Capacity is reviewed and held for the confirmed time instead of relying on a spot queue.
Why reserved
A higher-confidence
way to buy compute.
Topology, memory, storage, network, and access requirements are agreed before funding.
Bitcoin is live; card, wire, invoice, and purchase-order workflows are staged for business onboarding.
A real infrastructure review catches interconnect and data-movement constraints before a run starts.
GPU catalog
Right-sized for
every workload.
Focused current-generation GPU packages with CPU, RAM, storage allowance, advance booking, and business support.
Monthly package pricing · August 28, 2026
Public prices cover continuously reserved GPUs billed month to month. Availability and final node specifications are confirmed before funding. Taxes, if applicable, are additional.
Monthly packages
Dedicated capacity.
Month to month.
Choose one GPU family and reserve 1, 2, 4, or 8 GPUs continuously. The subscription renews monthly until your company cancels.
Month-to-month reserved capacity. Cancel before the next renewal date to stop the following month; the current paid month remains available through its billing period.
Dedicated GPU capacity
Request your capacity.
Choose the GPU, quantity, subregion, and preferred start date. Vodania confirms availability, then configures the recommended environment with your team.
3-month minimum commitmentMonthly billingConfiguration after confirmation
01
02
1GPU
03
04
1 × H100 · CanadaCalculating monthly price…3-month initial commitment · monthly billing · continues monthly until cancelled
Reservation model
Compute without
the capacity lottery.
Submit one clear capacity request. Vodania confirms the hardware and monthly price before payment or technical configuration.
Choose
Select one GPU family, quantity, subregion, and start date in the capacity navigator.
Confirm
Vodania verifies and holds the requested continuous monthly capacity.
Configure
Use the recommended environment or customize drivers, storage, container, and network after confirmation.
Run
Connect by SSH. Capacity renews monthly until your company cancels.
Reservation builder
Price the run.
Hold the window.
Build an indicative quote now. No payment is taken until capacity is confirmed.
Built for B2B workloads
The essentials,
already considered.
Dedicated capacity
Your reserved GPU is assigned to your workload for the confirmed window—no spot interruptions or queue roulette.
Root-level control
SSH access, custom containers, secure environment variables, and familiar CUDA development workflows.
Persistent storage
Keep models, datasets, and checkpoints available between reserved sessions on separately billed NVMe volumes.
Predictable monthly billing
One package covers continuously reserved capacity and renews until the company cancels.
Workload images
Start with Ubuntu, CUDA, PyTorch, TensorFlow, Jupyter, vLLM, or bring your own container image.
Human capacity desk
Complex topology, extended windows, and multi-node requirements are reviewed before funds are committed.
API foundation
Automate the capacity conversation.
Vodania now publishes versioned GPU catalog, capacity, service-status, and reservation endpoints. Instance lifecycle and API-key authentication remain on the platform roadmap.
GET https://vodania.com/v1/gpusQuestions
Before you reserve.
Clear answers for procurement, engineering, and finance teams.
How long does a reservation run?+
Every reservation has a three-month initial commitment, billed monthly. After month three, it continues month to month until your company cancels before the next renewal date.
When does billing begin?+
Billing begins at the confirmed reservation start time and follows the reserved hourly window. Your quote is indicative until Vodania confirms the exact capacity and node specification.
Can we reserve multiple GPUs?+
Yes. The builder supports 1, 2, 4, or 8 GPUs. Larger clusters, high-bandwidth GPU interconnects, and long-term capacity are handled by the capacity desk.
What operating systems and images are available?+
Launch options include Ubuntu 22.04 or 24.04 with CUDA, PyTorch and TensorFlow Jupyter images, a vLLM stack, and custom OCI-compatible container images.
How can our company fund a package?+
Bitcoin funding is available through the secure City of Hats payment portal. Approved companies may request invoice, purchase-order, or bank-wire handling. Stripe card checkout is prepared and will appear after the business account is connected.
How does cancellation work?+
Cancel before the next monthly renewal date to stop the following billing month. The currently paid reservation remains available through the end of its billing period.
What network and multi-GPU details are confirmed?+
The confirmation states GPU topology, supported interconnect, CPU and RAM profile, local and persistent storage, public/private networking, bandwidth, exposed ports, transfer treatment, facility region, and access method.
Reservations are open
Know when your
compute will be ready.
Configure a workload, hold a capacity window, and fund only after the details are confirmed.
Build your reservation