Solar
Australia's abundant rooftop and large-scale solar resources create an important opportunity for renewable-powered compute.
SolarCompute was founded and built in Australia with a simple belief: the growth of AI and computing should not have to come at the expense of the environment that powers our future.
Australia has abundant solar and other renewable energy resources, alongside a growing base of homes and businesses generating and storing energy locally.
SolarCompute is putting that opportunity into practice through a live distributed compute network. Rather than relying only on ever-larger centralised data centres, the platform connects useful compute capacity across participating homes, businesses and other approved sites.
As the network develops, that creates an opportunity to bring computing capacity closer to where renewable energy is generated, stored and available.
SolarCompute operates connected compute machines at an initial Australian residential solar-powered site. The network is running supported workloads while its applications and operational systems continue through alpha testing and refinement.
Electricity usage and renewable-energy costs are managed using on-site metering and a custom monitoring dashboard. This provides an operational foundation for understanding energy use as the network develops.
Measured electricity consumption is not, by itself, proof of avoided carbon emissions. The source and timing of electricity supply, grid imports, system boundaries and an appropriate comparison method all matter.
SolarCompute is designed around a future in which distributed machines can increasingly operate alongside locally available renewable generation and storage.
Australia's abundant rooftop and large-scale solar resources create an important opportunity for renewable-powered compute.
Stored energy can extend the availability of renewable supply beyond the period in which it was generated.
Wind generation can form another part of the renewable energy mix supporting distributed infrastructure.
Hydroelectric generation can provide renewable electricity alongside solar, wind and storage resources.
Our long-term ambition is to work towards a verifiably lower-emissions computing network. Achieving that requires more than an environmental claim — it requires architecture, measurement, evidence and continuous improvement.
Renewable generation and stored energy can support participating sites and machines.
Homes, businesses and approved locations provide geographically distributed infrastructure.
Compatible machines make useful AI and compute capacity available to the network.
Workloads can be allocated to suitable available capacity as the network grows.
AI, training, simulation and other compute-intensive workloads require physical machines, electricity and supporting infrastructure. Large concentrations of computing capacity can also require substantial cooling and associated resources.
Distributed computing does not eliminate those impacts. Our goal is to build an alternative architecture that can make better use of existing sites, available machines and renewable energy while measuring the environmental characteristics of the network honestly.
Computing requires energy. Increasing the role of renewable supply is central to reducing its operational emissions.
Concentrated computing infrastructure creates heat and requires systems to manage it safely and efficiently.
Some data-centre cooling systems and energy infrastructure can place additional demand on water and natural resources.
Hardware, buildings and supporting infrastructure all carry environmental impacts that should be considered rather than ignored.
More intelligence should not require more environmental damage.
SolarCompute is an Australian-founded and built technology company. We believe Australia has an opportunity not only to use the next generation of AI and computing technology, but to help demonstrate how the infrastructure behind it can evolve responsibly.
Our commitment is to keep developing a network that can make better use of Australia's renewable energy opportunity while protecting the natural environment on which our future depends.
As monitoring expands, SolarCompute aims to distinguish electricity consumed, renewable generation and storage, grid imports, and estimated emissions. Any carbon-reduction claim will require a clearly defined baseline and suitable methodology.
Record energy used by machines and, where possible, the wider site and supporting infrastructure.
Identify the contribution of solar, battery and grid electricity without assuming all usage is renewable.
Apply appropriate emissions factors, time periods and a defensible comparison baseline.
Review methodology and supporting evidence before publishing quantitative environmental claims.
Environmental ambition needs evidence. SolarCompute intends to base future carbon and environmental reporting on measurable network activity and appropriate reporting methodologies.
Gather operational information from approved machines, workloads, infrastructure participants and energy-related records.
Check that reporting inputs are based on trusted and traceable network records.
Apply appropriate methodologies to verified information before environmental figures are produced.
Provide public-safe summaries that communicate progress without exposing private participant information.