Hybr® Usage-Based & Metered Billing

Bill what customers actually use, across every platform you run

Usage-based billing with Hybr®: it collects usage from the platforms you already run, matches it to the right customer, prices it on their price list, and bills it on one invoice with the seats, licences and services they also buy.

One engine, every meterMicrosoft CSPAzureVMwareAzure LocalHyper-VKubernetes & GPUAI tokensOne consolidated invoice per customer.
Illustration with example figures.
The problem

Where usage-based billing actually breaks

Seats are easy to bill. Consumption is not. Once you sell Azure, virtual machines, Kubernetes capacity or AI tokens, the question becomes: who used what, at which rate, in which currency, and can we prove it? Most teams answer that with exports and spreadsheets. It works for a handful of customers and quietly loses margin after that.

Usage-based billing attribution

Attribution

The meter knows a resource consumed something. It does not know which of your customers to charge, or which of their departments.

With Hybr®Every reading is matched to a customer, and to their department or project, by subscription, cluster, namespace or API key.
Rating icon

Rating

Raw consumption is not a price. It has to meet the right price list, tier, reseller margin and currency before it means anything to finance.

With Hybr®Each reading is priced on that customer’s price list, with tiers, reseller margin and currency applied in a fixed order.
Timing icon

Timing

Finding out what a customer spent when the invoice runs is too late to do anything about it.

With Hybr®Usage is rated as it arrives, so you and your customer see the bill build up while the month is still open.
Proof icon

Proof

When a customer disputes a line, you need the records behind it, not a recollection of a spreadsheet.

With Hybr®Every invoice line traces back to the usage records and the price list that was applied.
Sources

What Hybr® meters

The same rating engine covers every platform under management, so a customer buying three different things still gets one bill.

Hybr collects usage from the systems you already run, such as Microsoft Partner Center, your hypervisors, your registered Kubernetes clusters and the LLM Gateway. Nothing is installed inside your workloads.

Seats, licences, consumption and tokens land on one invoice per customer, in that customer’s currency.

Microsoft CSP and Azure

Subscriptions, NCE terms and Azure consumption, priced per customer.

Microsoft CSP billing →

Virtual machines and storage

VMware vCenter, Azure Local, Hyper-V, Azure Stack Hub and OpenStack, metered per tenant.

Azure Local →

Kubernetes and GPU capacity

Node and GPU hours per registered cluster, shared or dedicated to one tenant.

Kubernetes & GPU-as-a-Service →

AI tokens

Token usage per tenant, per model and per key from the Hybr LLM Gateway.

LLM Gateway →

Your own services

Anything you publish to the catalog with an offer and a SKU, including managed services, managed applications and custom metrics.

Pricing models

Every way you charge, on one invoice

Pay-as-you-go, commitments, tiers, flat fees or a mix. Each model works for usage, licences and tokens, and a customer can have several on the same invoice.

  • Per customer: each customer gets its own price list, including negotiated rates.
  • Per reseller tier: resellers sell the same metered service at their own price, with their own margin.
  • Per currency: the same service is billed in each customer’s currency, without a second system.

Globex Retail · NovemberGPU usage: 180 hours
0 h100 h200 h
Invoice total$0.00
Example figures.
Before the invoice

Visible before the invoice, not after

Usage is rated while the month is still open, so you see what each customer owes before the invoice run. That changes what you can do with it.

  • Quotas: set limits per customer and per API key, so a runaway workload stops before it becomes an unbilled write-off.
  • Customers see their own usage: in the self-service portal, instead of asking your support team.
  • Margin per service: cost against price, while there is still time to act on it.
  • Budget alerts Roadmap on rated usage, when a customer reaches a threshold you set.
Globex RetailBilling period · November · Enterprise price list
Day 1 of 30
Azure consumptioncost + 12%
VM-hours · 4 vCPU$0.13 / hour
GPU-hours · A100$3.00 / hour
LLM tokens$4.80 / 1M
Microsoft 365 seatsfixed · $22.00 / seat
40 seats$880.00
GPU quota0 / 120 h
Invoice issuedPriced so far$880.00
Example figures.
Who it is for

Built for anyone who bills by consumption

Microsoft CSPs and MSPs

Adding consumption services such as Azure, GPU and AI to a seat-based portfolio.

Hosters and cloud providers

Selling infrastructure by the hour, the core or the gigabyte, through your own channel or resellers.

GPU clouds and AI providers

Selling GPU-hours, models and tokens per tenant instead of renting racks.

Enterprises

Charging shared platform, GPU and AI costs back to business units and cost centres.

FAQ

Questions about usage-based billing

What is usage-based billing?

Usage-based billing charges a customer for what they measurably used in a period, rather than only a flat subscription fee. It needs three things a subscription model does not: a meter that records consumption, attribution of that consumption to the right customer, and a rating step that turns raw units into a priced invoice line.

Does Hybr meter the workload itself?

No. Hybr collects usage from the systems you already run, such as Microsoft Partner Center, your cloud accounts, your hypervisors, your registered Kubernetes clusters and the LLM Gateway, and does the attribution, rating and invoicing on top. Nothing is installed inside your workloads.

Which platforms can Hybr collect usage from?

Microsoft CSP and Azure, VMware vCenter, Azure Local, Hyper-V, Azure Stack Hub, OpenStack, registered Kubernetes clusters with GPU capacity, and AI token usage from the LLM Gateway. Your own services can be metered too, through the catalog.

Can metered and fixed charges appear on one invoice?

Yes, and that is the normal case. One invoice can carry Microsoft CSP seats, a fixed managed-service fee, Azure consumption, GPU-hours and AI tokens, per customer and in their currency.

How do we stop a customer running up a bill they cannot pay?

Set quotas per customer and per API key. Usage is rated as it arrives, so you and your customer can see the bill building while the month is still open. Budget alerts on rated usage are on the roadmap.

What happens when a customer disputes a line?

Each rated line traces back to the underlying usage records and the price list that was applied, so a dispute is a lookup rather than a reconstruction.

Can resellers sell metered services under their own pricing?

Yes. Reseller tiers carry their own price lists and margins, so the same metered service reaches the end customer at a different price without duplicating the service. See reseller management.

Next step

Model it against your own price list

Bring one service you sell and the way you want to charge for it. We will build the offer, rate a period of real usage against it, and show you the invoice it produces.