A little more control. A lot more flexibility.
Starting today, Compute with Hivenet gives you full control over your running instances. You can stop any of them—midway through a project, during a long training job, or while you're taking a break—and start them again later without losing your setup.
Your files, packages, and configuration all stay exactly as you left them. No manual saves. No starting from scratch.
Compute just got a memory.
Pay nothing while stopped
As a launch promotion, stopped instances cost 0€ per hour. Pause your environment for a few hours (or a few weeks… we won’t judge) without paying a cent. Just start it back up when you're ready. This offer won’t last forever, but for now, it’s our way of saying thanks for building with us early.
Who’s this for?
Honestly, anyone who hates redoing work.
Model training with checkpoints
Run training jobs in stages. Save your progress. Pick it up later.
It’s perfect for PyTorch, TensorFlow, or anything with .ckpt and .pt files.
Academic and research projects
Need GPU for your thesis or semester-long experiment? Set it up once, stop it over the holidays, and resume without starting from scratch.
Persistent dev environments
CUDA. PyTorch. System packages. Drivers. That one weird dependency.
Build your dream stack, stop it when you're done, and come back later to exactly the same setup.
It’s live now
You’ll see new Stop and Start buttons next to all your running instances in the dashboard. Try it out, and if you’ve got feedback or ideas, we’d love to hear from you—message us on Discord or email support@hivenet.com.
More improvements are on the way. But for now, this one’s all about control.
