Snowflake vs Redshift

Quick verdict

Snowflake is a managed data platform that runs on AWS, Microsoft Azure and Google Cloud. Its documentation describes three layers: compressed columnar storage organised into micro-partitions, compute in virtual warehouses that you create, size, suspend and resume, and cloud services for metadata, security and optimisation. A Snowflake account on AWS runs in an AWS region but is a Snowflake service: you sign in, pay and manage access through Snowflake.

Amazon Redshift is AWS's data warehouse. A provisioned cluster has a leader node and compute nodes; AWS now recommends RG nodes (AWS Graviton based, with an integrated data lake query engine) or RA3 nodes, both of which use Redshift managed storage so compute and storage scale separately. DC2 nodes with local SSD storage are listed as previous generation on the pricing page. Redshift Serverless drops clusters altogether: capacity is measured in Redshift Processing Units (RPUs, 16 GB of memory each) and scales automatically from a base capacity you set.

Side by side

Key differences

Operating model: separate warehouses against clusters or serverless workgroups

Snowflake makes compute a lightweight object: you can create a warehouse per team, per tool or per job, each with its own size and auto-suspend, and all of them read the same stored data. Resizing a warehouse or adding clusters does not move data.

Redshift offers two models. A provisioned cluster is sized by node type and count; with RG and RA3 you size compute for the work you process and pay separately for managed storage, which offloads to Amazon S3 as it grows. Clusters can be paused, in which case AWS documents that you pay only for backup storage. Redshift Serverless organises compute into workgroups: you set a base capacity (4 RPUs, or 8 to 512 in steps of 8, and up to 1,024 in five large Regions), a maximum, and optionally a price-performance target, and AWS's AI-driven scaling adjusts RPUs between them. To isolate workloads on Redshift you typically run several workgroups or clusters and connect them with data sharing.

Scaling for concurrency

Snowflake's multi-cluster warehouses (Enterprise edition and above) add clusters of the same size when queries queue, under a Standard or Economy scaling policy, and remove them when the load drops. Each running cluster consumes credits.

Redshift's Concurrency Scaling sends eligible queries from a workload management (WLM) queue to extra clusters when the queue is full. AWS documents support for reads and for common writes (COPY, INSERT, DELETE, UPDATE, CTAS and VACUUM) on RG and RA3 nodes, with limitations: no queries on temporary tables, interleaved sort keys or tables with identity columns (for writes), and no Python or Lambda UDFs. Clusters earn one hour of free Concurrency Scaling credit every 24 hours, up to 30 hours. On Serverless, scaling is automatic within the base and maximum RPUs you set.

Semi-structured data: VARIANT against SUPER

Snowflake's VARIANT holds up to 128 MB of uncompressed data per value. Redshift's SUPER type holds up to 16 MB per value, and AWS documents that SUPER values over 1 MB can be ingested only from Parquet, JSON, TEXT or CSV. Redshift uses PartiQL: dot and bracket navigation, and unnesting by iterating over an array in the FROM clause. Navigation is lax by default, so invalid paths return NULL rather than an error, and AWS recommends enabling case-sensitive SUPER attribute names.

Snowflake:

SELECT e.payload:customer.id::STRING AS customer_id,
       i.value:sku::STRING          AS sku
FROM events e,
     LATERAL FLATTEN(INPUT => e.payload:items) i;

Amazon Redshift (PartiQL unnesting over a SUPER column):

SELECT e.payload.customer.id AS customer_id,
       i.sku                 AS sku
FROM events e, e.payload.items i;

SQL dialect differences

Redshift's SQL descends from PostgreSQL, and Snowflake's is its own ANSI-style dialect, but much everyday SQL looks alike. Both support QUALIFY; AWS notes that when QUALIFY follows the FROM clause directly, the table must have an alias.

-- Snowflake
SELECT customer_id, order_id, order_ts
FROM orders
QUALIFY ROW_NUMBER() OVER (PARTITION BY customer_id ORDER BY order_ts DESC) = 1;

-- Amazon Redshift (alias required here)
SELECT o.customer_id, o.order_id, o.order_ts
FROM orders o
QUALIFY ROW_NUMBER() OVER (PARTITION BY o.customer_id ORDER BY o.order_ts DESC) = 1;

Date arithmetic is close: both DATEDIFF functions take the unit first and return the second date minus the first.

-- Snowflake and Amazon Redshift
SELECT DATEDIFF(day, start_date, end_date) AS days_open,
       DATE_TRUNC('month', start_date)     AS start_month
FROM tickets;

The larger porting effort is usually in physical design and procedural code rather than SELECT syntax: Redshift has distribution and sort keys on tables, and AWS has announced that Redshift no longer supports Python UDFs after 30 June 2026, with enforcement in phases, so Python UDFs must be migrated (AWS published migration options alongside the announcement). Snowflake has no distribution keys; it uses micro-partitions with optional clustering keys.

Data sharing and the AWS ecosystem

Snowflake Secure Data Sharing gives consumers read-only access without copying data; consumers pay their own compute. Direct shares are limited to one region, and cross-region or cross-cloud sharing uses listings with auto-fulfilment or replication. Snowflake Marketplace spans AWS, Azure and Google Cloud accounts.

Redshift data sharing shares live data across provisioned clusters and Serverless workgroups, AWS accounts and Regions, again without copies, and can grant writes (INSERT, UPDATE) as well as reads, with transactional consistency for consumers. Datasets can be licensed through AWS Data Exchange. Redshift also sits directly in AWS: IAM roles, VPC networking, the Glue Data Catalog and S3 data lakes. On RG clusters and Serverless, AWS documents that data lake queries on Iceberg tables run on the warehouse's own compute with no separate charge, while RA3 and DC2 clusters use Redshift Spectrum, which is billed per TB scanned.

Pricing and licensing

Snowflake. According to Snowflake's pricing page in October 2026, compute is billed in credits and storage monthly on the average compressed volume, either on demand or as pre-paid capacity. The credit price depends on edition (Standard, Enterprise, Business Critical, Virtual Private Snowflake), cloud and region. Warehouses consume 1 credit per hour at X-Small up to 256 at 5X-Large, billed per second with a 60-second minimum each start. We could not read per-credit prices reliably, so none are quoted. The trial lasts 30 days or until its free usage balance runs out.

Amazon Redshift. The AWS Redshift pricing page, checked in October 2026, lists provisioned clusters billed per node-hour by node type (RG, RA3 and previous-generation DC2), with one- and three-year reserved nodes; RG and RA3 also pay for Redshift managed storage per GB-month. Redshift Serverless is billed in RPU-hours per second with a 60-second minimum; AWS's pricing examples use USD 0.375 per RPU-hour in US East (N. Virginia). Concurrency Scaling earns one free hour per 24 hours per cluster (up to 30 hours), and Redshift Spectrum is charged per TB scanned. New Redshift Serverless users get a USD 300 credit that expires after 90 days.

A Serverless workgroup at its default base of 128 RPUs consumes RPU-hours quickly while busy, so set the base capacity and the maximum RPU-hours limit deliberately. For both products, use the vendor's calculator with your own workload pattern.

Pricing checked on the vendors' official pages on 7 October 2026. Prices change; confirm before buying.

Where each one leads

Limitations

When to choose each

When neither is right

• You are on Google Cloud: compare Snowflake vs BigQuery and BigQuery vs Redshift.
• Your workload centres on Spark, notebooks and machine learning over a lakehouse: see Snowflake vs Databricks and Databricks vs Redshift.
• Your data still fits in a well-indexed PostgreSQL database: see Redshift vs PostgreSQL and Data warehouse vs database.
• You want a wider shortlist: see Snowflake alternatives.

Final recommendation

Frequently asked questions

Sources

• Snowflake pricing options
• Snowflake documentation: Key concepts and architecture
• Snowflake documentation: Overview of warehouses
• Snowflake documentation: Multi-cluster warehouses
• Snowflake documentation: Semi-structured data types
• Snowflake documentation: About Secure Data Sharing
• Amazon Redshift pricing
• Amazon Redshift Management Guide: Provisioned clusters and node types
• Amazon Redshift Management Guide: Serverless compute capacity
• Amazon Redshift Developer Guide: Concurrency scaling
• Amazon Redshift Developer Guide: SUPER type
• Amazon Redshift Developer Guide: Querying semi-structured data
• Amazon Redshift Developer Guide: Data sharing
• Amazon Redshift Developer Guide: Apache Iceberg tables
• Amazon Redshift Developer Guide: QUALIFY clause

Checked October 2026.

How we research comparisons: our editorial method.