Pricing & Billing
1. Pricing Overview
Heata offers a transparent, consumption-based pricing model. There are 3 core elements to our pricing:
-
Base Compute: users are charged for the compute resources (CPU, GPU, Memory etc.) provisioned or used during job execution.
-
Platform / Application: free for basic VMs or containerised services, then charged per minute for more advanced services.
-
Data Ingress / Egress: first 500GB / month free, thereafter charges apply for cost recovery.
2. Compute Rates
All prices listed in this document are the maximum prices for the heata platform. Lower rates may be available through reserved discounts.
2.1 Base Compute Pricing
Every instance starts with a base compute layer. This covers the physical hardware required to run a workload on our network.
Managed Compute
Production-grade reliability without the public cloud premium. Fully managed service with control plane + ephemeral worker nodes. SLA guaranteeed.
| Component | Specification | Unit | Unit Price |
|---|---|---|---|
| Base vCPU | Standard x86 Core | / core hour | £0.017 (1.70p) |
| RAM | High-Speed DDR5 | / GB hour | £0.0025 (0.25p) |
| Ephemeral Storage | Local SSD-backed | / GB hour | £0.00008 (0.008p) |
Spot Compute
Self managed service with better-than-spot persistence. No SLA.
| Component | Specification | Unit | Unit Price |
|---|---|---|---|
| Base vCPU | Standard x86 Core | / core hour | £0.010 (1.00p) |
| RAM | High-Speed DDR5 | / GB hour | £0.0018 (0.18p) |
| Ephemeral Storage | Local SSD-backed | / GB hour | £0.00008 (0.008p) |
2.2 Accelerator Pricing
For workloads requiring GPUs, such as AI InferenceI, GPU accelerators can be attached to any base compute instance. These are billed as an additional hourly surcharge on top of the compute they are attached to..
| Hardware Tier | Accelerator Profile (VRAM) | Example Models | Base Specs (vCPU / RAM / Ephemeral NVMe) | Suitable Workloads | Unit | Price (Hourly) |
|---|---|---|---|---|---|---|
| Tier 1: Entry | 16GB - 24GB | NVIDIA RTX 4090, RTX 5080, RTX A5000 | 4 vCPU / 16GB RAM / 100GB | Inference, small model tuning, development environments. | GPU-Hour | + £0.89 / hour |
| Tier 2: Mid | 24GB - 40GB | NVIDIA RTX 5090, A100 (40GB), RTX 6000 Ada | 8 vCPU / 32GB RAM / 200GB | Model fine-tuning, CAD/CAE, medium-scale data processing. | GPU-Hour | + £1.19 / hour |
2.3 Platform fees
Platform tiers define the software environment and management tools pre-installed on your compute. We distinguish between raw containerised environments and managed platforms. Fees are applied per core-hour to cover orchestration and specialised software licensing, and are charged on top of the base compute fees.
| Tier | Interfaces / Platforms | Platform Fee (per core hour) |
|---|---|---|
| Standard | Command Line Interface (CLI), REST API, Web Management Portal & Container Orchestration API (e.g. kubectl) | FREE |
| Data Science | Examples include Interactive Notebooks, Distributed Computing APIs (Ray, Spark) | 0.71p per CPU hour + 0.09p per GB RAM hour* |
* Example: an instance with 4 CPU and 8GB Ram would cost 4p per hour.
3. DevOps & CI/CD Services
heata provides hosted GitHub Actions runners for automated builds, testing, and deployment. These are ephemeral, isolated Linux environments that scale on demand. To mirror existing market offerings, these are billed differently to our other services. You pay per minute (rounded up to the nearest minute) based on the configuration of your runner.
| Runner Tier | Specs (vCPU / RAM)* | Price per Minute | GitHub Equivalent | Saving vs. GitHub |
|---|---|---|---|---|
| heata Slim | 1 vCPU / 5GB | £0.0011 | actions_linux_slim | \~25% |
| heata Standard | 2 vCPU / 8GB | £0.0034 | actions_linux | \~25% |
| heata Large | 8 vCPU / 32GB | £0.0120 | linux_8_core | \~25% |
* Custom configurations available on request.
4. Reserved Instance (RI) Discounts
Reserved Instances provide a lower hourly rate in exchange for a committed term of one or three years. These apply to the base compute rates only. This allows you to lock in your core hardware costs while maintaining flexibility in which platforms you use on top. Billing is handled via a fixed monthly invoice based on the reserved capacity.
How it works:
- Fixed price: a predictable monthly bill based on your reserved pool.
- Capacity Guarantee: your compute resources are "locked" and always available; they are never subject to market fluctuations.
| Commitment Term | Savings vs. On-Demand |
|---|---|
| 1-Year Reserved | 30% Off |
| 3-Year Reserved | 55% Off |
*Monthly cost is based on a standard 730-hour month. Pay monthly with no initial commitment fee.
5. Networking & Storage
Beyond compute, we charge for any persistent in-cloud storage (scratch space on worker nodes is free) and data transfer above a set threshold. Every account includes a baseline monthly data allowance to cover standard operational traffic.
| Service | Allowance / Rate | Note |
|---|---|---|
| Ingress | FREE | - |
| Egress | £20.00 per TB | 100GB free per month |
| Persistent Storage | £0.06 per GB / Month | SSD backed storage |
6. Support Packages
Support fees are calculated as a percentage of monthly consumption charges to scale with usage.
| Support Tier | Channels | Response (P1) | Support Hours | Engineering Support | Monthly Fee |
|---|---|---|---|---|---|
| Standard | Email, Ticketing | 8 hours | 9am-5pm on Business Days | None | Included |
| Priority | Email, Ticketing, Slack | 2 hours | 9am-5pm on Business Days | Engineering support for troubleshooting, configuration & optimisation | 5% of spend (min £500) |
| Enterprise | Email, Ticketing, Slack, Phone | 30 minutes | 24/7, 365 | Priority engineering support including integration & deployment guidance | 10% of spend (min £1,000) |
You can review our full SLA here.
7. Additional Information
-
VAT: all prices exclude VAT
-
Billing frequency: invoices are issued monthly in arrears unless otherwise agreed
-
Contact: email accounts@heata.co with any questions