35% faster on Gen2 but costs 35% more per second, you break even. You only save credits if the speed gain outpaces the rate increase. Short or light queries often show no savings on Gen2. Benchmark before switching.
Gen2 standard warehouse credits per hour by cloud
| Cloud | XS | S | M | L | XL | 2XL | 3XL | 4XL |
| AWS | 1.35 | 2.7 | 5.4 | 10.8 | 21.6 | 43.2 | 86.4 | 172.8 |
| Azure | 1.25 | 2.5 | 5.0 | 10.0 | 20.0 | 40.0 | 80.0 | 160.0 |
| GCP | 1.35 | 2.7 | 5.4 | 10.8 | 21.6 | 43.2 | 86.4 | 172.8 |
Gen2 is not available for 5X-Large or 6X-Large warehouses. Snowpark-optimized warehouses do not use Gen2.
b) Snowpark-optimized virtual warehouses
Snowpark-optimized warehouses provide 16x more memory per node compared to standard warehouses. They’re built for memory-intensive workloads: machine learning training, large Snowpark data science jobs, or any process that needs a large in-memory footprint on a single node.
Snowpark-optimized warehouses are generally available across AWS, Azure, and Google Cloud in all regions. The 1 TB memory option (MEMORY_64X resource constraints) is currently in preview on AWS only.
Snowpark-optimized warehouse credits per hour
| Resource Constraint | XS | S | M | L | XL | 2XL | 3XL | 4XL | 5XL | 6XL |
| MEMORY_1X | 1.00 | 2.00 | 4.00 | 8.00 | 16.00 | 32.00 | 64.00 | 128.00 | — | — |
| MEMORY_1X_x86 | 1.10 | 2.20 | 4.40 | 8.80 | 17.60 | 35.20 | 70.40 | 140.80 | — | — |
| MEMORY_16X | — | — | 6.00 | 12.00 | 24.00 | 48.00 | 96.00 | 192.00 | 384.00 | 768.00 |
| MEMORY_16X_x86 | — | — | 6.25 | 12.50 | 25.00 | 50.00 | 100.00 | 200.00 | — | — |
| MEMORY_64X (preview, AWS only) | — | — | — | 15.00 | 30.00 | 60.00 | 120.00 | 240.00 | — | — |
| MEMORY_64X_x86 (preview, AWS only) | — | — | — | 16.00 | 32.00 | 64.00 | 128.00 | 256.00 | — | — |
Check out the article below to learn more in-depth on Snowflake’s compute costs and how they are calculated:
Snowflake Compute Costs 101: An In-Depth Guide
c) Snowflake Serverless cost
Serverless compute in Snowflake covers features managed by Snowflake directly, rather than through user-controlled virtual warehouses. Snowflake automatically provisions and scales these resources based on workload requirements.
Serverless charges are measured in compute hours, calculated per second and rounded up to the nearest whole second. One compute hour is roughly equivalent to running an X-Small virtual warehouse for one hour. Each serverless feature has its own credit multiplier applied to that base rate.
This is where a lot of unexpected costs originate. Materialized views and Search Optimization run continuously once enabled. Automatic clustering keeps running on clustered tables regardless of whether your warehouses are active. These features don’t show up in warehouse dashboards; you need to query SNOWFLAKE.ACCOUNT_USAGE views to track them.
Serverless feature credit multipliers and unit charges
| Feature | Snowflake-managed compute multiplier | Cloud services multiplier | Unit charges |
| Archive Storage Retrieval (File Processing) | — | — | 0.05 credits per 1,000 files |
| Archive Storage Write | — | — | 0.05 credits per 1,000 files |
| Automated Refresh and Data Registration | 1.25 | — | 0.06 credits per 1,000 files |
| Backup | 2 | 1 | — |
| Clustered Tables | 2 | 1 | — |
| Copy Files | 2 | — | — |
| Data Quality Monitoring | 2 | 1 | — |
| Failsafe Recovery | 0.9 | 1 | — |
| Logging | 1.25 | — | 0.28 credits per 1,000 file batches |
| Materialized Views maintenance | 2 | 1 | — |
| Open Catalog | — | — | 0.5 credits per 1 million requests |
| Query Acceleration | 1 | — | — |
| Replication | 2 | 0.35 | — |
| Search Optimization Service | 2 | 1 | — |
| Sensitive Data Classification | 0.9 | 1 | — |
| Serverless Alerts | 0.9 | 1 | — |
| Serverless Tasks | 0.9 | 1 | — |
| Serverless Tasks Flex | 0.5 | 1 | — |
| Snowpipe | — | — | 0.0037 credits per GB |
| Snowpipe Streaming | — | — | 0.0037 credits per uncompressed GB |
| Snowpipe Streaming Classic | 1 | — | 0.01 credits per client instance per hour |
| Storage Lifecycle Policy Execution | 0.50 | 1 | — |
| Table Optimization | 0.75 | 1 | — |
| Telemetry Data Ingest | — | — | 0.0212 credits per GB |
| Trust Center | 1 | 1 | — |
d) Snowpark Container Services (SPCS) costs
Snowpark Container Services (SPCS) is a fully managed container environment that runs containerized workloads directly within Snowflake. It became generally available on AWS in August 2024, on Azure in February 2025, and on GCP in August 2025. It’s available on all paid Snowflake editions in commercial regions. Government regions and the Google Cloud me-central2 region are not yet supported. Trial accounts are excluded.
SPCS uses Compute Pools, which are separate from virtual warehouses. You pay for Compute Pool nodes based on instance type and running time; including idle nodes. If a node exists in IDLE or ACTIVE state, you’re paying for it.
First-generation SPCS instance credits per hour
| Instance Family | SPCS Compute | Openflow Compute (SPCS Deployment) |
| CPU_X64_XS | 0.06 | — |
| CPU_X64_S | 0.11 | 0.11 |
| CPU_X64_M | 0.22 | — |
| CPU_X64_SL | 0.41 | 0.41 |
| CPU_X64_L | 0.83 | 0.83 |
| HIGHMEM_X64_S | 0.28 | — |
| HIGHMEM_X64_M | 1.11 | — |
| HIGHMEM_X64_SL | 2.93 | — |
| HIGHMEM_X64_L | 4.44 | — |
| GPU_NV_XS | 0.25 | — |
| GPU_GCP_NV_L4_1_24G | 0.43 | — |
| GPU_NV_S | 0.57 | — |
| GPU_NV_SM | 1.70 | — |
| GPU_GCP_NV_L4_4_24G | 1.94 | — |
| GPU_NV_M | 2.68 | — |
| GPU_NV_2M | 3.50 | — |
| GPU_NV_3M | 3.55 | — |
| GPU_GCP_NV_A100_8_40G | 11.68 | — |
| GPU_NV_SL | 13.50 | — |
| GPU_NV_L | 14.12 | — |
SPCS Compute credits per hour (next-generation instances)
| Instance Family | AWS | Azure | GCP |
| GEN_ARM_G1_2 | 0.084 | N/A | N/A |
| GEN_ARM_G1_4 | 0.168 | N/A | N/A |
| GEN_ARM_G1_8 | 0.336 | N/A | N/A |
| GEN_ARM_G1_16 | 0.672 | N/A | N/A |
| GEN_ARM_G1_32 | 1.344 | N/A | N/A |
| GEN_X64_G2_2 | 0.092 | 0.086 | N/A |
| GEN_X64_G2_4 | 0.184 | 0.172 | N/A |
| GEN_X64_G2_8 | 0.368 | 0.344 | N/A |
| GEN_X64_G2_16 | N/A | 0.688 | N/A |
| GEN_X64_G2_32 | 1.472 | 1.376 | N/A |
| GPU_L40S_G1_8 | 1.580 | N/A | N/A |
| GPU_L40S_G1_16 | 3.160 | N/A | N/A |
| GPU_L40S_G1_48 | 9.480 | N/A | N/A |
| GPU_L40S_G1_192 | 37.920 | N/A | N/A |
| GPU_R6K_G1_8 | 2.537 | N/A | N/A |
| GPU_R6K_G1_16 | 5.074 | N/A | N/A |
| GPU_R6K_G1_32 | 10.148 | N/A | N/A |
| GPU_R6K_G1_48 | 15.222 | N/A | N/A |
| GPU_R6K_G1_96 | 30.444 | N/A | N/A |
| GPU_R6K_G1_192 | 60.888 | N/A | N/A |
| GPU_A100_G1_12 | N/A | N/A | 5.051 |
| GPU_A100_G1_48 | N/A | N/A | 20.204 |
| MEM_X64_G2_8 | 0.392 | 0.311 | N/A |
| MEM_X64_G2_32 | 1.568 | 1.244 | N/A |
| MEM_X64_G2_64 | 3.136 | 2.488 | N/A |
| MEM_X64_G2_96 | N/A | 3.732 | N/A |
| MEM_X64_G2_192 | 9.408 | N/A | N/A |
e) Snowflake Openflow cost
| Deployment type | Price |
| Openflow Compute — BYOC Deployment | 0.0225 credits per vCPU per hour |
| Openflow Compute — Snowflake Deployment (SPCS) | See first-generation SPCS instance table above |
f) Snowflake Postgres compute cost
| Instance Family | AWS | Azure | AWS (High Availability) | Azure (High Availability) |
| STANDARD_M | 0.0356 | 0.0376 | 0.0712 | 0.0752 |
| STANDARD_L | 0.0712 | 0.0752 | 0.1424 | 0.1504 |
| STANDARD_XL | 0.1424 | 0.1504 | 0.2848 | 0.3008 |
| STANDARD_2X | 0.2848 | 0.3008 | 0.5696 | 0.6016 |
| STANDARD_4XL | 0.5696 | 0.6016 | 1.1392 | 1.2032 |
| STANDARD_8XL | 1.1392 | 1.2032 | 2.2784 | 2.4064 |
| STANDARD_12XL | 1.7088 | 1.8048 | 3.4176 | 3.6096 |
| STANDARD_24XL | 3.4176 | 3.6096 | 6.8352 | 7.2192 |
| HIGHMEM_L | 0.1024 | 0.1088 | 0.2048 | 0.2176 |
| HIGHMEM_XL | 0.2048 | 0.2176 | 0.4096 | 0.4352 |
| HIGHMEM_2XL | 0.4096 | 0.4352 | 0.8192 | 0.8704 |
| HIGHMEM_4XL | 0.8192 | 0.8704 | 1.6384 | 1.7408 |
| HIGHMEM_8XL | 1.6384 | 1.7408 | 3.2768 | 3.4816 |
| HIGHMEM_12XL | 2.4576 | 2.6112 | 4.9152 | 5.2224 |
| HIGHMEM_16XL | 3.2768 | 3.4816 | 6.5536 | 6.9632 |
| HIGHMEM_24XL | 4.9152 | 5.2224 | 9.8304 | 10.4448 |
| HIGHMEM_32XL | 6.5536 | 6.9632 | 13.1072 | 13.9264 |
| HIGHMEM_48XL | 9.8304 | 10.4448 | 19.6608 | 20.8896 |
| BURST_XS | 0.0068 | — | 0.0136 | — |
| BURST_S | 0.0136 | 0.0144 | 0.0272 | 0.0288 |
| BURST_M | 0.0272 | 0.0288 | 0.0544 | 0.0576 |
3) Snowflake Cloud Services cost
The cloud services layer handles authentication, authorization, query compilation, SQL API execution, performance tuning, and metadata management. These services run continuously in the background.
Snowflake doesn’t charge you separately for cloud services up to a point. If cloud services usage stays within 10% of your daily virtual warehouse credit consumption, it’s included at no additional cost. If it exceeds that threshold, the excess is billed at the standard credit rate.
If you ran virtual warehouses consuming 100 credits in a day, up to 10 credits of cloud services usage is free. If cloud services consumed 12 credits that day, you’d be billed for 2 credits of overage. Most teams stay under the 10% threshold, so this rarely shows up as a meaningful cost driver.
4) Snowflake Data Transfer cost
Snowflake calculates data transfer costs based on factors such as the volume of data transferred, the direction of transfer (ingress or egress), the source or destination region and the cloud provider involved. Understanding these factors is essential for managing data transfer costs effectively.
Data Ingress cost
Data ingress (loading data into Snowflake) is always free. Data transfers within the same cloud provider and region are also free. Your internal queries, result sets and warehouse-to-warehouse movement within the same region don’t cost anything.
Data Egress cost
Data egress, or transferring data out of Snowflake, may incur charges depending on the destination and the cloud provider involved. When data is transferred within the same cloud provider and region, egress is typically free. However, if data is transferred to a different region or cloud provider, egress charges apply.
The egress charges are calculated based on the volume of data transferred and are typically billed on a per-gigabyte or per-terabyte basis. The rates vary among different cloud providers and regions. For example, transferring data from Snowflake to an Amazon S3 bucket in a different AWS region may incur egress charges at a different rate compared to transferring data to an Azure Blob Storage account in a different Azure region.
Take a look at some Snowflake pricing tables for comparison to get an idea of what the difference in cost could look like:
a) Snowflake Data Transfer costs on AWS (per TB)
| Region | Same Region | SPCS Same Region | Same Provider, Different Region | Different Provider or Internet |
| US East (N. Virginia) | $0.00 | $3.07 | $20.00 | $90.00 |
| US West (Oregon) | $0.00 | $3.07 | $20.00 | $90.00 |
| EU Dublin | $0.00 | $3.07 | $20.00 | $90.00 |
| EU Frankfurt | $0.00 | $3.07 | $20.00 | $90.00 |
| AP Sydney | $0.00 | $3.07 | $140.00 | $140.00 |
| AP Singapore | $0.00 | $3.07 | $90.00 | $120.00 |
| Canada Central | $0.00 | $3.07 | $20.00 | $90.00 |
| US East 2 (Ohio) | $0.00 | $3.07 | $20.00 | $90.00 |
| AP Northeast 1 (Tokyo) | $0.00 | $3.07 | $90.00 | $114.00 |
| AP Mumbai | $0.00 | $3.07 | $60.00 | $90.00 |
| US East 1 Commercial Gov | $0.00 | $3.07 | $20.00 | $90.00 |
| Europe (London) | $0.00 | $3.07 | $20.00 | $90.00 |
| Asia Pacific (Seoul) | $0.00 | $3.07 | $80.00 | $126.00 |
| US Gov West 1 | $0.00 | $7.17 | $30.00 | $155.00 |
| US Gov West 1 (FedRAMP High Plus) | $0.00 | $7.17 | $30.00 | $155.00 |
| Europe (Stockholm) | $0.00 | $3.07 | $20.00 | $90.00 |
| Asia Pacific (Osaka) | $0.00 | $3.07 | $90.00 | $114.00 |
| South America East 1 (Sao Paulo) | $0.00 | $3.07 | $138.00 | $150.00 |
| EU (Paris) | $0.00 | $3.07 | $20.00 | $90.00 |
| Asia Pacific (Jakarta) | $0.00 | $3.07 | $100.00 | $132.00 |
| US Gov East 1 (FedRAMP High Plus) | $0.00 | $7.17 | $30.00 | $155.00 |
| EU (Zurich) | $0.00 | $3.07 | $20.00 | $90.00 |
| US Gov West 1 (DoD) | $0.00 | $7.17 | $30.00 | $155.00 |
| US West (Commercial Gov – Oregon) | $0.00 | $3.07 | $20.00 | $90.00 |
| Africa (Cape Town) | $0.00 | $3.07 | $147.00 | $154.00 |
| Middle East (UAE) | $0.00 | $3.07 | $85.00 | $110.00 |
| Asia Pacific (Malaysia) | $0.00 | $3.07 | $80.00 | $108.00 |
| Asia Pacific (Thailand) | $0.00 | $3.07 | $80.00 | $108.00 |
| Asia Pacific (New Zealand) | $0.00 | $3.07 | $98.00 | $114.00 |
b) Snowflake Data Transfer costs on Azure (per TB)
| Region | Same Region | SPCS Same Region | Same Provider, Same Continent | Same Provider, Different Continent | Different Provider or Internet |
| East US 2 (Virginia) | $0.00 | $0.00 | $20.00 | $50.00 | $87.50 |
| West US 2 (Washington) | $0.00 | $0.00 | $20.00 | $50.00 | $87.50 |
| West Europe (Netherlands) | $0.00 | $0.00 | $20.00 | $50.00 | $87.50 |
| Australia East (New South Wales) | $0.00 | $0.00 | $80.00 | $80.00 | $120.00 |
| Canada Central (Toronto) | $0.00 | $0.00 | $20.00 | $50.00 | $87.50 |
| Southeast Asia (Singapore) | $0.00 | $0.00 | $80.00 | $80.00 | $120.00 |
| Switzerland North | $0.00 | $0.00 | $20.00 | $50.00 | $87.50 |
| US Gov Virginia | $0.00 | $0.00 | $20.00 | $50.00 | $87.50 |
| US Central (Iowa) | $0.00 | $0.00 | $20.00 | $50.00 | $87.50 |
| North Europe (Ireland) | $0.00 | $0.00 | $20.00 | $50.00 | $87.50 |
| Japan East (Tokyo) | $0.00 | $0.00 | $80.00 | $80.00 | $120.00 |
| UAE North (Dubai) | $0.00 | $0.00 | $80.00 | $80.00 | $120.00 |
| South Central US (Texas) | $0.00 | $0.00 | $20.00 | $50.00 | $87.50 |
| Central India (Pune) | $0.00 | $0.00 | $40.00 | $50.00 | $87.50 |
| UK South (London) | $0.00 | $0.00 | $20.00 | $50.00 | $87.50 |
| US Gov Virginia (FedRAMP High Plus) | $0.00 | $0.00 | $20.00 | $50.00 | $87.50 |
| Mexico Central | $0.00 | $0.00 | $20.00 | $50.00 | $87.00 |
| Korea Central | $0.00 | $0.00 | $80.00 | $80.00 | $120.00 |
| Sweden Central | $0.00 | $0.00 | $20.00 | $50.00 | $87.00 |
| East US (Virginia) | $0.00 | $0.00 | $20.00 | $50.00 | $87.00 |
c) Snowflake Data Transfer costs on GCP (per TB)
Same cloud provider transfers:
| Region | Same Region | SPCS Same Region | North America | Europe | Asia | Indonesia | Middle East | Oceania | Africa | South America |
| US Central 1 (Iowa) | $0 | $6.22 | $20 | $50 | $80 | $100 | $110 | $100 | $110 | $140 |
| US East 4 (N. Virginia) | $0 | $6.22 | $20 | $50 | $80 | $100 | $110 | $100 | $110 | $140 |
| Europe West 4 (Netherlands) | $0 | $6.22 | $50 | $20 | $80 | $100 | $110 | $100 | $110 | $140 |
| Europe West 2 (London) | $0 | $6.22 | $50 | $20 | $80 | $100 | $110 | $100 | $110 | $140 |
| Europe West 3 (Frankfurt) | $0 | $6.22 | $50 | $20 | $80 | $100 | $110 | $100 | $110 | $140 |
| Middle East Central 2 (Dammam) | $0 | $6.22 | $110 | $110 | $110 | $110 | $80 | $110 | $110 | $140 |
| Australia Southeast 2 (Melbourne) | $0 | $6.22 | $100 | $100 | $100 | $80 | $110 | $80 | $140 | $140 |
Different cloud provider or internet transfers:
| Region | North America | Europe | Asia | AU/Indonesia/Korea/South America/Saudi Arabia | Middle East & Africa | China |
| US Central 1 (Iowa) | $120 | $120 | $120 | $190 | $150 | $230 |
| US East 4 (N. Virginia) | $120 | $120 | $120 | $190 | $150 | $230 |
| Europe West 4 (Netherlands) | $120 | $120 | $120 | $190 | $150 | $230 |
| Europe West 2 (London) | $120 | $120 | $120 | $190 | $150 | $230 |
| Europe West 3 (Frankfurt) | $120 | $120 | $120 | $190 | $150 | $230 |
| Middle East Central 2 (Dammam) | $190 | $190 | $190 | $190 | $190 | $230 |
| Australia Southeast 2 (Melbourne) | $190 | $190 | $190 | $190 | $190 | $230 |
Check out the article below to learn more in-depth on Snowflake’s data transfer costs and how they are calculated:
Snowflake Data Transfer Costs 101: An In-Depth Guide
5) Snowflake AI Services Cost
Snowflake AI Features run on Snowflake-managed compute resources and bill based on:
- Tokens processed (input/output)
- Compute time (hours)
- Units processed (pages, messages, GB, etc.)
Note: When prices are listed per-quantity (e.g., per 1 million tokens), you’re billed proportionally for all units used, not only in whole-quantity increments.
AI credit pricing by region
| Region | On-Demand | Tier 1 | Tier 2 | Tier 3 | Tier 4 | Tier 5 | Tier 6 | Tier 7 |
| Global | $2.00 | $2.00 | $1.96 | $1.96 | $1.94 | $1.92 | $1.90 | $1.88 |
| Regional | $2.20 | $2.20 | $2.16 | $2.16 | $2.13 | $2.11 | $2.09 | $2.07 |
- Tier 1 = $0-$1,199,999
- Tier 2 = $1,200,000-$2,999,999
- Tier 3 = $3,000,000-$4,999,999
- Tier 4 = $5,000,000-$9,999,999
- Tier 5 = $10,000,000-$19,999,999
- Tier 6 = $20,000,000-$39,999,999
- Tier 7 = $40,000,000+
a) Cortex AI function credits (per million tokens)
| Model | Input | Output |
| AI_COMPLETE — claude-haiku-4-5 | 0.55 | 2.75 |
| AI_COMPLETE — claude-opus-4-5 | 2.75 | 13.75 |
| AI_COMPLETE — claude-opus-4-6 | 2.75 | 13.75 |
| AI_COMPLETE — claude-opus-4-7 | 2.75 | 13.75 |
| AI_COMPLETE — claude-opus-4-8 | 3.00 | 15.00 |
| AI_COMPLETE — claude-sonnet-4-5 | 1.65 | 8.25 |
| AI_COMPLETE — claude-sonnet-4-5-long-context | 3.30 | 12.38 |
| AI_COMPLETE — claude-sonnet-4-6 | 1.65 | 8.25 |
| AI_COMPLETE — deepseek-r1 | 0.68 | 2.70 |
| AI_COMPLETE — gemini-2-5-flash | 0.15 | 1.25 |
| AI_COMPLETE — gemini-2-5-flash-lite | 0.05 | 0.20 |
| AI_COMPLETE — gemini-3.1-pro | 1.10 | 6.60 |
| AI_COMPLETE — gemini-3.1-pro-long-context | 2.20 | 9.90 |
| AI_COMPLETE — gemini-3.5-flash | 0.90 | 5.40 |
| AI_COMPLETE — llama3.1-405b | 1.20 | 1.20 |
| AI_COMPLETE — llama3.1-70b | 0.36 | 0.36 |
| AI_COMPLETE — llama3.1-8b | 0.11 | 0.11 |
| AI_COMPLETE — llama3.3-70b | 0.36 | 0.36 |
| AI_COMPLETE — llama4-maverick | 0.12 | 0.49 |
| AI_COMPLETE — llama4-scout | 0.09 | 0.33 |
| AI_COMPLETE — mistral-large2 | 1.00 | 3.00 |
| AI_COMPLETE — mistral-7b | 0.08 | 0.10 |
| AI_COMPLETE — mixtral-8x7b | 0.23 | 0.35 |
| AI_COMPLETE — openai-gpt-4.1 | 1.00 | 4.00 |
| AI_COMPLETE — openai-gpt-5 | 0.69 | 5.50 |
| AI_COMPLETE — openai-gpt-5-mini | 0.14 | 1.10 |
| AI_COMPLETE — openai-gpt-5-nano | 0.03 | 0.22 |
| AI_COMPLETE — openai-gpt-5.1 | 0.69 | 5.50 |
| AI_COMPLETE — openai-gpt-5.2 | 0.97 | 7.70 |
| AI_COMPLETE — openai-gpt-5.4 | 1.38 | 8.25 |
| AI_COMPLETE — openai-gpt-5.4-long-context | 2.75 | 12.38 |
| AI_COMPLETE — openai-gpt-5.5 | 2.75 | 16.50 |
| AI_COMPLETE — openai-gpt-5.5-long-context | 5.50 | 24.75 |
| AI_COMPLETE — pixtral-large | 1.00 | 3.00 |
| AI_COMPLETE — qwen3-32b | 0.09 | 0.36 |
| AI_COMPLETE — qwen3-next-80b-a3b | 0.09 | 0.72 |
| AI_COMPLETE — qwen3-vl-235b-a22b | 0.32 | 1.60 |
| AI_COMPLETE — snowflake-llama-3.1-405b | 0.96 | 0.96 |
| AI_COMPLETE — snowflake-llama-3.3-70b | 0.29 | 0.29 |
| AI_AGG | 1.60 | — |
| AI_CLASSIFY | 1.39 | — |
| AI_EMBED — voyage-multimodal-3 | 0.06 | — |
| AI_EMBED — multilingual-e5-large | 0.05 | — |
| AI_EMBED — nv-embed-qa-4 | 0.05 | — |
| AI_EMBED — snowflake-arctic-embed-l-v2.0 | 0.05 | — |
| AI_EMBED — voyage-multilingual-2 | 0.07 | — |
| AI_EMBED — e5-base-v2 | 0.03 | — |
| AI_EMBED — snowflake-arctic-embed-m | 0.03 | — |
| AI_EMBED — snowflake-arctic-embed-m-v1.5 | 0.03 | — |
| AI_EXTRACT — arctic-extract | 5.00 | — |
| AI_FILTER | 1.39 | — |
| AI_GUARDRAILS | 0.35 | — |
| AI_REDACT | 0.63 | — |
| AI_SENTIMENT | 1.60 | — |
| AI_SUMMARIZE_AGG | 1.60 | — |
| AI_TRANSCRIBE | 1.30 | — |
| AI_TRANSLATE | 1.50 | — |
| Extract Answer | 0.08 | — |
| Guard | 0.25 | — |
| Sentiment | 0.08 | — |
| Summarize | 0.10 | — |
Legacy Cortex features:
| Model | Credits per million tokens |
| AI_COMPLETE — gemma-7b | 0.12 |
| AI_COMPLETE — jamba-instruct | 0.83 |
| AI_COMPLETE — jamba-1.5-large | 1.40 |
| AI_COMPLETE — jamba-1.5-mini | 0.10 |
| AI_COMPLETE — llama2-70b-chat | 0.45 |
| AI_COMPLETE — llama3-70b | 1.21 |
| AI_COMPLETE — llama3-8b | 0.19 |
| AI_COMPLETE — llama3.2-1b | 0.04 |
| AI_COMPLETE — llama3.2-3b | 0.06 |
| AI_COMPLETE — mistral-large | 5.10 |
| AI_COMPLETE — reka-core | 5.50 |
| AI_COMPLETE — reka-flash | 0.45 |
| AI_EXTRACT — arctic-tilt-entity | 9.50 |
| AI_EXTRACT — arctic-tilt-table | 28.40 |
| AI_COMPLETE — claude-4-sonnet | 1.50 (input) / 7.50 (output) |
b) REST API with prompt caching ($ per million tokens)
| Model | Inference Region | Input | Output | Cache Write | Cache Read |
| claude-4-sonnet | AWS Regional | $3.00 | $15.00 | $3.75 | $0.30 |
| claude-sonnet-4-5 | AWS Regional | $3.30 | $16.50 | $4.13 | $0.33 |
| claude-sonnet-4-5 | AWS Global | $3.00 | $15.00 | $3.75 | $0.30 |
| claude-sonnet-4-5-long-context | AWS Regional | $6.60 | $24.75 | $8.25 | $0.66 |
| claude-sonnet-4-5-long-context | AWS Global | $6.00 | $22.50 | $7.50 | $0.60 |
| claude-sonnet-4-6 | AWS Regional | $3.30 | $16.50 | $4.13 | $0.33 |
| claude-sonnet-4-6 | AWS Global | $3.00 | $15.00 | $3.75 | $0.30 |
| claude-haiku-4-5 | AWS Regional | $1.10 | $5.50 | $1.38 | $0.11 |
| claude-haiku-4-5 | AWS Global | $1.00 | $5.00 | $1.25 | $0.10 |
| claude-opus-4-5 | AWS Regional | $5.50 | $27.50 | $6.88 | $0.55 |
| claude-opus-4-5 | AWS Global | $5.00 | $25.00 | $6.25 | $0.50 |
| claude-opus-4-6 | AWS Regional | $5.50 | $27.50 | $6.88 | $0.55 |
| claude-opus-4-6 | AWS Global | $5.00 | $25.00 | $6.25 | $0.50 |
| claude-opus-4-7 | AWS Regional | $5.50 | $27.50 | $6.88 | $0.55 |
| claude-opus-4-7 | AWS Global | $5.00 | $25.00 | $6.25 | $0.50 |
| claude-opus-4-8 | AWS Regional | $5.50 | $27.50 | $6.88 | $0.55 |
| claude-opus-4-8 | AWS Global | $5.00 | $25.00 | $6.25 | $0.50 |
| openai-gpt-4.1 | Azure Regional | $2.20 | $8.80 | — | $0.55 |
| openai-gpt-4.1 | Azure Global | $2.00 | $8.00 | — | $0.50 |
| openai-gpt-5 | Azure Regional | $1.38 | $11.00 | — | $0.14 |
| openai-gpt-5 | Azure Global | $1.25 | $10.00 | — | $0.13 |
| openai-gpt-5-mini | Azure Regional | $0.28 | $2.20 | — | $0.03 |
| openai-gpt-5-mini | Azure Global | $0.25 | $2.00 | — | $0.03 |
| openai-gpt-5-nano | Azure Regional | $0.06 | $0.44 | — | $0.01 |
| openai-gpt-5.1 | Azure Regional | $1.38 | $11.00 | — | $0.14 |
| openai-gpt-5.1 | Azure Global | $1.25 | $10.00 | — | $0.13 |
| openai-gpt-5.2 | Azure Regional | $1.93 | $15.40 | — | $0.19 |
| openai-gpt-5.2 | Azure Global | $1.75 | $14.00 | — | $0.18 |
| openai-gpt-5.4 | Azure Regional | $2.75 | $16.50 | — | $0.28 |
| openai-gpt-5.4 | Azure Global | $2.50 | $15.00 | — | $0.25 |
| openai-gpt-5.4-long-context | Azure Regional | $5.50 | $24.75 | — | $0.55 |
| openai-gpt-5.4-long-context | Azure Global | $5.00 | $22.50 | — | $0.50 |
| openai-gpt-5.5 | Azure Regional | $5.50 | $33.00 | — | $0.55 |
| openai-gpt-5.5 | Azure Global | $5.00 | $30.00 | — | $0.50 |
| openai-gpt-5.5-long-context | Azure Regional | $11.00 | $49.50 | — | $1.10 |
| openai-gpt-5.5-long-context | Azure Global | $10.00 | $45.00 | — | $1.00 |
c) REST API credits ($ per million tokens)
| Model | Input | Output |
| deepseek-r1 | $1.35 | $5.40 |
| llama3.1-405b | $2.40 | $2.40 |
| llama3.1-70b | $0.72 | $0.72 |
| llama3.1-8b | $0.22 | $0.22 |
| llama3.2-1b | $0.10 | $0.10 |
| llama3.2-3b | $0.15 | $0.15 |
| llama3.3-70b | $0.72 | $0.72 |
| llama4-maverick | $0.24 | $0.97 |
| mistral-large | $4.00 | $12.00 |
| mistral-large2 | $2.00 | $6.00 |
| mistral-7b | $0.15 | $0.20 |
| snowflake-llama-3.3-70b | $0.72 | $0.72 |
d) Snowflake Intelligence, Cortex Agents and Cortex Analyst (AI credits per million tokens)
| Model | Input | Output | Cache Write | Cache Read |
| claude-haiku-4-5 | 0.65 | 3.25 | 0.81 | 0.06 |
| claude-opus-4-5 | 3.25 | 16.26 | 4.07 | 0.33 |
| claude-opus-4-6 | 3.25 | 16.26 | 4.07 | 0.33 |
| claude-opus-4-7 | 3.25 | 16.26 | 4.07 | 0.33 |
| claude-opus-4-8 | 3.25 | 16.26 | 4.07 | 0.33 |
| claude-sonnet-4-5 | 1.95 | 9.76 | 2.44 | 0.20 |
| claude-sonnet-4-6 | 1.95 | 9.76 | 2.44 | 0.20 |
| gemini-2-5-flash | 0.18 | 1.48 | — | 0.02 |
| gemini-3.1-pro | 1.30 | 7.81 | — | 0.13 |
| openai-gpt-4.1 | 1.30 | 5.20 | — | 0.33 |
| openai-gpt-5 | 0.81 | 6.51 | — | 0.08 |
| openai-gpt-5-mini | 0.16 | 1.30 | — | 0.02 |
| openai-gpt-5.1 | 0.81 | 6.51 | — | 0.08 |
| openai-gpt-5.2 | 1.14 | 9.11 | — | 0.11 |
| openai-gpt-5.4 | 1.63 | 9.76 | — | 0.16 |
| openai-gpt-5.5 | 3.25 | 19.50 | — | 0.33 |
| openai-gpt-5.5-long-context | 6.50 | 29.25 | — | 0.65 |
e) Cortex Code credits (AI credits per million tokens)
| Model | Input | Output | Cache Write | Cache Read |
| claude-opus-4-5 | 2.75 | 13.75 | 3.44 | 0.28 |
| claude-opus-4-6 | 2.75 | 13.75 | 3.44 | 0.28 |
| claude-opus-4-7 | 2.75 | 13.75 | 3.44 | 0.28 |
| claude-opus-4-8 | 2.75 | 13.75 | 3.44 | 0.28 |
| claude-sonnet-4-5 | 1.65 | 8.25 | 2.07 | 0.17 |
| claude-sonnet-4-6 | 1.65 | 8.25 | 2.07 | 0.17 |
| openai-gpt-5.2 | 0.97 | 7.70 | — | 0.10 |
| openai-gpt-5.4 | 1.38 | 8.25 | — | 0.14 |
| openai-gpt-5.5 | 2.75 | 16.50 | — | 0.28 |
| openai-gpt-5.5-long-context | 5.50 | 24.75 | — | 0.55 |
f) Fine-tuning credits (per million tokens)
| Feature | Training | Inference (Cortex Complete) |
| AI EXTRACT — arctic-extract — finetuned | 0 | 10 |
| Cortex Fine-tuning — llama3.1-70b | 3.40 | 2.42 |
| Cortex Fine-tuning — llama3.1-8b | 0.64 | 0.38 |
| Cortex Fine-tuning — mistral-7b | 0.64 | 0.24 |
| Cortex Fine-tuning — mixtral-8x7b | 3.40 | 0.44 |
| Cortex Fine-tuning — llama3-70b (legacy) | 3.40 | 2.42 |
| Cortex Fine-tuning — llama3-8b (legacy) | 0.64 | 0.38 |
g) Other AI feature credits
| Feature | Rate |
| AI_COMPLETE — twelvelabs-pegasus-1-2 | 0.294 credits per 1,000 video seconds + 4.5 credits per million output tokens |
| AI_MULTI_EMBED — twelvelabs-marengo-embed-3-0 | 0.420 credits/1,000 video seconds / 0.084/1,000 audio seconds / 0.060/1,000 image requests / 0.042/1,000 text requests |
| AI Parse Document — Layout | 3.33 credits per 1,000 pages |
| AI Parse Document — OCR | 0.5 credits per 1,000 pages |
| Batch Cortex Search | 0.12 credits per GB/hr of indexed data |
| Cortex Analyst | 67 credits per 1,000 messages |
| Cortex Search | 6.3 credits per GB/mo of indexed data |
| Document AI | 8 credits per hour of compute |
h) Provisioned throughput credits
| Cloud Provider | Credits per PTU per hour | Term Length |
| AWS | 0.08 | 1 month |
| Azure | 0.10 | 1 month |
Optimizing Snowflake costs
Here are some tips and tricks for optimizing Snowflake costs.
Right-size your warehouses. An X-Large warehouse isn’t automatically better than a Large. Snowflake’s performance scales with warehouse size, but many analytical queries don’t get faster just because you added more compute. Test your workloads at different sizes. Many teams run queries on a Large that ran just as well on a Medium.
Set auto-suspend aggressively. 60 seconds is a reasonable default for most interactive query workloads. If a warehouse resumes for a 5-second query and then stays running for 5 minutes because auto-suspend is set to 300 seconds, you’ve wasted 295 seconds of compute at whatever your warehouse credit rate is.
Use result caching. Snowflake caches query results and serves identical repeat queries from cache at zero credit cost. Design dashboards and scheduled reports to take advantage of this — consistent query text and parameters are what trigger cache hits.
Monitor serverless features. Materialized views, automatic clustering, and search optimization run on serverless compute regardless of whether your warehouses are active. Use SNOWFLAKE.ACCOUNT_USAGE.SERVERLESS_TASK_HISTORY and similar views to track what’s consuming credits outside your warehouses.
Separate workloads into dedicated warehouses. ELT pipelines, interactive dashboards, and ad hoc queries have very different compute requirements. Mixing them on a single warehouse means either over-provisioning for interactive users or throttling your pipelines. Dedicated warehouses let you size and suspend each workload independently.
Watch your Time Travel window. On Enterprise, 90-day Time Travel retains a copy of every version of your data for 90 days. On frequently updated tables, this compounds quickly. If you don’t need 90 days, lower it. Most operational tables work fine with 1-7 days.
Use query tags and resource monitors. Snowflake’s built-in Resource Monitor can cap credit usage per warehouse or per account and trigger alerts when thresholds are hit. Query tags let you attribute costs to specific teams, pipelines, or use cases so you can see where spend is actually going.
Use third-party observability tools. Snowflake’s native dashboards give you usage data, but connecting that data to actual business context like which team, which pipeline, which query pattern is driving cost, which often requires more granular tooling. Platforms like Flexera One’s Data Cloud Optimization solutions can surface anomalies and attribute costs in ways that the default Snowflake UI doesn’t.
If you’re looking to further optimize your Snowflake spending and cut costs significantly, be sure to check out our article “Best Practices to Reduce Snowflake Costs“. It’s packed with practical tips and strategies that can help you make the most of your Snowflake investment and keep your expenses in check.
Tired of feeling in the dark about your Snowflake spending? Schedule a demo with us today and see the difference it can make.
Save up to 30% on your Snowflake spend in a few minutes!
References
- Snowflake Credit Consumption Table
- Snowflake documentation — Understanding compute cost
- Snowflake documentation — Warehouse overview
- Snowflake documentation — Gen2 standard warehouses
- Snowflake documentation — Understanding overall cost
- Snowflake documentation — Time Travel and Fail-safe
- Snowflake blog — Consumption-based pricing
- Snowflake blog — Gen2 warehouses generally available
- Snowflake blog — Snowpark Container Services GA
Conclusion
And that’s a wrap! Snowflake’s consumption-based pricing model gives you real flexibility; no idle hardware costs, no upfront provisioning and the ability to scale compute and storage completely independently. The other side is that the bill is entirely proportional to how well you manage it. Teams that run warehouses without auto-suspend, default to full-refresh ingestion pipelines or forget that serverless features run continuously can see 70% year-over-year cost growth as they consolidate workloads. Teams that actively right-size, set aggressive auto-suspend and track serverless usage have documented 30 to 50% bill reductions. This guide covered Snowflake’s four editions and their features, the full breakdown of compute, storage and data transfer costs and the most effective cost optimization strategies. The consumption model rewards management effort directly and that’s either a feature or a risk, depending on how much attention you pay to it.
FAQs
How does Snowflake pricing work?
Snowflake uses a consumption-based model. You pay for three separate meters: compute (credits consumed by virtual warehouses, billed per second with a 60-second minimum), storage (per compressed TB per month), and data transfer (for egress across regions or cloud providers). There’s no flat subscription fee.
Can I use Snowflake for free?
Snowflake doesn’t have a permanent free tier. It offers a 30-day free trial with $400 in credits for new accounts, no credit card required. The trial ends when you exhaust the credits or after 30 days, whichever comes first.
What are the Snowflake pricing editions?
Four editions: Standard (~$2/credit on-demand, AWS US East), Enterprise (~$3/credit), Business Critical (~$4/credit), and Virtual Private Snowflake (custom pricing). Each adds progressively more security, compliance, and governance features.
What does a Snowflake credit cost in 2026?
On-demand US AWS rates are approximately $2/credit (Standard), $3/credit (Enterprise), and $4/credit (Business Critical). Non-US regions add roughly 10-50% depending on location. Capacity contracts reduce these rates by 20-45% depending on volume and term length.
How much does Snowflake storage cost?
Storage is billed in dollars per TB per month (not credits). On AWS US East, the on-demand storage rate per Snowflake’s official consumption table is $23/TB per month. Some external sources cite $40/TB as the true on-demand list price before any capacity contract terms apply. If you haven’t signed a capacity contract, use $40/TB for conservative budgeting. The capacity Tier 1 rate for AWS US East is $23/TB per month.
What’s the difference between on-demand and capacity storage in Snowflake?
On-demand is month-to-month with no commitment, billed at the full rate. Capacity involves pre-purchasing a storage commitment in exchange for a lower rate — roughly 40% cheaper on AWS US East at higher tiers. The minimum capacity contract is typically $25,000/year.
How does the virtual warehouse billing minimum work?
Each time a warehouse starts or resumes, you’re billed for a minimum of 60 seconds regardless of how long the query takes. After the first 60 seconds, billing continues per second. Resuming a warehouse multiple times for short queries means paying the 60-second minimum each time.
What are Gen2 warehouses and do they cost more?
Snowflake Gen2 standard warehouses run on faster hardware and became generally available in November 2025 across AWS, Azure, and GCP. They cost 1.35x the credits per hour of Gen1 on AWS and GCP, and 1.25x on Azure. They’re only available for sizes X-Small through 4X-Large. Whether you save money depends on how much faster Gen2 finishes your specific workloads.
Does Snowflake charge for idle warehouses?
No. Suspended warehouses consume no credits. Auto-suspend must be configured — it’s not enabled by default on all warehouse types, and leaving it disabled means you pay for idle time.
How does the cloud services layer billing work?
Cloud services (query parsing, metadata operations, authentication) are free up to 10% of your daily virtual warehouse credit consumption. If cloud services exceed that threshold, Snowflake bills the excess at the standard credit rate. Most teams stay under this threshold.
Does Snowflake offer discounts?
Yes, through capacity contracts. Annual or multi-year commitments with sufficient volume (typically $25,000+ per year minimum) qualify for discounted credit and storage rates. Volume discounts increase with commitment size — organizations committing to 100,000+ credits annually can reduce per-credit costs by 25-45%.
Is Snowflake expensive?
It depends on how it’s managed. Unmanaged, with warehouses running without auto-suspend, full-refresh ingestion pipelines, and serverless features left on, costs grow fast. With right-sizing, proper auto-suspend, and query optimization, teams have documented 30-50% bill reductions. The consumption model rewards management effort directly.
How do I estimate Snowflake costs before committing?
Start with the trial account ($400 in credits over 30 days) and run your actual workloads. Pay attention to which warehouse sizes complete your queries without queue delays, how much data you’re storing after compression, and whether any serverless features are consuming credits. Real usage data from a trial gives you a far more accurate estimate than any calculator.
What are the biggest hidden costs in Snowflake?
The most common budget surprises: materialized views and search optimization running continuously after being enabled; full-refresh data ingestion pipelines loading unchanged data every few hours; Time Travel storage on frequently updated tables; Cortex AI features generating token-based charges that don’t appear in warehouse dashboards; and cross-region or cross-cloud data egress fees that show up as cloud provider charges rather than Snowflake line items.
How do I reduce my Snowflake bill?
The highest-impact actions: set auto-suspend to 60 seconds on all warehouses, right-size warehouses by testing workloads at smaller sizes before defaulting to larger ones, turn off serverless features you’re not actively using, switch SaaS ingestion pipelines from full-refresh to incremental loads, use result caching by keeping query text consistent, and implement resource monitors to catch unexpected usage spikes before they become large invoices.
Is Snowflake better than BigQuery?
Different tools for different contexts. Snowflake runs natively on AWS, Azure, and GCP, making it cloud-agnostic — a significant advantage for multi-cloud organizations or teams that want portability. BigQuery runs natively only on GCP and has a different pricing model (on-demand per-byte scanned for queries, or flat-rate slot pricing). Snowflake separates compute and storage; BigQuery does too at the architecture level, but the pricing abstraction differs. The right choice depends on your existing cloud environment and workload characteristics.
Is Snowflake better than Databricks?
They serve different primary use cases. Snowflake is optimized for SQL-based analytics and structured data warehousing. Databricks is built around data engineering, machine learning, and the Lakehouse architecture with Apache Spark and Delta Lake. Many organizations run both: Databricks for data engineering and ML pipelines, Snowflake for business analytics and reporting. Comparing costs directly is difficult because the overlap in capabilities doesn’t cover the full use case.
Is Snowflake better than AWS Redshift?
Snowflake and Redshift have meaningfully different architectures. Snowflake separates compute and storage from the ground up, enabling instant scaling without data movement. Redshift traditionally coupled them, though Redshift Serverless has closed some of that gap. Snowflake has stronger native support for semi-structured data (JSON, Avro, Parquet) and built-in data sharing capabilities. Redshift has tighter native integration with the AWS ecosystem. If you’re already heavily invested in AWS services, Redshift may have integration advantages. If you need multi-cloud portability or data sharing, Snowflake has an edge.
What happens when my Snowflake trial ends?
When your $400 in trial credits is exhausted, or after 30 days, you can no longer run compute or access storage until you upgrade to a paid account. Your data isn’t immediately deleted, but your account is suspended. Upgrading converts the account to your chosen edition and billing model.
Can I switch Snowflake editions after signing up?
Yes. You can upgrade your edition at any time. Downgrading is also possible but requires coordination with Snowflake support to avoid losing access to features that depend on the higher edition tier (like 90-day Time Travel data retained under Enterprise).
