Long-running virtual machines with root SSH, persistent disks, snapshots and private networking. CPU or GPU-backed, on the decentralized TRD network, billed by the second.
Virtual servers designed for modern applications. Unlimited bandwidth, automatic scaling, and transparent pricing — no surprise bills ever.
The same VMs you launch from the Compute dashboard — pick a GPU, choose a cluster size, configure, deploy. Distinct from short-lived instances: these run until you stop them.
Small LLM (7–13B, RAG), Large LLM (70B+, training), image gen (SDXL, Flux), 3D render (Blender, Octane) — the catalog filters to GPUs that fit.
Single card to ×4 consumer/pro clusters and ×8 HGX/SXM datacenter nodes with NVLink — scale to the job.
Template images (PyTorch 2.8 / CUDA 12.4, Ubuntu 22.04, ComfyUI), container disk + data volume, root SSH key — your machine, your setup.
Set an auto-stop budget cap per VM. Per-second metering with a signed receipt attached to every charge.
Join thousands of developers and businesses that rely on TRD Network for rock-solid, affordable web hosting.
Start Hosting → Back to Home