Databricks review

Databricks is a platform for data engineering, analytics, and AI, bringing analytical and operational workloads onto a shared, governed foundation. Teams can use it to build and run batch or streaming data pipelines, analyze data, build applications and AI agents, and run business intelligence.

Its product range includes Lakeflow for data pipelines, serverless data warehousing on open lake data, and Lakebase, which combines Postgres with lakehouse data, governance, and analytics. Genie lets business users ask questions about enterprise data in plain language. Databricks also offers a governance layer for data, models, dashboards, and agents.

That breadth comes with usage-based pricing rather than a published standard monthly price for paid plans. The pricing information lists Community Edition at $0 for learning Apache Spark and a 14-day free trial for Premium pay-as-you-go. Paid usage is metered by DBU or CU, and cloud infrastructure costs during trials are separate.

Databricks’ main advantage is that it covers several connected jobs in one platform. That can suit organizations that need to build data pipelines, analyze data, manage access, and develop AI applications on the same governed foundation.

• Data engineering: Lakeflow supports batch and streaming ETL pipelines. Premium pricing lists Data Engineering from $0.15/DBU, billed by the second.
• Analytics and warehousing: Serverless data warehousing runs on open lake data; listed pricing starts at $0.22/DBU for Data Warehousing.
• Governance: The platform’s governance layer covers data, models, dashboards, and agents, rather than limiting governance to tables alone.
• AI and natural-language access: Databricks supports production-ready agents grounded in organizational data, while Genie is designed to let business users ask questions in plain language. Genie includes 150 free monthly DBUs, then starts from $0.07/DBU, according to the pricing information.
• Low-cost learning: Community Edition costs $0 and includes a single-node cluster, though its micro-cluster and storage functionality are limited.

This breadth is most compelling when a team has a real need for several of these workloads. A Reddit commenter argued that Databricks can be more horsepower than needed for smaller data workloads; that is a useful prompt to check fit, not a universal verdict.

The first drawback is cost visibility. Premium has no published flat monthly price: Data Engineering starts at $0.15/DBU, Data Warehousing at $0.22/DBU, and Genie at $0.07/DBU after 150 free monthly DBUs. Those rates do not tell you the total bill. Pay-as-you-go usage is billed monthly, while annual or multi-year prices require custom commitments. Enhanced Security and Compliance adds 10% of list price to Databricks product spend in a workspace; Mission Critical pricing is not published. Trial cloud infrastructure costs are separate.

Users on Reddit report high bills and migrations to AWS-native services, though those accounts do not establish what Databricks will cost your team. Community Edition is also a learning option, not a full substitute: it has limited micro-cluster and storage functionality, and seat limits are not published. Standard is discontinued for new sign-ups on AWS, GCP, and Azure, with existing customers automatically upgraded to Premium.

Treat Genie as something to evaluate with your own data and questions. Reddit discussions include mixed reports, and some prominent criticism concerns Genie Code, a separate coding assistant, rather than Genie’s natural-language data analytics.

Databricks has no published starting price for its paid plans. Community Edition costs $0 per month, but it is for learning Apache Spark and comes with a single-node cluster and limited micro-cluster and storage functionality. Seat limits are not published.

Premium is metered, billed by the second, and priced per DBU or CU. The pricing details list Data Engineering from $0.15/DBU and Data Warehousing from $0.22/DBU. Genie includes 150 free monthly DBUs, then starts at $0.07/DBU. That is a usage bill, not a predictable flat subscription. Enterprise contracts have custom committed spend; annual or multi-year discounts are available, but prices are not published. Classic jobs are listed at around $0.20/DBU.

Budget for add-ons: Enhanced Security and Compliance adds 10% of list price to Databricks product spend in the workspace. Mission Critical pricing is not published. Cloud infrastructure costs during the 14-day trial are separate. Standard is discontinued for new sign-ups on AWS, GCP, and Azure.

Switching effort depends on what you are consolidating. Databricks covers data engineering, warehousing, analytics, governance, and AI, so a move can touch pipelines, dashboards, agents, and their governance—not just storage. The published material does not publish migration timelines, export steps, or a compatibility matrix. Treat the work and any data-transfer costs as items to confirm before committing.

Databricks describes its foundation as open and says its serverless warehouse runs on open lake data. That may matter when assessing portability, but the published sources do not specify what exports cleanly or what requires rebuilding. Map your current workloads and check each one against the target services before moving.

We can’t summarize why Reddit users leave: no Reddit posts or comments were included in the published material. What is clear from the pricing details is that usage-based billing can be a poor fit for teams that need flat, predictable spend. Compare your expected workload costs and contract terms before switching in either direction.