Rafay
Platform fee plus a per-node unit fee, covering fleets, GPUs and multi-tenancy. Our kubernetes management platforms ranking puts Rafay at #7 of 13.
A vendor never writes its own entry here, Rafay included, although paid placement exists; see the disclosure for how links are marked.
Visit Rafay See it in the kubernetes management platforms ranking
At a glance
| Headquarters | United States |
|---|---|
| Pricing model | Platform fee plus per-node or per-GPU unit fee, billed on running average; support fee on top |
| Free plan | No free tier on record |
| API | Unknown from our sources |
| Deployment | Cloud or hosted |
| Best for | Platform teams offering governed self-service clusters and GPU capacity internally |
What is Rafay?
Rafay Systems, Inc. sells the Rafay Platform from a Sunnyvale, California notice address. The homepage describes six offerings: AI Token Factory for token-metered, API-driven inference on multi-tenant GPU infrastructure, AI Factory for a self-service governed AI platform, Rafay Observability with synthetic monitoring and AI-assisted root-cause analysis, enterprise-wide multi-tenancy, SLURM-as-a-Service, and single-click Kubernetes cluster launch with approval trails.
Fleet management spans public cloud, private cloud and on-premises, with hosting named on EKS, AKS, GKE and OKE, and operation in in-region, private or air-gapped environments. Pricing is a platform fee plus a unit fee per GPU, CPU or node, billed on the running average of units in use. The pages we read show an AICPA SOC compliant badge and make no ownership statement.
Rafay key features
- AI Token Factory. Token-metered, API-driven inference delivered on multi-tenant GPU infrastructure, aimed at operators who sell or share inference capacity.
- AI Factory. A self-service, governed AI platform for internal teams, so data scientists can request resources within limits set by the platform owner.
- Rafay Observability. Synthetic monitoring and AI-assisted root-cause analysis, with workflows that include human approvals before changes are applied.
- Multi-tenancy. Hard and soft multi-tenancy enforced across customers, teams, projects and regulated environments, which is the core of the platform's governance story.
- SLURM-as-a-Service. Offered alongside Kubernetes, for organisations whose researchers still schedule work through Slurm rather than containers.
- Cluster launch with approvals. Single-click Kubernetes cluster launch with approval trails, so provisioning is quick and still leaves a record of who approved what.
- Fleet management. Kubernetes fleet management across public cloud, private cloud and on-premises, with hosting on EKS, AKS, GKE and OKE.
Who Rafay suits
- GPU cloud or neocloud operator. Token metering on multi-tenant GPU infrastructure and a per GPU unit fee line up usage and billing for providers selling inference to external customers.
- Enterprise AI platform team. AI Factory offers a governed self-service layer, with multi-tenancy across teams and projects, so internal users get GPUs without direct cluster access.
- Regulated or sovereign deployment. In-region, private and air-gapped operation, together with the SOC badge and hard multi-tenancy, suit teams with data location and separation requirements.
Getting started with Rafay
The pricing page sets out no free tier or trial terms, so a start means speaking to the vendor. The terms of service show two agreements, one for a self-hosted product and one for SaaS, so decide early which model you need.
The homepage describes single-click Kubernetes cluster launch with approval trails and names EKS, AKS, GKE and OKE as hosting targets, plus integrations with Slack, Microsoft Teams, PagerDuty, Jira and ServiceNow for approvals and alerts. The pages we read give no step-by-step onboarding procedure or timeline. Ask what a pilot involves, how long it runs and which units are counted during it.
The case for and against Rafay
- Rafay Systems, Inc., with a Sunnyvale notice address, charges a platform fee plus a unit fee per node, GPU or CPU, averaged over what runs.
- Unit-based fee is among the traits we single out for it.
- Multi-tenancy, one of the three traits the catalogue lists for Rafay.
- Marked in our records as air-gapped.
- Suits: Platform teams offering governed self-service clusters and GPU capacity internally.
- The pricing page lists no plan names or amounts, and does not set out trial terms.
- Enterprise Support costs 20 percent of the cluster or node subscription on top.
- Less suited to: Buyers wanting a fixed monthly figure for a small estate.
Verdict: Rafay
Rafay Systems, Inc., with a Sunnyvale notice address, charges a platform fee plus a unit fee per node, GPU or CPU, averaged over what runs. Standard support in US business hours is included; Enterprise Support adds 20 percent of the subscription.
The homepage now sells an AI Token Factory and SLURM-as-a-Service beside cluster management; the pricing page names no plans, amounts or trial.
In our kubernetes management platforms ranking Rafay stands at #7 of 13: Platform9 sits directly above it and Northflank directly below. Challengers hold places 1 to 3 where a category has them; the number here is a place in the order, not a mark out of ten.
Home market: the United States, one of 7 North America bases among the 13 products here.
Rafay pros and cons in practice
- Unit-based pricing follows the running average of units in use.
- Standard Support is included during U.S. business hours.
- Hosting on EKS, AKS, GKE and OKE, plus air-gapped operation.
- Approval trails and human approvals in workflows.
- No plan names or amounts are listed, so a bill cannot be estimated from public pages.
- Enterprise Support adds 20 percent of the cluster or node subscription.
- Standard Support hours are U.S. business hours only.
- SSO, RBAC and encryption details are not given on the pages we read.
Rafay in practice
Rafay launches Kubernetes clusters in one click with approval trails and runs fleets across EKS, AKS, GKE, OKE, private cloud and on-prem sites, including air-gapped ones.
Hard and soft multi-tenancy covers customers, teams and projects.
Integrations include Slack, Microsoft Teams, PagerDuty, Jira, ServiceNow and Datadog.
Separate terms of service cover the self-hosted and the SaaS product.
Paying for Rafay
Charging model: Platform fee plus per-node or per-GPU unit fee, billed on running average; support fee on top.
The pricing page lists no plan names or amounts. The model is a platform fee plus a unit fee per GPU or CPU, or per node, billed on the running average of units in use. Standard Support during U.S. business hours, 8AM to 8PM EST, is included. Enterprise Support costs an additional fee equal to 20 percent of the cluster or node subscription. Volume discounts exist, without figures, and EDU and GOV discounts require contacting sales. Free tier and trial terms are not set out on the page, and no currency is named. Because billing follows a running average, the bill tracks usage rather than peak capacity, which is worth modelling against your own GPU utilisation pattern.
Source: the pricing material on rafay.co, as read for Best Kubernetes Management Platforms in 2026 and last reviewed on 15 November 2026.
Rafay: data and security
The terms of service name Rafay Systems, Inc. with a Sunnyvale, California notice address and state that the agreement is governed by California law and controlling U.S. federal law; the country of incorporation is not stated. The privacy policy says data may be transferred to offices and servers located in the United States.
The homepage shows an AICPA SOC compliant badge, a Kubernetes Certified Service Provider badge and Nvidia Cloud Validated status. It states that hard and soft multi-tenancy is enforced, and offers air-gapped operation. SSO, RBAC and encryption details are not given on the pages we read, so request them.
Fact sheet: Rafay
| Product | Rafay |
|---|---|
| Vendor website | rafay.co |
| Rank there | #7 of 13 |
| Alternatives page | No dedicated page |
| Head-to-head pages | None yet |
| Headquarters | United States (North America) |
| Catalogue tags | Unit-based fee, Multi-tenancy, Air-gapped |
| Primary guide | Best Kubernetes Management Platforms in 2026 |
| Gaps in our records | company size and funding |
| Price visibility | Model recorded, no figure |
Integrations and ecosystem
Integrations the vendor names (9):
- Slack
- Microsoft Teams
- PagerDuty
- Jira
- ServiceNow
- Datadog
- Elasticsearch
- Loki
- NVIDIA
Also in our records:
- Integrations include Slack, Microsoft Teams, PagerDuty, Jira, ServiceNow and Datadog.
Also ranked in
| Guide | Position |
|---|---|
| Best Kubernetes Management Platforms in 2026 | #7 of 13 |
Close alternatives to Rafay
These are the nearest neighbours of Rafay in the same list, picked on shared traits and price model.
- Kubermatic (#4 of 13): Hamburg company's multi-cluster platform for multi-cloud and edge estates. Watch out: the homepage shows no prices, plans or trial terms, only Try It Now and Book a Demo.
- Platform9 (#6 of 13): Aimed at: Teams leaving a virtualisation vendor who want VMs and containers on one plane. (United States)
- Civo Kubernetes (#9 of 13): Managed Kubernetes with a free control plane and hourly worker billing.
- Nirmata (#10 of 13): Not for: teams looking for a cluster provisioning or upgrade tool. Kyverno's creators selling policy control hubs and per-developer agent governance. (United States)
- Headlamp (#11 of 13): Not for: buyers who need vendor support or a service contract. Extensible Kubernetes interface as desktop app or in-cluster web service.
Questions about Rafay
9 answersWhich country is Rafay from?
Rafay contracts from the United States, so it sits in North America in our grouping. Country here means the contracting company, which can differ from where servers run.
Is Rafay the right fit for me?
Platform teams offering governed self-service clusters and GPU capacity internally. It does not fit as well for buyers wanting a fixed monthly figure for a small estate.
How much does Rafay cost?
Platform fee plus per-node or per-GPU unit fee, billed on running average; support fee on top is how we record the pricing of Rafay. On free access, a paid plan is required, per our records.
What are the main alternatives to Rafay?
The closest products in the same guide are Kubermatic, Platform9, Civo Kubernetes, Nirmata and Headlamp. All of them are reviewed in the same ranking.
How does Rafay rank for kubernetes management platforms?
In our kubernetes management platforms list Rafay stands at place 7 of 13.
How is Rafay priced?
A platform fee plus a unit fee per GPU, CPU or node, billed on the running average of units in use. The pricing page lists no amounts.
What does support cost?
Standard Support during U.S. business hours is included. Enterprise Support costs an additional fee equal to 20 percent of the cluster or node subscription.
Can it be self-hosted?
The terms of service cover a self-hosted product and a SaaS product, each under its own agreement, and the homepage mentions air-gapped operation.
Where is the company based?
Rafay Systems, Inc. gives a notice address in Sunnyvale, California. The country of incorporation is not stated, and the privacy policy mentions transfers to the United States.
Rafay: sources
Read from the vendor’s own pages on 14 November 2026.
- rafay.co/pricing: Pricing model, support fees, discounts, absence of plan amounts and trial terms
- rafay.co: Offerings, hosting targets, integrations, badges, multi-tenancy
- rafay.co/terms-of-service: Self-hosted and SaaS agreements, Sunnyvale notice address, California governing law
- rafay.co/privacy-policy: Operator, US data transfers, no ownership statement
How we rank
Four tests decide a position: how soon the product is useful, what the bill really becomes, how data leaves and who it suits. On those tests Rafay lands at place 7 of 13. Read the editorial process for the method and where paid visibility fits in.