# Faster than an X-Large Snowflake warehouse

Source: https://www.hotdata.dev/compare/snowflake
Site index: https://www.hotdata.dev/llms.txt

Without paying for one.

## At a glance

| | Hotdata | Snowflake |
| --- | --- | --- |
| What it is | A Rust query engine built for low latency at high concurrency. | A cloud data warehouse. |
| Where data lives | Queries your sources in place, or Parquet in an instant database. | Loaded into Snowflake storage. |
| Deployment | In your VPC, run by us. Or fully managed. | Managed service. |
| Pricing | Compute runs only while queries run. | Per-warehouse credits. Bigger warehouses cost more. |
| Concurrency | 5× faster than Snowflake with 10 concurrent clients. | Latency rises past 10 streams. |
| Snapshots and copies | Snapshots, time travel, and forks. | Zero-copy clones. |
| Search | Vector, full-text, and geospatial in one SQL query. | Vector and keyword search functions. |
| SQL | Postgres-familiar HotSQL. PostgreSQL, DuckDB, Snowflake, and ClickHouse SQL run as written. | Snowflake SQL. |

[16×: Bigger Snowflake warehouse, still slower than Hotdata.](/blog/benchmarking-concurrency-on-hotdata-snowflake-and-bigquery#sizing-and-cost-implications)

## Performance

TPC-H SF1 · 22 queries · 10 concurrent clients.

| System | Hardware | p50 | p95 | Queries/s |
| --- | --- | --- | --- | --- |
| Hotdata | 4 vCPU · 16 GiB | 0.17 s | 0.38 s | 51.1 |
| Snowflake | ~8 vCPU · 16 GiB | 0.85 s | 1.74 s | 10.9 |

## When to choose Hotdata

- You need answers 5× faster, on half the vCPUs.
- Concurrency keeps climbing, and sizing up to an X-Large warehouse isn't the answer.
- Your apps need vector, full-text, and SQL filters in one query.
- You want it in your own VPC, run for you.
- Your Snowflake SQL should run as written.

## When Snowflake still fits

- Your analysts and BI tools need a governed warehouse as the source of truth.
- You run large batch transformations, and query concurrency is low.

## Using Hotdata with Snowflake

Connect Snowflake as a read-only source and join it with everything else. Snowflake SQL runs as written. See [Connecting a source](/docs/pull-data).

## Get started

The [quick start](/docs/quick-start) gets a table into Hotdata in a few minutes.