What a cutting-edge distributed AI proposal might tell us about renewable energy, regional prosperity and the next frontier of digital inclusion

What if some of the most valuable infrastructure of the artificial intelligence age did not sit only inside enormous overseas data centres, but was also distributed through ordinary homes, businesses and community buildings?

And what if some of that computing could be powered by renewable electricity generated close to Whanganui?

That is not currently a plan for our region. It is the proposition set out in a recently published American white paper for a system called XFRA.

Whether XFRA itself succeeds remains to be seen. It is an ambitious commercial proposal at an early stage, not an independently proven model ready to be transferred directly to New Zealand.

But the possibility it illustrates deserves our attention.

It brings together two conversations that are already gathering momentum around Whanganui: the growth of renewable-energy generation and the rapidly increasing demand for the computing power required by AI.

Until now, these may have appeared to be separate discussions.

They may not be.

From renewable energy to renewable compute

Whanganui and the wider region are being considered for substantial new renewable-energy investment.

Meridian’s proposed Waiinu Energy Park, northwest of Whanganui, would combine wind generation, solar generation and grid-scale battery storage. The current proposal includes up to 56 wind turbines with 392 megawatts of capacity, a 230-megawatt solar farm and a battery energy-storage system. Meridian estimates that the development could generate around 1,600 gigawatt-hours of electricity each year if it proceeds. It has entered the fast-track consenting process, but significant assessment and decision-making still lie ahead.

Whanganui & Partners has also publicly identified a Whanganui-led regional data-centre hub as a strategic infrastructure opportunity capable of attracting AI and cloud-computing activity.

Even Meridian’s own explanation of why new generation is needed refers not only to the electrification of transport, heating and industry, but also to new demands such as the data centres required to support New Zealand’s use of AI.

That makes the central regional question larger than how much electricity we can generate.

It is also about what we do with it.

Will Whanganui primarily export electricity into the national grid while most of the high-value economic activity happens somewhere else?

Or could some of that energy support useful digital production here, creating local capability, enterprise, resilience and community benefit?

A different kind of computing infrastructure

The XFRA white paper proposes a distributed computing cloud made up of powerful GPU-based computing nodes installed in homes and small businesses.

Each node would be combined with a smart electrical panel, a whole-home battery, specialised cooling and a secure system for directing computing jobs across the network. Optional rooftop solar could also be incorporated.

Instead of relying entirely on a single enormous data centre, computing work would be divided between many smaller sites, which would operate together as one coordinated cloud.

The proposition is based partly on the observation that household electrical systems are designed to meet occasional peaks but normally operate well below their maximum capacity. XFRA aims to use some of that available electrical headroom without compromising the household’s own electricity needs.

During periods of pressure on the grid, the computing load could be reduced or transferred elsewhere. Batteries would help manage interruptions and could remain available to support the household during an outage.

The company says participating homeowners would receive the smart electrical panel, battery and computing equipment without paying for their installation. They would also receive a monthly hosting payment intended to reduce their electricity and high-speed broadband costs.

XFRA is not intended to replace the large, tightly connected data centres needed to train the biggest AI models. Its proposed role is complementary, supporting activities such as AI inference, cloud gaming, rendering and other computing work that can be divided across geographically distributed equipment.

The company says it already has revenue-generating test units operating and plans a 100-home American proof of concept later in 2026, followed by wider deployment from 2027. These remain the company’s plans and claims rather than evidence of proven large-scale delivery.

That distinction matters.

This is not a ready-made solution for Whanganui. American household electricity systems, regulation, housing patterns, energy markets and commercial arrangements differ from ours.

But as an example, XFRA changes what we can imagine.

Not something from nothing

There is no free energy in this proposition.

Computing equipment still requires electricity, cooling, minerals, manufacturing, skilled maintenance, secure communications and customers willing to pay for the work it performs.

Renewable infrastructure also has real environmental, landscape, transmission and community impacts. Calling an energy source renewable does not remove the need for careful assessment and responsible decision-making.

The opportunity comes from using existing and future systems more intelligently.

Electrical capacity that is normally idle, renewable generation that does not always align perfectly with ordinary demand, batteries, fibre connections and geographically distributed buildings may be coordinated in ways that were previously difficult or uneconomic.

The additional value does not appear from nowhere. It comes from making better use of assets and capacity that would otherwise be underused.

In that sense, energy becomes more than a commodity transported away through power lines.

It becomes an input into a high-value digital product: compute.

From technology consumer to economic participant

This is where the proposal becomes especially relevant to digital inclusion.

For many years, the digital divide has been understood primarily in terms of access and ability:

Can someone afford an internet connection?

Do they have a suitable device?

Can they use online services safely and confidently?

Can they obtain help when something goes wrong?

Those questions remain urgent. Many people in our community are still struggling with them.

But a new divide may be emerging.

It is the divide between people who merely consume digital technology and those who can own, host, govern or otherwise share in the value created by digital infrastructure.

Under the XFRA model, a qualifying homeowner would not simply use the internet. Their property, electricity connection and broadband service would become part of the physical infrastructure of the AI economy.

They might receive better household energy resilience, new equipment and a regular payment.

But who would qualify?

A person who owns a suitable modern home, has reliable fibre, understands the agreement and is confident dealing with a technology company may be well placed to participate.

A renter may not be.

A household in older or unsuitable accommodation may not be.

A person who lacks digital or financial confidence may struggle to assess the risks and benefits.

Social-housing tenants, people with insecure housing, low-income families and older people could again find themselves watching others enter an opportunity that was never designed with them in mind.

That would be a new form of digital exclusion—not exclusion from using technology, but exclusion from the wealth and resilience that technology creates.

Growth does not automatically produce inclusion

In an earlier article, Power, Place and Participation, I argued that a smart region should be measured not only by what it generates, but by who gets to take part.

That question becomes even more important when renewable energy, AI and regional infrastructure converge.

A major energy or computing development can create investment, construction work, rates revenue and national economic value. Those are genuine benefits.

But they do not automatically produce broad or lasting community prosperity.

A highly automated data centre may consume substantial electricity while employing relatively few people once construction is complete. Distributed computing equipment may generate valuable services while being owned and controlled elsewhere. Hosting payments may benefit selected property owners while bypassing renters and households experiencing the greatest hardship.

The important question is therefore not simply whether new technology creates value.

It is how that value is distributed.

What would community benefit look like?

Whanganui should not approach this opportunity by asking only how we can attract a data centre or copy XFRA.

We should ask:

How could Whanganui convert its emerging renewable-energy advantage into locally beneficial computing, capability and prosperity?

That might involve a conventional regional data-centre hub, smaller distributed installations, or a combination of both.

Computing nodes could potentially be located in suitable industrial premises, businesses, community facilities, schools, marae, housing developments or other sites—not only in individually owned homes.

Community trusts, iwi, hapū, social enterprises or co-operatives might hold ownership interests.

A share of commercial revenue could support lower-cost household energy and connectivity, local digital services, education, community resilience or long-term regional investment.

Schools, researchers, small businesses and community organisations could receive access to computing capacity that would otherwise be beyond their reach.

Local training pathways could prepare people for work in electrical systems, fibre, cybersecurity, equipment maintenance, energy management, data governance and AI services.

None of these outcomes would happen automatically. They would have to be deliberately negotiated and designed.

Questions to ask before the infrastructure arrives

The time to consider inclusion is not after the investment decisions have been made and the ownership arrangements have been locked in.

It is at the beginning.

Before supporting any energy-to-compute model, we should ask:

Who owns it?

Will the physical and digital assets be entirely externally owned, or will local communities have a meaningful stake?

Who receives the income?

Will value flow mainly to investors and technology companies, or will households and communities receive a fair and continuing share?

Who can participate?

Could renters, social-housing residents, rural communities and people on lower incomes participate, perhaps through shared or community-owned facilities?

Who carries the risk?

Who is responsible for equipment failure, fire, noise, heat, cybersecurity, insurance, household disruption and future removal of obsolete equipment?

Who controls the data?

What information would be collected about household electricity use, equipment operation and computing activity? How would privacy, security, Māori data sovereignty and community trust be protected?

What happens to electricity affordability?

Would new computing demand strengthen the economics of renewable generation and improve the use of the network, or could it increase costs and competition for electricity at difficult times?

What remains in Whanganui?

Beyond construction work, what permanent capability, ownership, employment, revenue and community resilience would remain here?

These are not arguments against development.

They are the questions that make good development possible.

AI can help us think, but it cannot choose our values

Artificial intelligence can help model electricity demand, coordinate distributed equipment, predict grid pressure and compare possible economic arrangements.

It may help us understand complex systems that would otherwise be extremely difficult to manage.

But AI cannot decide what is fair.

It cannot determine who should own critical infrastructure, how much value should remain in a host community, what obligations are owed to people who cannot participate, or how competing interests should be balanced.

Those remain human and community decisions.

The technology may be new, but the underlying questions are familiar: ownership, power, fairness, trust and responsibility.

An opportunity to lead the conversation

There is no reason to assume XFRA will become the dominant model, nor that it could be transferred unchanged to New Zealand.

But there is every reason to pay attention to the direction of travel.

Artificial intelligence is creating rapidly increasing demand for computing and electricity. Whanganui and its wider region may host significant new renewable-generation capacity. Local economic-development planning is already considering data centres and digital infrastructure.

Those conversations are beginning to meet.

The point is not to chase a shiny technology or to assume that every available unit of renewable electricity should be consumed by AI.

It is to recognise the possibility early enough to shape it.

Whanganui has an opportunity to ask whether renewable electricity can become more than an exported commodity—whether it can help create locally useful digital infrastructure, shared economic value and greater resilience.

We also have an opportunity to ensure that people who already experience digital, financial or social exclusion are not excluded again from the next layer of the digital economy.

The next step is not to purchase an XFRA system or announce that Whanganui will become an AI-computing centre.

It is to bring the right people into the same conversation: mana whenua, energy generators, electricity networks, economic-development organisations, technology specialists, local government, educators, housing providers, businesses and communities.

Together, we can ask what is technically possible, what is environmentally responsible, what is economically credible and what would genuinely serve the people of this region.

Because the future is not inclusive simply because it is renewable.

And it will not become fair simply because it is powered by artificial intelligence.

If energy is going to become compute, the question Whanganui should begin asking now is:

Who benefits?

The piece was developed collaboratively, blending  Alistair’s lived experience with AI-assisted reflection.