AGL’s latest investor update puts its virtual power plant fleet at something north of 200 megawatts of aggregated capacity across home batteries, pool pumps and controllable loads. Origin runs a comparable programme. Neither number sounds like much next to Eraring’s 2,880 megawatts of nameplate coal capacity. But a virtual power plant isn’t trying to be Eraring, and that’s precisely where most explanations of the technology go wrong. They compare it to a power station when it behaves more like a conductor standing in front of ten thousand small orchestras, trying to get them to play the same note at the same moment.
Let’s be careful with that word “plant” to start with, because nothing is being generated that wasn’t already going to be generated or stored. A virtual power plant, or VPP, is a software and contractual layer sitting over a fleet of distributed energy resources – rooftop solar inverters, home batteries, smart hot water systems, sometimes EV chargers – that lets an aggregator call on them collectively, as if they were one coordinated unit, and offer that aggregated capacity into the market or into a network service. The “plant” is the orchestration, not the hardware.
What’s actually being aggregated #
Start with the asset mix, because it varies more than most pitch decks let on. AGL’s VPP leans heavily on home batteries paired with its own retail plan, where the customer agrees to let AGL discharge or charge their battery within defined limits in exchange for a lower tariff or a sign-on credit. Origin runs something similar through its Loop platform. Energy Locals, Amber and a handful of smaller retailers have built their own variants, some built around wholesale price pass-through rather than a VPP discharge event at all, which is a genuinely different product even though it gets lumped into the same conversation.
The common thread is a household or business asset that was purchased for a private reason (cutting a power bill, backing up during an outage) getting a second job as a grid resource. That’s the elegant part of the idea. The battery exists regardless; the VPP just finds a second use for capacity that would otherwise sit idle most of the day.
Power versus energy, and why the distinction decides whether this works #
Here’s where I get fussy, because the trade press conflates these two units constantly and it matters. A VPP’s value proposition rests on megawatts of power – how much it can inject or withdraw from the grid at a given instant, typically for a short, defined window during a peak demand event or a price spike. That’s a different claim entirely from megawatt-hours of energy, which is how much total electricity it can shift over a period.
A fleet of 5,000 home batteries averaging 10 kilowatts each gives you a notional 50 megawatts of power. But if the average usable capacity per battery is 10 kilowatt-hours, and AEMO or the aggregator only wants an hour of discharge, you’ve got 50 megawatt-hours available once – then the fleet needs to recharge before it can do it again. Compare that to the Waratah Super Battery’s 850 megawatts of power and 1,680 megawatt-hours of energy sitting behind one gate, dispatchable on one instruction. The VPP’s number looks competitive on paper and behaves completely differently in practice, because you’re coordinating thousands of individually owned, individually sited, individually maintained assets rather than one.
How dispatch actually happens #
AEMO’s framework for this is the wholesale demand response mechanism, alongside a newer category of registration specifically for aggregated distributed energy resources. In practice, most VPPs today aren’t bidding directly into the five-minute dispatch engine the way a scheduled generator does – I’ve written before about how the NEM actually dispatches power every five minutes, and a VPP largely sits outside that loop, responding instead to price signals, network requests or its own retailer’s internal trigger.
The sequence usually runs like this: the aggregator’s platform watches the spot price forecast or a network operator’s request, decides a discharge event is worth calling, sends a signal to the fleet’s inverters or battery management systems, and the assets respond within a window, often minutes, sometimes closer to real time. The household doesn’t do anything; the contract they signed already authorised it, usually with caps on how often and how deep the discharge can go. AEMO’s own published material on this sits under its distributed energy resources programme, and it’s worth reading directly rather than through a retailer’s marketing page, because the constraints are spelled out more honestly there.
The reliability gap nobody likes discussing #
This is the part that deserves a straight answer rather than a sales pitch. A VPP’s biggest structural weakness isn’t technical, it’s behavioural and contractual. A customer can opt out of a discharge event, sell the house, have their battery offline for warranty service, or simply not have charged it that day because the week was cloudy. Multiply that across ten thousand participants and the aggregator’s forecast of available capacity has a confidence interval that a single large battery or gas peaker simply doesn’t carry.
Compare that to Kurri Kurri’s gas peaker or a big grid-scale battery, both of which I’ve covered on this site – Kurri Kurri: the peaker that took the long way round is one example – where the operator controls one asset end to end, with maintenance schedules and fuel contracts they can plan around. A VPP operator is, in effect, running a logistics business on top of someone else’s garage.
My own measured view, and I’ll own this as an opinion rather than settled fact: I think the industry oversold VPPs for a few years as a replacement for firm capacity, when the more honest pitch was always that they’re a demand-shaving and network-deferral tool first, and a wholesale market participant a distant second. The 2026 reality check on virtual power plants I wrote earlier this year covers how the aggregate numbers have landed versus the targets set a few years back, and the gap between pitch and delivered megawatts has been real.
Where the economics genuinely stack up #
None of that means the technology is pointless. It means the use case has narrowed to where it actually works. Network operators value VPPs most for peak shaving in constrained parts of the distribution network, where deferring a substation upgrade by coordinating a few hundred batteries in one suburb is cheaper than digging up the street. That’s a genuinely good trade, and it’s quieter than the headline “VPPs save the grid” framing because it’s happening suburb by suburb rather than state by state.
On the household side, whether a VPP makes financial sense depends enormously on the tariff structure you’re already on and what the aggregator pays for access to your battery. I went through the maths in home battery payback: does it stack up yet – the short version repeated here is that a VPP credit is usually worth somewhere between a modest discount and a few hundred dollars a year, which helps the payback period but rarely transforms it. Compare that against rooftop solar versus a home battery payback comparison and the VPP income sits as a secondary benefit, not the reason to buy the battery in the first place.
| Metric | Typical home battery VPP fleet | Grid-scale battery (e.g. Waratah-class) |
|---|---|---|
| Power (MW) | Tens to low hundreds, aggregated | Hundreds, single asset |
| Energy (MWh) | Limited, degrades fast on repeat calls | High, within one dispatch instruction |
| Dispatch reliability | Variable, opt-out risk | Contracted, operator-controlled |
| Primary value | Network deferral, peak shaving | Frequency control, firming, arbitrage |
Why the Consumer Energy Resources framework matters more than the hardware #
The technical capability of a modern battery inverter to respond to a dispatch signal in under a second has been solved for years. What’s held VPPs back from scaling faster is standardisation: how a retailer talks to an inverter from a different manufacturer, how consent is recorded, how AEMO visibility into aggregated fleets gets built without every aggregator running a bespoke integration. The Consumer Energy Resources roadmap sets out where the AEMC and AEMO are trying to land this, and it’s duller reading than a VPP press release but more consequential, because it decides whether a customer’s battery can talk to any retailer’s platform or only the one that sold it.
There’s an analogy I keep coming back to from long training rides, oddly enough: a peloton can move faster than any one rider, but only if the signals between riders are fast and trusted. A VPP fleet is the same problem at a different scale: the orchestration layer, not the individual battery, is what determines whether the whole thing moves as one.
What this looks like by the end of the decade #
AEMO’s Integrated System Plan treats distributed energy resources, VPPs included, as a meaningful contributor to the 2030s NEM, not a rounding error – rooftop solar alone is edging toward the kind of scale discussed in what the AEMO Integrated System Plan actually says, and a growing share of behind-the-meter batteries sitting on those same roofs is the natural next step for aggregation. The direction of travel is clear. What’s less clear, and what I’d watch over the next couple of years, is whether aggregators can close the gap between modelled fleet capacity and what actually turns up reliably on a 40-degree February afternoon when every air conditioner in the state is running and half the fleet’s owners quietly opted out of the discharge event because nobody wants their home battery flat before dinner.
That’s not a reason to write the technology off. It’s a reason to read the capacity numbers the way you’d read any aggregated forecast – with the confidence interval attached, not just the headline megawatt figure.
– Anjali Rao, Grid & Storage Correspondent
Photo by Rahul Karyal on Unsplash