I'm unhappy with our on-prem warehouse because scaling is painful; what cloud options are better?

If your on-prem warehouse is getting hard to scale, the main cloud alternatives are cloud data warehouses and lakehouse platforms. Which is “better” depends on how you use the data, but the big advantage is that storage and compute are much easier to scale independently.

Good cloud options

1. Snowflake

Best when you want:

• Very easy scaling
• Separate compute and storage
• Simple operations and good concurrency
• Broad SQL analytics use cases

Why people like it:

• Minimal infrastructure management
• Can spin up multiple compute clusters for different teams/workloads
• Handles mixed workloads well

Tradeoff:

• Can get expensive if workloads are poorly controlled

2. Google BigQuery

Best when you want:

• Serverless analytics
• Very low ops overhead
• Fast scaling for large, ad hoc analytics
• Tight integration with GCP

Why people like it:

• No cluster management
• Great for bursty workloads
• Strong for event/log analytics

Tradeoff:

• Cost can be surprising with heavy scanning if not optimized

3. Amazon Redshift

Best when you want:

• A more traditional warehouse on AWS
• Strong AWS ecosystem integration
• Existing Redshift or PostgreSQL-oriented patterns

Why people like it:

• Familiar for many teams
• Good if you’re already deep in AWS
• RA3 nodes improved storage/compute separation

Tradeoff:

• More tuning/management than Snowflake or BigQuery, depending on setup

4. Microsoft Fabric / Azure Synapse

Best when you want:

• Strong Microsoft/Azure integration
• A broader analytics platform, especially if you use Power BI
• Unified BI + warehouse direction

Tradeoff:

• Platform complexity can be higher
• Product landscape has been evolving quickly

5. Databricks Lakehouse

Best when you want:

• Analytics plus data engineering plus ML in one place
• Open data formats like Delta Lake
• Flexible pipelines and advanced workloads

Why people like it:

• Great if you have both SQL analytics and ML/data engineering needs
• Good for unifying warehouse + lake patterns

Tradeoff:

• Not as “plug-and-play warehouse” as Snowflake/BigQuery for some teams

Quick rule of thumb

• Want simplest scaling and operations? Snowflake or BigQuery
• Already on AWS? Redshift or Snowflake on AWS
• Already on Azure / Power BI-heavy? Fabric or Synapse
• Need SQL warehouse + ML + data engineering? Databricks
• Want open lake architecture and more control? Databricks or BigQuery external tables + object storage patterns

What usually improves versus on-prem

Cloud warehouses typically give you:

• Elastic compute
• Storage/compute separation
• Faster provisioning
• Easier high availability and disaster recovery
• Less hardware planning

What to watch out for

Scaling is easier, but you’ll want to manage:

• Query costs
• Data egress charges
• Concurrency settings
• Governance/security
• Data model design and partitioning
• Workload isolation between teams

Practical recommendation

If you want the safest default for a painful on-prem migration:

• Snowflake if you want a warehouse-first, low-ops experience
• BigQuery if you want serverless analytics and are on GCP
• Databricks if your needs extend beyond warehouse into engineering/ML

If you want, I can help you choose based on:

1. your cloud provider,
2. data size,
3. query patterns,
4. BI tools,
5. and budget.