A retailer rang a Blackburn North customer at 4.40pm on a stinking hot Tuesday in late January this year and asked, politely, to borrow four kilowatt-hours out of her Tesla Powerwall for the next ninety minutes. She said yes, because she’d ticked a box eighteen months earlier and mostly forgotten about it. That one exchange, repeated across a few thousand households simultaneously, is the entire concept of a virtual power plant in miniature. No new poles, no new generator, just permission to use batteries that already exist.
I’ve spent a fair bit of this year going through VPP enrolment numbers from the retailers and from the Clean Energy Regulator’s small-scale data, and the honest picture is more interesting than the marketing. These things work, technically. Whether they’re working hard enough to matter at grid scale is a separate question, and it’s the one worth sitting with.
What a VPP actually is, and what it isn’t
Strip away the branding and a virtual power plant is a software layer sitting over a fleet of distributed batteries – home batteries mostly, sometimes EVs, occasionally pool pumps and hot water systems – that lets an operator dispatch them as one coordinated resource. The “plant” doesn’t exist as a single piece of infrastructure. It’s an aggregation contract plus a comms protocol plus a control algorithm, pointed at assets sitting in garages and meter boxes from Blacktown to Bentleigh.
The distinction that gets flattened in most coverage, and it’s one I keep having to correct in conversations with people who aren’t steeped in this stuff, is between power and energy. A VPP might advertise 200 megawatts of capacity. That’s power – the rate at which it can discharge at a single instant. It says nothing about how long it can sustain that rate. A fleet of 200MW built from Powerwalls averaging 10kWh of usable storage each, assuming something like 5kW of continuous discharge per unit, might only hold enough energy for a 60 to 90 minute full-rate event before the batteries are drained. AEMO’s own Integrated System Plan modelling treats distributed storage duration as a genuinely different input to distributed storage power, and for good reason – what the AEMO Integrated System Plan actually says about this split is worth reading if you want the detail behind the headline megawatt figures.
Where the money actually comes from
There are essentially three revenue streams stacked on top of each other, and most operating VPPs are chasing all three at once because none on its own pays for the battery.
The first is wholesale arbitrage – charge when prices are low or negative, discharge into the evening peak. The second is frequency control ancillary services, the FCAS markets AEMO runs in roughly 4-to-6-second and longer windows to keep the grid at 50Hz; aggregated batteries are well suited to this because the response has to be fast and short, which plays to a battery’s strength and a gas peaker’s weakness. The third, and increasingly the one retailers spruik hardest, is demand response: getting paid, directly or via avoided network costs, to shave a network’s peak demand on the handful of days a year that actually threaten supply.
None of those three pays enough by itself to make the household battery stack up financially inside, say, five years. I went through this ground carefully in a separate piece on home battery payback and whether it stacks up yet, and the VPP revenue is best thought of as shaving a year or two off an otherwise long payback, not transforming the economics outright.
A worked example, because the numbers get thrown around loosely
Take a fleet of 5,000 enrolled Powerwall-class batteries, each rated at 5kW continuous discharge and 13.5kWh usable capacity, with the retailer reserving 70% of each household’s charge for VPP dispatch and leaving the rest for the owner’s own evening use.
Available power: 5,000 × 5kW = 25MW.
Available energy at that reserved fraction: 5,000 × 13.5kWh × 0.7 ≈ 47.25MWh.
Run flat out at 25MW, that energy bank lasts: 47.25MWh ÷ 25MW ≈ 1.9 hours.
So the honest advertised figure for that fleet during a tight evening is “25MW for just under two hours,” not “25MW VPP.” That distinction is exactly the one that matters when a network operator is trying to work out whether a VPP can cover a four-hour evening ramp on a 40-degree day, or only the sharpest hour of it. It’s the same discipline I’d apply to reading a capacity factor on a wind farm: the nameplate number and the delivered number are different animals, and conflating them is how people get surprised. There’s a good explainer on this site covering what a capacity factor is and why it actually matters if the analogy’s useful.
Who is actually running these in Australia right now
Tesla’s South Australian VPP, built initially off Housing Trust properties and expanded through AGL and other partners, remains the most cited example and genuinely was early: the programme has been running since 2018 and has been through several iterations. I covered the broader Tesla Energy Australia footprint, including the Powerwall fleet and the Hornsdale battery it’s unrelated to but often confused with, in an earlier profile. AGL, Origin, EnergyAustralia and Amber Electric all run some version of a VPP or battery-coordination product now, with different splits between fixed credits and dynamic dispatch payments. Zen Energy’s involvement in Whyalla ties a retail VPP offer to a specific industrial and community context that’s worth its own look, which I gave it here.
What’s changed since the early South Australian trials is scale and standardisation. The Australian Energy Market Commission has been working through rule changes to make it easier for aggregated distributed resources to participate directly in NEM markets rather than through bespoke bilateral deals with a single retailer, and AEMO has been building out the technical standards for what it calls Distributed Energy Resource (DER) registration. That’s plumbing work, unglamorous, but it’s the difference between VPPs as a retailer marketing exercise and VPPs as a real market participant category.
The limits nobody likes dwelling on in the press release
Enrolment doesn’t equal participation. A household can sign the VPP contract and still have a fully charged battery that never gets called on, either because the retailer’s dispatch algorithm didn’t need it that particular evening or because the household opted out of a specific event through the app. Churn is real too: people switch retailers, batteries get decommissioned, warranty disputes freeze participation. The headline “X thousand homes enrolled” figure that gets quoted in announcements is a ceiling, not a delivered capacity.
There’s also a coordination problem that’s structurally awkward and I don’t think gets resolved cleanly any time soon. A household battery sits behind a distribution network’s meter. The retailer running the VPP wants to dispatch it for a wholesale or FCAS signal. The distribution network operator, separately, might want the same battery held in reserve to manage a local voltage or thermal constraint on its own feeder. These two instructions can conflict, and the rules for resolving that conflict, who has priority, who compensates whom – are still being worked out at AEMC and AER level rather than fully settled. If you want the fuller context on how that same tension plays out with solar export limits, I’d point to the piece on rooftop solar and grid stability, because the physics underneath is the same: a lot of small devices behind meters, a grid built around the assumption of one-way power flow from a handful of big generators.
Where I think the sector oversells it
Here’s my actual read, and it’s mildly against the grain of most VPP coverage I see: the household-battery VPP model, as currently marketed, is being asked to do a job it’s only partly suited for. The industry pitch is that VPPs are a cheap, fast alternative to grid-scale batteries and gas peakers. For the FCAS and short-duration ancillary markets, fair enough, that’s true and the aggregated response is genuinely competitive. But for covering the four-to-six-hour evening firming gap that the grid actually needs filled as coal retires: the job central to debates like the one in gas peakers versus big batteries – a fleet of household batteries with under two hours of usable reserve per cycle just doesn’t have the energy depth. Big grid-scale batteries like Waratah Super Battery or the Hornsdale expansion are built with duration in mind from day one. Most home battery VPPs weren’t, because the household’s own evening consumption competes directly with the retailer’s dispatch need for the same stored energy. You can’t sell the same kilowatt-hour twice.
None of that makes VPPs a bad idea. It makes them a genuinely useful, cheap-to-deploy tool for a specific slice of grid services (frequency response, short peak shaving, some wholesale arbitrage) stacked on top of assets households were buying anyway for backup and bill savings. Treating that as equivalent to new firm grid-scale capacity is where the sales pitch outruns the engineering.
What’s actually worth watching from here
The Capacity Investment Scheme has started drawing in aggregated distributed resources alongside grid-scale projects in some tender rounds, which is a meaningful shift in how AEMO and the Commonwealth think about counting this capacity toward reliability obligations: the mechanics are covered in the piece on how the Capacity Investment Scheme actually works. Watch also for how AEMO’s draft standards for DER dispatch visibility bed in; right now a meaningful share of enrolled batteries still aren’t visible to AEMO’s control room in real time, which is a genuine blind spot for system security during tight periods, documented in AEMO’s own published materials on distributed energy resources integration.
And keep an eye on the AER’s ongoing work on network tariff reform, because the distribution-network-versus-retailer dispatch conflict I mentioned earlier won’t resolve through goodwill. It’ll resolve through pricing signals and rule changes, the kind the Australian Energy Regulator is slowly working through for DER export and import tariffs.
I’ll admit I used to be more sceptical of the whole category than I am now: a few years of watching the FCAS numbers come in changed my mind on that slice of the business, if not on the headline capacity claims. Worth remembering, next time a retailer quotes a big VPP megawatt figure, to ask the follow-up question: for how long, and compared to what.
– Anjali Rao, Grid & Storage Correspondent
Photo by Andrey Metelev on Unsplash