Serverless GPUs

Run machine learning inference at scale

Priced by the hour, billed by the minute — so you only pay for what you use.

Launch instance

HOW IT WORKS

  • Launch
    ready

    Always available, pre-configured NVIDIA GPU clusters and ML frameworks

  • Safe
    &     secure

    Complete isolation via a separate control plane to keep your data private

  • Easily
    automates

    Turnkey autoscaling, fully managed and load balanced

  • Cost
    saving

    Priced by the hour, billed by the minute — so you pay for only what you use

WHY SERVERLESS KUBERNETES

Pre-configured by experts to streamline ambitious builds

  • SPEED
    5s
    Or less to start-up and go
  • SCALE
    1000+
    GPUs to build with
Discover our latest

serverless clusters

Get started

  • NVIDIA
    H100 SXM

    80GB vRAM | 3.3TB/s Bandwidth

    For top-tier performance

  • NVIDIA
    H100 PCIe

    80GB vRAM | 2TB/s Bandwidth

    Perfect for large-scale AI and HPC workloads

  • NVIDIA
    L40s

    48GB vRAM | 864GB/s Bandwidth

    For cost-effective and performant inference

  • NVIDIA
    L4

    24GB VRAM

    For cost-efficient AI compute

Why developers love Ori

Ori’s GPU costs have been very competitive and customer support has been superior to many other cloud providers we’ve tried.

Vinay Maniam
Founding Engineer, nCompass

Serverless Kubernetes

GPU resources

RESOURCEPRICE ($ PER MINUTE)PRICE ($ PER HOUR)
NVIDIA H200 141GB0.07875/min4.73/hr
NVIDIA H100 80GB SXM0.06525/min3.92/hr
NVIDIA H100 80GB PCIe0.06525/min3.92/hr
NVIDIA L40S 48GB0.03488/min2.09/hr
NVIDIA L4 24GB0.019044/min1.14/hr

Other resources

RESOURCEPRICE ($ PER MINUTE)PRICE ($ PER HOUR)
Memory (MB)0.00000008/min0.000005/hr
CPU (1/100)0.00000167/min0.0001/hr
Load Balancer0.0003015/min0.01809/hr
CUSTOM PRICING

Need a custom cloud, or pricing for a large-scale project? Let's talk.

Talk to an expert

Chart your own
AI reality

Launch Now