The operator's view
of the energy transition.

Getting households off gas is less of a technology problem than it appears. The hardware works. What doesn't work — reliably, at scale — is the coordination: a qualified installer available within a reasonable timeframe, a financing option that isn't confusing, a service experience smooth enough that the customer doesn't give up halfway through. I ran AGL's home electrification business and built the platform that now serves 600,000+ households, and the consistent finding was that the people who didn't electrify weren't unconvinced — they were let down by friction, a quote that never arrived, or a workflow that assumed more customer effort than was reasonable. Sequencing and installer capacity matter as much as subsidies.

Most demand response programs lose customers not because the technology fails but because the experience asks too much of them. The best virtual power plant runs in the background, pays out reliably, and occasionally sends a notification that feels like good news — customers who notice the benefit, not the effort of participation, are your best customers. I scaled AGL's Peak Energy Rewards program to 300,000+ households. The lesson: you can't engineer the revenue without engineering the experience first. Getting FCAS, wholesale, and network value to stack without the customer needing to understand any of those terms is the real engineering challenge — and it's mostly a software and orchestration problem, not a hardware one.

Distribution networks are making multibillion-dollar investment decisions with surprisingly limited real-time visibility into what's actually on their network. The gap between what operators can see and what's happening at the edge is closing — but unevenly, and more slowly than the pace of DER uptake demands. I back GridSight because the value of proper grid analytics doesn't require an AI narrative to justify: if you can see your network accurately, you make better investment decisions, respond to faults faster, and avoid building infrastructure you don't yet need. The opportunity is less about frontier technology and more about finally acting on data that already exists but isn't being used.

Most EV conversations happen at the technology layer — range, charging speed, battery chemistry. The harder problems are system problems: how does a large fleet of vehicles participate in the grid without destabilising it, what does a smart charging tariff need to look like for customers to actually engage with it, and what does market entry look like for an OEM that has never operated in a regulatory environment like Australia's. I've worked on all three. At AGL I got insights into smart charging trials and EV-specific tariff design as part of the broader demand flexibility program — direct, operator-side experience of how EVs sit within an energy system under load. I've also provided formal advisory to leading Chinese OEMs on Australian market entry across retail energy partnerships, charging infrastructure, and the regulatory landscape. The thread between these engagements is the same one that runs through my broader DER work: EVs are not a technology problem, they're a coordination problem — between the vehicle, the network, the retailer, and the customer.

When I look at a climate-tech venture, I'm asking one question: does this make something that's currently hard, slow, or invisible into something fast, clear, and actionable? I'm most drawn to businesses where software compresses operational complexity — where a good product doesn't just automate an existing task but changes what decisions are even possible. I backed GridSight because good network analytics converts data that already exists into operational clarity that previously required expensive fieldwork. I backed OpenSolar because solar installation remains more fragmented and manual than it should be, and a platform that connects installers, customers, and suppliers at scale creates durable value. The founders I find most credible understand the workflow they're replacing at a granular level — they've usually lived inside the problem before trying to solve it.