For space operators with GPU infrastructure

Host GPU at your space. Turn it into a compute business.

Turn dedicated GPU infrastructure into a profitable compute business. Presell upcoming capacity to our sales team and have customers ready before hardware arrives.

Revenue before hardware arrivesWe handle sales, billing & supportNo long-term commitment required
host-dashboard — GPU nodes

Active nodes

8 × H100

80 GB SXM5

Utilisation

94%

Last 30 days

This month

HK$41,200

Settled next cycle

Job queue6 running · 2 queued

GPU1

GPU2

GPU3

GPU4

GPU5

GPU6

GPU7

GPU8

Day 1

Revenue can start before hardware arrives

24 / 7

We handle customer support so you don't

Zero

Sales headcount needed on your side

One

Dashboard to track utilisation and earnings

A clear division of responsibility

You own the hardware and the space. We own the commercial layer. Neither side steps on the other.

What we handle

HeyOffice takes care of this

Customer acquisition & platform exposure

Your GPUs are listed on our platform and surfaced to businesses actively looking for compute. You do not need a sales team.

24/7 customer support

Our team handles every inbound question, ticket and complaint from compute customers. You never deal with end-user issues.

Billing, invoicing & payments

We collect payment, issue invoices and reconcile usage. Revenue lands in your account on a regular settlement cycle.

Job scheduling & resource optimisation

Workloads are routed, queued and balanced across available hardware so your GPUs are busy when demand exists.

Marketing & demand generation

We run the campaigns, partnerships and content that bring customers to the platform. Your hardware benefits from shared growth.

What you handle

Your side of the partnership

Hardware uptime & maintenance

Keep the GPUs powered, stable and reachable. Replace faulty components. We route jobs away during your maintenance windows.

Network connectivity

Provide reliable, low-latency internet. Our platform communicates the bandwidth requirements before onboarding.

Physical infrastructure

Cooling, power, rack space and physical security are your responsibility. We advise on common configurations.

Initial hardware setup

Install the required drivers and run our onboarding agent. Setup typically takes a few hours with our technical guide.

How it works

From unused hardware to earning compute in six steps

01

Register your hardware

Tell us your GPU model, quantity and location. We assess your setup and confirm compatibility with the platform.

02

Presell upcoming capacity

Our sales team starts selling your compute before the hardware arrives. Customers queue up so you have revenue from day one.

03

Install the onboarding agent

A lightweight agent connects your machines to the scheduling layer. Takes a few hours following our step-by-step guide.

04

Go live

Jobs start flowing to your GPUs. You watch utilisation in your host dashboard. We handle the rest.

05

Receive settlement

Revenue from completed jobs is reconciled and paid out on a regular cycle. Full usage breakdown in your dashboard.

06

Scale when ready

Add more GPUs as demand grows. Our sales pipeline grows with supply — your hardware is never listed idle for long.

Revenue model

Your hardware earns while you focus on running your space

Compute revenue is metered per job and settled regularly. You see a full breakdown of GPU hours consumed, jobs completed and gross revenue before our platform fee in your host dashboard. No surprises — the split is agreed upfront and does not change.

  • Revenue-share model — you keep the majority, we take a platform fee
  • Settlement on a fixed cycle with itemised usage breakdown
  • Minimum commitment: none — scale hardware up or down as you choose
  • Presales start before your hardware is live — no idle period

Sample earnings — 8× H100 node

Gross compute revenue

/mo at 90% utilisation

HK$52,000

Platform fee (illustrative)

20% example — agreed at onboarding

− HK$10,400

Host settlement

/mo to your account

HK$41,600

Illustrative only. Actual earnings depend on GPU model, utilisation rate, agreed split and market demand. Revenue share percentage is confirmed during onboarding.

Minimum requirements

What we look for in a host partner

GPU tier

NVIDIA data-centre grade — H100, A100, A6000 or equivalent

Connectivity

Dedicated high-bandwidth internet with low contention and stable latency

Power & cooling

Sufficient power headroom and cooling for continuous full-load operation

Physical security

Access-controlled environment with appropriate monitoring

Not sure if you qualify? Reach out and we will assess your setup — no obligation.

Ready to turn your GPU infrastructure into a compute business?

Speak to our team about your hardware, location and goals. We will confirm fit, agree the revenue split and start preselling before your hardware is even racked.