Mega ducklings: For Demanding Workloads at a Larger Scale

Mega ducklings are designed for when your workloads have outgrown Jumbo and you need more power — not eventually, but right now.

"An extremely large instance for when you need complex transformations done quickly."

Use a Mega when:

• Your queries are too complex or your data volume is too high for Jumbo to handle — especially in crunch time
• You’re running a weekly job that rebuilds all your tables, and it has to run in minutes, not hours
• One customer has 10x the data of everyone else, and they still expect subsecond response times

Under the hood, Mega unlocks more in-memory execution, handles larger joins and aggregations without spilling, whether you’re reading from your MotherDuck storage, Parquet files or your shiny new DuckLake.

Giga ducklings: When Nothing Else Will Work

Giga ducklings are our largest instance sizes, purpose-built for the toughest of transformations.

"Largest instances enable the toughest of transformations to run faster."

Request a Giga when:

• Your data workload is so complex or so massive that nothing else will work
• You’re running a one-time job to restate revenue for the last 10 years — and it needs to be correct and fast
• You need a growth path beyond Mega, because your data volume and complexity just grew 10x

Giga gives DuckDB an environment with maximum compute and memory — ideal for very complex joins, deeply nested CTEs, and long-range analytical backfills. It’s not for every job — but when you need it, you really need it.

Scaling up isn't the only way

Scaling up to larger instance sizes (ducklings) is only one of the many ways MotherDuck scales data warehousing workloads.

Most data warehouses are built as monoliths, where every user in the organization shares the same data warehouse compute resources. These monoliths often begin to crack under high concurrency. At the core of MotherDuck's architecture is per-user tenancy, in which each user (or customer, in the case of customer-facing analytics) gets their own duckling that's configurable in size. So you might use one of the new Mega instances for some complicated transformations in your data pipelines, but still rely upon Standard instances to serve most of your users. Each instance is provisioned on demand and managed for you.

There may be cases where per-user tenancy isn't as natural. For example, business intelligence (BI) tools typically share a single database connection but then may have dozens of users running queries at the same time. This would ordinarily break the "one-user-per-duckling" pattern.

MotherDuck’s read scaling is designed for these types of cases – providing an extra boost in compute through horizontal scaling and maintaining the pattern of “one-user-per-duckling!”

Available on the Business Plan

These new duckling sizes are available on the instance plan. Megas are completely self-serve. If you want access to Gigas, please quack with us about what you're building.