-
Your expensive GPU spends a lot of time doing absolutely nothing.
Which is fine. It is yours. But what if "nothing" became optional?
Use your machine normally. Play. Render. Build. Run local models. Then, when you are not using all of it, decide whether the spare capacity should do something productive.
Your machine. Your rules. Your upside.
-
"I want that GPU."
We all know the next sentence: "But do I really need it?" Rutli changes the question.
A better GPU can mean better gaming, faster creation, bigger local AI models, more capable software, and potentially more productive spare capacity when you are not using all of it. That is not a promise that a GPU pays for itself. It is something simpler: better hardware can become useful in more hours of the day.
Turn "I want that GPU" into "I can justify that GPU."
-
When you need your GPU, you win.
Start a game. Launch a render. Run your own model. Do something more interesting than serving somebody else's batch job. The network gives way.
Rutli uses the capacity you choose to make available. It does not get first claim on your computer.
Your machine serves you first.
Play
Use it when you want it.
GPU 97% · your game · 144 fps
Idle
When you do not need all of it, decide what happens next.
GPU 2% · idle · most of the day, most days
Network
Put spare capacity to work.
GPU 70% of your 80% ceiling · a paid AI job · yours to stop
Value
Customer pays $1. Your node gets $1. Rutli fee $0.
the whole compute price, to the owner
Upgrade
Wanted the better one anyway? Now it has another job.
the same rules, on a bigger card
The AI cloud nobody owns.
Want that new GPU? Now it can pay you back.
Game harder. Create faster. Run your own AI. And when you are not using all that power, Rutli is being built to let you put spare capacity to work on paid AI workloads.
You keep 100% of the compute price your node generates.
The AI boom sent gamers the bill.
Your GPU can send one back.
AI turned compute into one of the world's hottest resources. Meanwhile, gamers are paying more for powerful hardware. Rutli gives GPU owners a way to participate in that economy instead of only paying for it.
An RTX 5090 lists at $1,999 and sells for $4,699 on the street in September 2026; AI data centres are forecast to take up to 70% of the world's memory output this year. CNBC, Tech Insider, IDC via VGTimes.
- Customer pays
- $1
- Your node gets
- $1
- Rutli fee
- $0
PLANNED
One machine. Three jobs.
Put spare capacity to work.
You choose whether to contribute. You choose when. You choose the limits. You choose the resource and power policy. Then you go back to using your computer.
Your local work comes first.
Run your own AI.
Private assistants. Local models. Research. Images. Voice. Home intelligence. Business tools.
Your machine should be useful even if network contribution is switched off forever.
Need more than your machine has?
Rutli is being built so your applications can reach into compatible network capacity when local hardware is not enough. Use your own power first. Reach for more when you need it.
Strong alone. Stronger together.
The hardware is yours. The compute value should be too.
PLANNED
A customer pays $1 for compute performed by your node.
You receive $1.
Rutli compute marketplace fee: $0.
What the owner keeps
Not included in the dollar
- Payment processing, at cost, shown separately.
- Payout to your account, at cost, shown separately.
- Your electricity.
- Your taxes.
Rutli does not make its money on your compute. It makes it on the services around the network, and it says so.
No platform taking 25% because it can. No mystery spread. No provider maths that needs a calculator.
Rutli should make money by making the network more useful and valuable, not by permanently owning a percentage of output from hardware it never bought.
Electricity, taxes, payment and payout costs and separately disclosed third-party costs are not included in the 0% compute marketplace fee statement. Public settlement is not yet live.
Why now.
The centralised path is running out of money, power and permission at the same moment that open models and owned hardware have become good enough.
- $725B
The centre has never been this expensive.
Yahoo Finance and Futurum, 2026 hyperscaler capex - 950 TWh
The power is not there.
IEA figures as reported, 2026; Brookings on grid strain - 70%
The public is saying no.
Data Center Watch Q1 2026; NBC News; Gallup via Brookings - 4 months
The models are open, and they are good.
Epoch AI open-closed gap; UK AISI; Semafor, 2026 - 32 GB
The hardware is already under the desk.
Northflank, GPUs for open-weight models, 2026; the founding nodes - $4,699
Gamers are paying for the data centres' memory.
CNBC, April 2026; IDC via VGTimes; TrendForce via PC Game Check; Tech Insider price tracking, September 2026
Powerful hardware has a new job.
Wanted the better one anyway? Now it has another job.
A high-end PC used to be mainly a consumption device. Then it became a creation device. Now it can become an AI device. Rutli adds one more role: infrastructure. Gaming rig. Creative workstation. Private AI computer. Rutli node. The same hardware, more reasons to own it.
And that creates a flywheel: more useful hardware, more people justify powerful machines, more nodes, more network capacity, more workloads, more value for node owners, another reason to own powerful hardware.
Hardware demand and distributed AI can reinforce each other.
Gaming is the wedge.
Gamers bring the first nodes. Software brings the rest.
Gamers are a natural starting point because they already buy and upgrade the exact hardware Rutli needs. But the strongest network is not made from one type of owner. Offices go quiet at different times. Home AI systems stay online for different reasons. Self-hosters already run private infrastructure. Businesses create real AI demand.
Every other kind of owner needs a reason to keep a machine on. We built those reasons, and they run today. A Rutli node is useful before it contributes one GPU cycle to the network.
Different owners. Different idle hours. One network underneath.
-
Offices
Private AI for accounting and business intelligence, on the company's own machine.
LIVE
-
Homes
Automation, security and a voice that answers from the house.
LIVE
-
Self-hosters
Messaging, calls, files and an assistant on infrastructure you control.
LIVE
-
Specialised business
Compliance and traceability where ownership of data matters.
LIVE
-
Anyone with a GPU
Chat, cited search, research and images on your own hardware.
LIVE
Some nodes bring compute. Some applications need it. Rutli is building both sides.
Pre-network-scale. Not pre-product.
Useful local software exists. The founding network is being built and tested. Every meaningful capability is labelled LIVE , PILOT , PLANNED or RESEARCH . No fake node maps. No imaginary GPU capacity. No "millions of users" because somebody registered a wallet.
Trust is infrastructure too.
Your next GPU could be part of the first network.
The public network is still being founded. That means the first participants get to help test the controls, the contribution policy, local priority, workload behaviour, the economics, hardware support, and the rules that should become difficult to break later.
Bring the machine. Help found the network.
PILOT Windows, Linux, macOS planned
A Founding Node is participation in an early technical cohort. It is not an equity interest, an investment product, or a guarantee of payment.