Every DCDATA Node pairs energy, compute and orchestration in a single deployable unit. DCOS — the DCDATA Orchestration System — is what turns thousands of independent, household-scale Nodes into one coherent, verifiable, 24×7 carbon-free compute cloud.
Every DCDATA Node is built from the same five components, wherever in the world it's deployed — the consistency is what makes the model repeatable across five very different regulatory markets.
Rooftop or ground-mount solar, sized for household/business need plus compute load plus resilience margin.
Backup power, demand-spike buffering and grid-services capacity, sized to exceed the Node's own compute draw.
One or more modular compute units, purpose-built for quiet, reliable residential and small-business operation.
Connects the Node to DCOS over a secure, encrypted network link for monitoring and control.
Circuit-level visibility and control, enabling DCOS to safely and dynamically manage the Node's live energy position.
A Node that meets its host's own energy needs first, powers its compute load second, and leaves resilience margin third — always in that order.
DCOS is the software layer that makes a fleet of independent household Nodes behave like a single, reliable cloud. For every Node, it continuously monitors solar generation, battery state of charge, household consumption, compute load and grid conditions — and acts on that live picture, second by second.
To the Nodes best placed to serve them, based on customer latency requirements and each Node's real-time energy and hardware capacity.
Home or business energy needs are always met before compute load is allowed to draw on-site power.
Workloads move to other Nodes in the fleet during a local outage, maintenance window or safety event — zero customer-visible downtime at the fleet level.
Generation, storage and consumption are matched and published hourly, Node by Node — not offset annually.
Nodes are engineered and contractually provisioned so that compute load is served first from on-site solar and battery. DCOS is designed to prevent a Node from drawing uncontracted power from the public grid to serve compute workloads.
Where a Node's on-site generation and storage are temporarily insufficient — several consecutive low-sunlight days, for example — DCOS throttles or fails the workload over to another Node in the fleet, rather than increasing draw from the local grid connection. Net-net, DCDATA adds generation and storage capacity to the communities it operates in; it never competes with households and businesses for scarce grid headroom.
Individual Nodes are small; the fleet is not. DCOS treats every Node as expendable and every workload as portable — customers contract for fleet-level service levels, typically failing over within the same metro region to preserve latency guarantees. That is, structurally, a more resilient promise than a single large data centre can make.
Every dollar of compute revenue is split in real time, on a shared, auditable ledger, between the host, the local energy retailer or utility partner, the hardware financier, and DCDATA. Nobody waits on a quarterly statement to know they were paid fairly.
| Stakeholder | What they contribute | How they are paid |
|---|---|---|
| Homeowner / small business | Rooftop, land, electrical connection, hosting | Fixed monthly hosting fee + share of compute revenue, paid automatically |
| Energy retailer / utility | Grid connection, metering, demand-response coordination | Service fee + share of grid-services value created |
| Hardware & capital partner | Solar, battery and compute equipment financing | Amortised return via a fixed share of Node revenue |
| DCDATA | DCOS platform, fleet operations, compute sales, security | Platform margin on compute sold to offtake customers |
Illustrative revenue-share structure; final commercial terms vary by market, regulatory regime, and partner agreement.
DCOS timestamps generation, storage and consumption at the Node level and publishes matched-hour carbon accounting — not annual renewable energy certificates.
| Node | Region | Generation | Battery SoC | Compute draw | 24×7 CFE |
|---|
Encrypted communications, strong identity and access controls, and workload isolation appropriate for a multi-tenant compute environment.
Physical Nodes use tamper-evident enclosures and remote monitoring. A compromised or removed device has its credentials revoked immediately and is cut off from the fleet.
The payment ledger is append-only and independently auditable by every stakeholder in the revenue share — no party has to take another's word for it.
DCDATA is not intended to replace hyperscale training infrastructure — the largest frontier training runs need single-site, tightly coupled clusters. DCDATA's fleet is built for the far larger, faster-growing set of inference, rendering and interactive workloads that hyperscalers themselves are increasingly pushing to the edge.
| Workload | Why DCDATA fits |
|---|---|
| AI inference (incl. RAG) | Modular, latency-sensitive and growing faster than any other compute category — well suited to distributed, edge-adjacent capacity. |
| Cloud gaming & interactive streaming | Proximity to players directly improves experience and reduces backbone congestion. |
| Cloud rendering & content production | Bursty, parallelisable workloads that can be scheduled flexibly across a distributed fleet. |
| Enterprise & sovereign AI under 24×7 CFE mandates | Customers with ESG reporting requirements need hour-matched, auditable clean energy — not offsets. |
The DCDATA White Paper covers Node engineering, fleet orchestration, power events and our global rollout methodology in depth.