🎉 Read about our $5.4M funding announcement on TechCrunch! 🎉


Tue Jun 06 2023

Snowflake Pricing: Breaking Down Snowflake Costs


Graeme Caldwell

Snowflake is a cloud-based data warehouse that enables users to store, query, and analyze structured, semi-structured, and unstructured data. Snowflake is designed to be scalable, secure, and easy to use, with features such as automatic scaling, data sharing, and data lake integration.

Snowflake also claims to offer lower costs than traditional on-premises data warehouses or data lakes. But how does Snowflake pricing work, and what are the benefits and risks of using Snowflake for your data needs?

Understanding the Components of Snowflake Pricing

Snowflake pricing is based on three main components: storage, compute, and data transfer.


Storage costs vary according to the amount of data stored in Snowflake. The platform charges a monthly fee per terabyte of storage used, averaged over a month. The fee varies depending on the location where the data is stored and whether the user opts for premium support services. Pre-purchasing storage capacity can significantly lower the per-terabyte cost, but that depends on the users' ability to predict how much capacity they will need from month to month.


Compute is the processing power used to run queries and perform operations in Snowflake. There are three aspects to Snowflake's compute costs. The most important is the per-second fee for compute used by each virtual warehouse (a cluster of servers that execute queries). The additional compute costs come, firstly, from serverless operations such as search optimization and, secondly, from cloud services, which include various data management and processing services such as authentication and metadata control.

Compute is paid for in Snowflake credits. For virtual warehouses, the number of credits increases with the warehouse's size and what Snowflake refers to as its "Edition". An Edition is essentially a package of services, ranging from the Standard Edition with basic services through the Premier, Enterprise, and Enterprise for Sensitive Data tiers, which add services like round-the-clock support, federated authentication, and regulatory compliance.

Data transfer

Snowflake charges a per-byte fee for data egress—data transferred out of Snowflake, between cloud platforms, and between different regions of the same platform. The fee varies depending on the cloud provider and the data’s destination.

Snowflake Costs vs. On-Premises Data Warehouse Costs

One of the main advantages of Snowflake pricing is that it is pay-as-you-go: users only pay for resources they consume, and they can scale up or down as needed.

For example, a user on the Standard Edition can expect to pay $40 per terabyte per month for data storage in AWS’s US East region, $20 per TB for transfers between regions, and $90 per TB for transfers to a different cloud platform. Compute costs for a Standard account are $2.00 per credit, with a Small Standard Virtual Warehouse costing 2 credits per hour, with larger Warehouses costing additional credits.

These costs are known in advance and are straightforward to calculate.  Most importantly, the amount a business actually spends depends on their usage, which can be scaled up or down depending on their needs.

This contrasts with traditional on-premises data warehouses or data lakes, which require upfront capital expenditure for hardware and software licenses as well as ongoing scaling, maintenance,  and operational costs. The business must support large fixed costs, regardless of how its storage and data needs change over time.

Snowflake offers better performance and efficiency than on-premises solutions by leveraging the elasticity and availability of the cloud. Snowflake can automatically scale a virtual warehouse’s compute resources up or down based on need without affecting the storage tier or other warehouses. This elasticity helps users avoid overprovisioning or underutilizing infrastructure.

The Benefits of Snowflake Pricing

The main benefit of Snowflake pricing is that users can save money by paying only for what they use, scaling up or down as needed, and avoiding upfront and ongoing costs associated with on-premises solutions. The platform also includes features to help users manage and reduce costs, such as auto-suspension and resumption of data warehouses and the ability to pre-purchase credits for less than on-demand credits.

The Risks of Snowflake's Pricing Model

Snowflake pricing also comes with some potential risks that users should be aware of before adopting it. The biggest risks are unpredictability and the possibility of rapidly spiraling costs when the system scales up to accommodate growing workloads. Users may face unexpected or fluctuating costs due to factors such as query complexity, concurrency, data volume, or data transfer. They should monitor their usage and set budgets and alerts to avoid overspending.

Balancing Scalability and Cost Management in Snowflake

Snowflake is a powerful cloud-based data warehouse solution that offers numerous benefits. Its pay-as-you-go pricing model allows users to save money by paying only for what they use and to optimize their costs by leveraging the cloud's elasticity and availability. However, it is essential for users to keep a close eye on their costs to avoid rapidly increasing expenses and unpredictable bills.

Reducing Snowflake Costs, a How To

Data pipelines can be expensive, and companies that price based on compute time have incentive for all the data dirty work to be done on their platform. What if only a sample of the data is required to get the results you need? How much more could you get out of Snowflake if data was cleaned and better prepared before ingestion? With Streamdal, you can achieve this and much more.

Looking at the second example on Snowflake’s pricing guide, we see that the storage cost for 65TB (after compression) of data might cost $17,940 a year. Let's say you only need to transfer half of that data to snowflake.

Streamdal allows you to:

  • Tap into any data stream, and observe and search data regardless of encoding. When you use Streamdal to tap into your data, you are effectively creating a datalake logically indexed in parquet.
  • Run serverless functions on data. From ingest to egress, you can run functions on data to determine how and what data is stored.
  • Set up monitors and alerts. Set up monitors for rate, fields like PII, schema changes, and much more.

Before a table or a query breaks from bad data, you’ll be alerted by Streamdal.

By sending half of the data with this example, you could save $8,970. How much more savings could you unlock by preparing data before transferring? Would you still need the 2x virtual warehouse provisions if you could eliminate the need to run complex queries? Learn more about our snowflake cost optimization solution, give us a try today, or book a demo to see how much you could save.

Graeme Caldwell

Technical Writer

Graeme is a copywriter and technical writer who has spent over a decade helping businesses to translate complex ideas into engaging content. Graeme's writing spans numerous fields, including technology, finance, compliance, and marketing.

Continue Exploring

Data Protection

Thu Jul 27 2023

Data Protection: Challenges and Opportunities


Daniel Selans

Explore data protection strategies, key regulations, and the role of automation in safeguarding sensitive information in an ever-evolving digital landscape.

Read more >
Data Protection

Wed Jul 19 2023

Data Consistency in Distributed Enterprise Applications


Daniel Selans

Learn about data consistency in distributed enterprise apps, why it matters, and how to maintain it using validation and real-time data monitoring.

Unlock real-time data visibility today!

Get Started

backed by


Privacy PolicyTerms and Conditions

© 2023 Streamdal, Inc.