In December 2025, we met Brendan Banfield, Co-founder and CEO of Gridsight, in Sydney on a scorching 37-degree day. As air conditioners hummed across the city, New South Wales’ four electricity utilities were preparing for a peak demand of 12 gigawatts.Historically, grid operators planned for these moments: a handful of peak hours each year when demand spikes.

What they haven’t planned for is data centres waiting in line to connect. Just one of those four utilities currently has 12 gigawatts of demand sitting in their connection queue. Twelve gigawatts from data centres alone, all wanting to plug into a network whose entire peak capacity is…4 gigawatts. 

“It’s staggering numbers just in terms of the load growth that we’re expected to see,” says Brendan. “Load growth has been stagnant over the last 15 years, not only in Australia, but around the world.”

In our conversation with Brendan, we discuss how intelligent grid management will be the unlock for both AI advancement and decarbonisation, why Australia has become the unexpected testing ground for the world’s energy future and what it takes to build software that can orchestrate millions of dynamic resources in real-time. 

The grid is breaking

The electricity grid is under strain as three massive forces arrive simultaneously and the system was designed for none of them. 

AI’s big appetite

The most recent and dramatic shift is AI. Since ChatGPT launched in late 2022, data centres globally have added energy demand roughly equivalent to powering an entire Australia. Hyperscalers are queuing up to connect loads that would add 30-50% to a grid’s peak demand; often to networks already operating at capacity. 

“There’s a race for artificial intelligence supremacy,” says Brendan, “but the utilities who manage the network infrastructure aren’t set up to handle the sheer number of connection requests coming through. 

When every third house becomes a generator 

Next is renewables. Australia now has one of the most decentralised energy systems in the world. One in three homes has rooftop solar and by 2030 residential solar is expected to be the largest generator in the National Electricity Market on the East Coast. Its scale is unheard of—in the best way possible—but it’s also constrained by intermittency. 

“The sun isn’t always shining; the wind isn’t always blowing,” says Brendan. While batteries help smooth intermittency, they don’t eliminate it. The grid is now managing millions of small, unpredictable generators instead of a handful of large, controllable ones. 

The great electrification

Finally, electrification. After more than a decade of flat demand, electricity consumption is rising again as transport, industry and everyday appliances move off fossil fuels and onto the grid. On its own, that transition would be manageable. But arriving simultaneously with AI and renewables is where things get a bit tricky. 

“We’re seeing this coalescing of these really large problems,” says Brendan. “How do we get more capacity out of the grid? How can we connect more things quickly without overloading the network infrastructure?”. 

The answer isn’t just building more infrastructure. Instead, it requires rethinking the assumptions that have governed grid planning for a century. 

The orchestration opportunity

For decades, grid operators have optimised for these 37+ degree days, or the handful of winter days each year when sub-zero temperatures drive peak load; built enough capacity to handle that moment and called it a day. That model collapses when data centres arrive in gigawatt blocks, solar output swings hourly and EV chargers create new demand spikes at the edge of the network. 

“Utilities need to get a lot more intelligent about planning and operating these large loads,” says Brendan. 

Building more infrastructure isn't the answer and creates a new crisis of cost and delay. Doubling or tripling poles and wires takes years and drives electricity bills higher, which are already hurting homeowners and businesses. 

Instead, Brendan believes the answer is using existing infrastructure more intelligently. 

All the technology that we need to achieve this already exists: the inverters, solar and electric vehicles,” says Brendan. “It’s a matter of orchestrating it in a more intelligent way to make sure we’re keeping up with this rapid load growth we’re seeing.”

In other words, the grid isn’t lacking assets, it’s lacking intelligence. 

“For the last hundred years, grid operators have effectively been running blind,” says Brendan. Operators could see the transmission network clearly, but much of what happened in the distribution network remained opaque.

That’s all changing as smart meters, IoT monitoring, improved network models and AI have turned the grid into a data-rich system; if someone can make sense of it. 

That shift laid the foundation for Gridsight.

The digital twin advantage

Gridsight pulls all these previously siloed data sources into a unified, electrical digital twin of the network. 

“Bringing all of that data together into one system provides us with this foundational data layer in which we can build more dynamic and flexible models on top of,” explains Brendan. 

The digital twin not only visualises the grid but also models its physics in real-time, running optimisation algorithms to coordinate resources. Only recently have advances in AI and machine learning made it feasible to model a system this large and complex. It’s one of life’s beautiful ironies that the same AI workloads causing havoc for the grid are also powering the technology that can help save it. 

“We augment the entire end-to-end workflow of connecting and managing new resources to the grid, through to real-time management and orchestration of these devices in the days, weeks, months and years ahead,” says Brendan. 

For a data centre submitting a connection request, Gridsight can model whether there’s capacity during peak hours as well as whether flexible connection terms could enable faster interconnection. “If the grid’s under strain, we can send a signal: The grid’s under pressure at the moment. Can we wind down your load? Can we shift some of that load to a different time of day? It will allow you to connect today as opposed to connecting in 2–4 years time,”. 

It’s a shift from binary yes/no connection decisions based on worst-case planning to nuanced, data-driven orchestration that maximises grid utilisation. And there’s really no better place to build this software than from Australia. 

Greetings from the future, love Australia 

Australia’s unique combination of geography, policy and timing has made it a live testing ground for the future of electricity infrastructure. 

“Australia’s in a really unique position for a variety of reasons,” Brendan explains. The country has world-leading rooftop solar penetration, driven by generous government incentives that were introduced in the early 2010s, combined with exceptional solar resources. But there’s another factor that makes Australia’s grid challenges uniquely acute. 

“We have a really vast land area with a relatively small population comparatively,” says Brendan. “This has led to huge amounts of electrical network infrastructure to supply what is a small amount of people.”

These unique features have led to a system prone to constraints and congestion. “A lot of the challenges that distribution and transmission networks around the world are going to face as they see this load growth and rise in decentralised renewables, we’re seeing here first in Australia.” 

The same forces reshaping Australia’s grid are rising elsewhere, just on a delayed timeline. Europe is adding solar rapidly. California’s duck curve—the supply-demand mismatch created by midday solar abundance—grows steeper each year. EV adoption is accelerating across developed economies.

“People often call it a postcard from the future—Australia's energy system compared to what we're seeing in the US and Europe," says Brendan.

For Gridsight, this has given them a strategic advantage. They don’t have a theoretical vision of how grids may need to evolve; they’re operating proven solutions to problems utilities around the world will face within years.

It’s evidence of a flexible grid management working at scale that the market desperately needs, as Brendan and the team found out when they entered the US market 

“For the last five years, it’s been death by pilots,” says Brendan. “Just these trials that go nowhere. Whereas in Australia, we’re solving these meaningful problems and delivering solutions to utilities across their entire network. The US networks we’re working with really see that as a huge differentiating factor for us.”

Gridsight landed its first two major US customers within 12 months of market entry, a timeline Brendan notes would be remarkable even for vendors who’d spent five years in the market. 

“We’re not just a science experiment,” says Brendan. “Having our core team here, that has been able to see the future and start to build towards what the grid should be in 10, 15, 20 years’ time, has meant that we’ve really got a headstart.” 

From bottleneck to abundance 

If we go right back to the start, Brendan has always seen opportunities where others saw challenges. As a young electrical engineer witnessing the solar boom firsthand, his colleagues complained that renewables were hurting the grid. Unconvinced, Brendan pursued a PhD, building a smart solar-powered house to provide distributed resources that could benefit both homeowners and the grid. When the pace of the utility sector felt too slow, he began chatting with his childhood friend Kurt—now his co-founder—about how they could accelerate the transition themselves.

Now a father, Brendan’s sense of urgency has only intensified. “The grid itself is the biggest bottleneck to the decarbonisation of the economy,” he says. “I truly believe that energy abundance for humanity is one of the key indicators toward a better quality of life for all.”

“Building software that removes this bottleneck and allows us to scale incredible technology like AI and renewables is incredibly motivating.” 

For Brendan and for the grid, there’s no time to waste.