Dremio review

Dremio is a lakehouse platform for querying enterprise data and supporting analytics and AI. It combines a query engine with a semantic layer, an open catalog and an AI agent. The semantic layer is meant to define metrics, business context and relationships so analytics tools and agents can work with shared definitions.

The platform can query structured, semi-structured and unstructured data across sources without moving it, and it processes Apache Iceberg data natively. Dremio says it connects with Claude, ChatGPT, Gemini and coding agents, as well as its own built-in agent. Its optimization features include Reflections, caching and on-demand scaling; access controls include role-based and fine-grained permissions.

There are two deployment paths: Dremio Cloud, which manages the infrastructure, and Dremio Enterprise, which customers run themselves on Kubernetes, on-premises, AWS, Azure or GCP. That makes Dremio a platform choice as much as a query-engine choice.

Dremio’s clearest fit is teams that want to query Iceberg and other data where it already lives, rather than first moving it into a single warehouse. Native Iceberg processing and federation across data types are central to the product description. The semantic layer also gives teams a place to define business meaning for both agents and analytics, rather than treating agent connectivity as the whole AI strategy.

The deployment choice is useful if managed cloud is not the right fit: Enterprise supports Kubernetes, on-premises and the major cloud providers, while Cloud handles infrastructure, upgrades and scaling. The pricing page lists Cloud at $0.20 per DCU pay-as-you-go, plus a $400 credit for a 30-day trial. AWS is supported; Azure is listed as coming soon.

• Reflections, caching and on-demand scaling are designed to optimize query performance.
• Dremio lists connections for Claude, ChatGPT, Gemini and coding agents.
• Fine-grained and role-based controls are intended to carry permissions from clients to data sources.

These are relevant capabilities, but they are not proof that Dremio will be faster or cheaper for a particular workload. Reddit commenters discussing Dremio and Starburst say performance depends on the workload and recommend evaluating with real data.

The main buying drawback is price predictability. Cloud is billed at $0.20 per DCU, but the pricing information does not publish a monthly starting price, a typical DCU requirement or a spending cap. Annual committed contract prices are also not published. Enterprise pricing is consumption-based and requires contacting Sales; trial access is also through Sales. Budgeting therefore depends on getting a workload-specific estimate, not comparing a public monthly plan price.

Deployment control comes with operating work. The pricing page says Enterprise uses self-managed infrastructure with manual upgrades and scaling. One Reddit commenter describes Dremio as expensive to host and scale yourself, while another argues its enterprise catalog and governance features are basic. Those are individual opinions, not a definitive feature assessment, but they point to questions worth checking in a procurement review.

There may also be a portability trade-off. In a Dremio-versus-Starburst thread, some commenters favor Trino for its wider industry use and migration optionality. Dremio’s positioning has also confused some users, according to a Reddit discussion. Shortlist it for open lakehouse workloads, but verify operating effort, required DCUs and the specific governance features you need before switching.

Dremio Cloud is listed at $0.20 per DCU, billed pay-as-you-go. The pricing details also list a $400 trial credit for 30 days; they do not publish a monthly starting price or an estimate of typical consumption. Your bill depends on DCU use, so confirm expected workloads and costs before committing. Annual contracts are available, but their prices are not published.

Dremio Enterprise is also consumption-based, with pricing available from Sales rather than on a public rate card. It is self-managed and supports Kubernetes, on-premises, AWS, Azure, or GCP; upgrades and scaling are manual. The pricing details say to contact Sales for trial access.

There is no published price jump to compare: neither plan has a fixed monthly or annual price, and no add-ons are listed. Cloud includes Dremio-managed infrastructure and automatic upgrades and scaling; AWS is supported, while Azure is listed as coming soon. Compare the operational work as well as consumption before choosing.

Switching effort will depend on whether you choose managed Cloud or self-managed Enterprise. Cloud includes Dremio-managed infrastructure and automatic upgrades and scaling. Enterprise supports Kubernetes, on-premises, AWS, Azure, or GCP, but upgrades and scaling are manual. That makes deployment ownership a concrete decision, not just a pricing detail.

Dremio supports Apache Iceberg natively and can query data across sources without moving it, according to the published product description. Those capabilities may suit teams building around Iceberg or federation, but they do not establish how easy it is to migrate existing queries, permissions, or semantic definitions. The published sources do not document migration tools, data export procedures, or an exit process, so verify those with Dremio before switching.

We can’t summarize why customers leave based on Reddit: no Reddit posts or comments were provided. The clearest stated reasons to hesitate are consumption-based pricing, unpublished contract prices, and the operational demands of self-management.