Sometimes, you just need more RAM. That’s especially true when you want to run memory-hungry enterprise applications like SAP’s HANA database or high-performance computing workloads. Until now, if you wanted the Google Compute Engine to run applications like that, your options topped out at 624GB of memory. Starting today, though, the company is going beyond that by introducing three new tiers on top of this that top out at 3844GB and 160 virtual compute cores.

These three new machine types, dubbed ‘n1-ultramem,’ join Google’s existing ‘n1-megamem’ machines. Unsurprisingly, this kind of performance comes at a price. Running the “low-end” machine with 40 cores and 938GB of RAM for a month will set you back just over $3,221. The high-end machine with 160 cores and 3844GB of RAM is yours for $12,885.1716 per month.

You can see the hourly prices below:

With these new machines, Google now matches the top-end memory-optimized options on the AWS platform, though Google offers slightly more compute power thanks to a higher number of cores and newer processors.

Unsurprisingly, Google notes that the canonical use case for this kind of machine is running SAP HANA. “If you’ve delayed moving to the cloud because you have not been able to find big enough instances for your SAP HANA implementation, take a look at Compute Engine,” the company writes today. “Now you don’t need to keep your database on-premises while your apps move to cloud.”

The new ultramem machines are now available in three Google Cloud regions (us-central1, us-east1 and europe-west1), with more to follow.