GRID AI / Investors
Building a new
distributed layer of
AI infrastructure.
AI compute requires enormous amounts of electrical capacity. GRID AI intends to place compute behind suitable existing electrical connections in homes and buildings, allow the property to retain priority, and coordinate distributed Nodes as one compute network.
Don't bring all the power to the compute.
Bring the compute to the power.
Current stage
From assumption to infrastructure data
Concept → prototype → model → measurement → validation → scale. Measured and validated are reserved for real MF01 data.
The infrastructure
already exists.
The compute layer does not.
Homes and buildings already have electrical connections, and the full capacity of those connections is not used continuously. GRID AI intends to deploy compute behind suitable connections and dynamically adapt Node consumption to the capacity available at each location. The property retains priority.
The strategic opportunity is not electricity itself. It is converting distributed available electrical capacity into coordinated, sellable AI compute.
Available capacity varies by property, location and moment.
Existing electrical infrastructure
Available capacity
GRID AI Nodes
AI compute
GRID AI network
Sellable compute capacity
The product is not the box.
A single Node is hardware. The intended value sits in the system around it.
Node
The physical unit of compute behind a connection.
Control
Node consumption follows the capacity available at the property.
Orchestration
Distributed Nodes coordinated as one compute network.
Host network
Acquiring, qualifying and keeping suitable locations.
Compute demand
Routing workloads to capacity that is available now.
Operations
Installation, monitoring, maintenance and field reliability.
Data
Instrumented performance per Node across the portfolio.
Capital
Funding Node deployment against what the data shows.
The value is the system that can deploy, control, monetize and finance thousands of Nodes.
Defensibility would not come from buying GPUs. It would come from building the infrastructure and operating system required to repeatedly turn suitable distributed electrical capacity into usable compute. That system is what GRID AI is being built to create.
Compute creates
the value.
AI customers buy compute. Compute revenue is intended to fund the electricity the Node consumes, the infrastructure, operations, the host benefit, capital and GRID AI's margin.
The base economic model assigns zero financial value to heat. Where useful, heat recovery may create additional upside.
AI customers
Compute revenue
GRID AI network
Power
Electricity the Node consumes.
Infrastructure
Hardware and installation.
Operations
Software, monitoring, maintenance.
Host benefit
Target residential participation.
Capital
Financing of the deployed asset.
GRID AI margin
What the network keeps.
Distributed infrastructure needs distributed hosts.
Provide or finance the Node
No hardware purchase by the host.
Arrange installation
GRID AI intends to organise the physical install.
Operate and maintain
Remote operation, monitoring and service.
Bear Node electricity
Electricity attributable to the Node is a GRID AI cost.
Property keeps priority
The Node reduces consumption when the property needs power.
Host the compute.
The target: compute covers your power.
GRID AI is investigating whether the economics of a residential Node can support not only its own electricity consumption, but also the normal household electricity cost of the host. If it holds, it creates an unusually simple host proposition. MF01 is designed to measure it.
Residential scale.
Commercial density.
One network.
Residential
- Large number of existing electrical connections.
- Potentially enormous deployment surface.
Commercial
- Potentially greater capacity per suitable location.
- Potentially more compute per installation.
One network
- Different deployment environments.
- Same infrastructure architecture.
- Same control layer.
- Same compute network.
The thesis becomes a machine.
MF01 is the first physical GRID AI Node. Its purpose is to turn the most important assumptions into measurements.
Assumed compute rate
Model
€0.39 / rented GPU hour ($0.45)
Measured
Effective utilization
Model
75%
Measured
Electrical availability
Model
Per-property capacity
Measured
Power consumption
Model
approx. 2.55 kW under load
Measured
Dynamic power response
Model
Node follows property demand
Measured
Technical uptime
Model
Continuous operation
Measured
Operating cost
Model
€1.000 / year reserve
Measured
Maintenance
Model
Remote-first, field service on exception
Measured
Node contribution
Model
approx. €3.418 / year
Measured
Model values are illustrative underwriting assumptions. Underwriting assumptions are model inputs and are not live market prices.
MF01 turns the thesis into data.
Capital follows proof.
- 01Underway
MF01
Technical and economic proof
Does one Node work?
- 02
10 Nodes
Repeatability
Can it be repeated?
- 03
100+ Nodes
Portfolio data
Do the economics hold across a portfolio?
- 04
Financeable infrastructure
Asset underwriting
Can predictable Node cash flows support asset financing?
- 05
Network scale
Capital deployment
Can capital deployment accelerate?
Prove the asset.
Finance the asset.
Scale the network.
From equity to
infrastructure capital.
Capital funds the proof.
Proof unlocks infrastructure capital.
Infrastructure capital scales the network.
Intended long term capital architecture
Equity builds the company.
Infrastructure capital builds the asset base.
In the early phase, equity capital may need to fund engineering, software, MF01, testing and initial deployments.
If Node economics later become sufficiently predictable, GRID AI intends to investigate financing the physical infrastructure separately. That would allow corporate equity to focus increasingly on technology, software, orchestration, operations, compute demand, host acquisition and network development.
Equity
Builds MF01
MF01 creates proof
Proof enables portfolio data
Portfolio data may enable asset financing
Asset financing deploys Nodes
Nodes generate compute revenue
Network scales
The long term capital question
How much infrastructure can €1 of GRID AI equity ultimately support?
If every Node must always be funded entirely with GRID AI equity, scaling the network would be extremely capital intensive.
If proven Node economics eventually support dedicated infrastructure financing, the relationship between corporate equity and deployed compute capacity could change materially. That is an important part of the GRID AI investment thesis.
One Node has to work
as a financial object.
The investor summary of the current scenario is below. The full model, sensitivities and production targets live on the Economics page.
| Item | Value | |
|---|---|---|
| Prototype investment | approx. €25.500 | Model |
| Full Node power under load | approx. 2.55 kW | Model |
| Assumed compute revenue — underwriting assumption | €0.39 / rented GPU hour ($0.45) | Model |
| Compute utilization | 75% | Model |
| Node electricity price | €0.28 / kWh | Model |
| Household electricity benefit | €1.050 / year | Target |
| Indicative annual Node contribution | approx. €3.418 / year | Model |
Prototype investment
Model- Value
- approx. €25.500
Full Node power under load
Model- Value
- approx. 2.55 kW
Assumed compute revenue — underwriting assumption
Model- Value
- €0.39 / rented GPU hour ($0.45)
Compute utilization
Model- Value
- 75%
Node electricity price
Model- Value
- €0.28 / kWh
Household electricity benefit
Target- Value
- €1.050 / year
Indicative annual Node contribution
Model- Value
- approx. €3.418 / year
Illustrative scenario published on the Economics page. Compute pricing is an underwriting assumption, not a live market price. Before hardware financing, tax and central GRID AI overhead.
Measured
Capital should buy proof.
The first capital is not intended to fund theoretical scale. It is intended to reach the next level of evidence.
- MF01 engineering
- Hardware
- Power control
- Software and orchestration
- Installation
- Testing
- Security
- Compute market integration
- Operations
- First multi-Node pilot
Build.
Measure.
Validate.
Then scale.
The questions MF01 must answer.
Technical
- Can the Node reliably follow available electrical capacity?
- Can workloads tolerate dynamic power behaviour?
- Can Nodes be operated remotely and securely?
Economic
- What compute rate can GRID AI actually measure?
- What effective utilization can be achieved?
- What does one Node contribute after power and operating costs?
Host
- Which properties are suitable?
- What does installation cost?
- How much friction exists in hosting and maintaining a Node?
Capital
- How predictable are Node cash flows?
- How does hardware depreciate?
- When does the asset become financeable?
We build to answer these questions with data.
From model to measured.
Each pair holds the current model value and the field designed to replace it with real infrastructure data once MF01 operates.
Compute rate
Model
$0.45 / rented GPU hour
Measured
Utilization
Model
75%
Measured
Node electricity
Model
€0.28 / kWh
Measured
Node contribution
Model
approx. €3.418 / year
Measured
Model values are illustrative underwriting assumptions. Underwriting assumptions are model inputs and are not live market prices.
Current capital stage
Prototype and validation
Next milestone
MF01 technical and economic proof
Follow the proof.
Investor access is a separate private portal — not yet open
Portal in developmentBuild the first proof
with us.
GRID AI is developing MF01 and the infrastructure required to test the technical and economic case for distributed AI compute behind existing electrical connections. For investors interested in the development of this infrastructure layer, register your interest.
Registering interest does not constitute an offer, allocation or commitment to invest.
Investor disclosure
GRID AI is in prototype development. MF01 is in development and not yet operational. Financial figures on this site are modelled scenarios, not historical performance; measured values will only be displayed once real operating data exists.
The household electricity benefit is a target model in development, not a current offer. Asset financing is a potential future capital strategy and is not currently committed. No investor return is promised, and no investment terms are offered through this page. Nothing here is an offer, a solicitation or investment advice.
The infrastructure thesis starts with one Node.
Prove one.
Repeat it.
Finance it.
Scale it.