Storage
Your data, beside the GPUs.
Fast storage in the same datacentre as the GPUs, with S3-compatible object and file access. Datasets, checkpoints and models are kept safe, in Nepal.
Coming soonEnterprise early access is open
How it works
Three steps.
- 01
Bring your data
Upload once, with the S3 tools you already use.
- 02
Read it from every job
Notebooks, pipelines and endpoints work from the same copy.
- 03
Keep what matters
Checkpoints and models stay beside the GPUs, in Nepal.
What you get
Storage, in detail.
Fast, beside the GPUs
Storage in the same datacentre as the GPUs, so jobs read data without a long trip.
S3-compatible
Object storage that works with the S3 tools and libraries you already have, plus file access.
Safe with us
Isolated from other customers, and never copied abroad.
Under the hood
The key numbers.
- Location
- Kathmandu Valley, Nepal
- Location
- Kathmandu Valley, Nepal
- GPU
- NVIDIA H200 NVL
- GPU memory
- 141 GB
- Memory bandwidth
- 4.8 TB/s
- Currency
- Nepalese rupees (NPR)
- Ways to pay
- eSewa, Khalti, bank transfer or invoice
Billing
By capacity, in NPR
Commitment: none. Rates are being finalised; every request gets a firm quote in rupees.
How billing worksQuestions
About storage.
Can we use our existing S3 tools?
Yes. The object storage is S3-compatible, so the usual command-line tools and SDKs work.
Where is the data kept?
In our datacentre in Nepal, beside the GPUs.
Start with storage.
Tell us what you want to run. It takes about a minute.