Capabilities
Compute Rental
Short-term GPU capacity when buying hardware is not the right move — or not the right move yet.
What we help with
Some work is a burst: a training run, an evaluation, a bridge before owned hardware arrives. We quote short-term access to GPU capacity against the duration, isolation, and destination or region you specify.
This is not a self-serve cloud console and not a listed spot market. Every request is reviewed and quoted. We will be clear about what we can arrange and what we cannot.
If the rental is a bridge ahead of a hardware delivery, include that configuration in the same inquiry. The two quotes are easier to align when they are reviewed together.
Configurations we can quote
These are examples of configurations we can quote, not a catalogue of stock on hand. Availability is confirmed in the quotation. This site does not publish list prices.
-
Short-term H100- or H200-class capacity
Time-bounded access to H100-class or H200-class GPU capacity, quoted to the quantity and window you specify.
-
Training, fine-tuning, and evaluation windows
Capacity scoped to a defined job — a training run, a fine-tune, or an evaluation — rather than an open-ended reservation.
-
Bridge capacity ahead of a delivery
Short-term capacity quoted alongside a hardware configuration you are also asking us to supply, so the two can be reviewed together.
Practical considerations
-
Quoted access
This is arranged and quoted capacity, not a self-serve console and not a listed spot market with public prices.
-
Isolation and region
Isolation, data handling, and region often determine what can be arranged. State the constraints even if they are still being decided.
-
A defined window
A start, a duration, and a purpose are more useful than an open-ended request. We quote to the window you give us.
-
Bridge to owned hardware
Bridge capacity is easier to quote when the owned-hardware configuration is also in front of us.
What to include in an inquiry
- GPU class and how many GPUs you need, or a performance target we should quote against
- Duration, and the earliest start date that is still useful
- Isolation, tenancy, or data-handling constraints
- Preferred region or destination, if the work must run in a particular place
- Software or interconnect assumptions that would change what we can arrange
- Whether this is a standalone burst or a bridge ahead of a hardware purchase
- Who will operate the environment, if that is already decided
Who it is for
Labs and operators who need capacity on a defined schedule, without committing to a purchase for that window.
Related capabilities
Ready to specify a configuration?
Share the use case, the target build, quantity, destination country, and timeline. We will take it from the inquiry.