Virtual power plants explained: MW, MWh and the fine print

Here’s a number to sit with: AGL’s virtual power plant fleet and Origin’s Loop VPP each now coordinate tens of thousands of household batteries and solar systems across the grid, but on any given afternoon the amount of power they can actually push back into the network at the same moment is a fraction of what the marketing implies. That gap between “we have this many batteries signed up” and “we can dispatch this many megawatts right now” is the whole story of virtual power plants, and it’s the bit most explainers skip.

I’ve spent a fair few afternoons this year on the phone with aggregators trying to pin down exactly what a VPP promises versus what it delivers on a stinking hot Tuesday in February. The short version: it’s a genuinely useful piece of grid infrastructure, built on real hardware in real driveways, but the branding runs well ahead of the dispatch reality. Let’s get into why.

What a virtual power plant actually is #

A virtual power plant is not a power plant. There’s no turbine hall, no single site you could drive to and photograph. It’s a software layer that aggregates a fleet of small, distributed devices — home batteries, rooftop solar inverters, sometimes smart hot water systems and EV chargers — and coordinates them to behave, from the market’s perspective, like a single generator or load. AEMO’s own Virtual Power Plant Demonstrations programme, which ran trials with several retailers before folding into standard market rules, described it plainly as an aggregation of distributed energy resources operated as if it were a single dispatchable unit.

The mechanics are straightforward enough. A household signs up, usually because they’ve already bought a battery and want it to earn something beyond just shaving their own bill — I’ve covered the payback maths on that decision in our piece on rooftop solar versus a home battery. The retailer or aggregator gets a data connection into the battery’s inverter, and in exchange for a software subscription, a discounted upfront battery price, or a cut of the revenue, the operator earns the right to charge or discharge that battery according to what the grid needs, not just what the household wants.

The megawatt versus megawatt-hour problem #

This is where I get pedantic, because the distinction actually matters to what a VPP can promise. Power, in megawatts, is the rate at which a fleet can push electricity onto the grid at a given instant. Energy, in megawatt-hours, is how much it can sustain that for. A VPP press release loves to quote the power figure — “50 megawatts of dispatchable capacity” sounds like a small gas peaker. But most home batteries have only two to four hours of stored energy at full output, and once they’re empty, that MW figure drops to zero until the sun comes back up or the grid recharges them.

So when an aggregator says its VPP can deliver a certain output, the honest follow-up question is: for how long, and how often can it do it again the same day? A gas peaking plant can run flat out for hours and refuel. A battery VPP is closer to a very large, very fast, very shallow reservoir. Useful for smoothing a five-minute price spike or riding through the evening peak; not a substitute for firm generation across a multi-day wind lull.

A worked example: what 2,000 batteries actually look like #

Let’s put real numbers on it, because rounding this to “lots of batteries” is how people get misled.

A fleet like that is a genuinely handy tool for AEMO during a tight five-minute dispatch interval — I’ve written before about how the NEM dispatches power every five minutes, and a VPP of this size can matter in that window. But 6 or 7 MW for two hours is not comparable to a grid-scale battery like the ones Akaysha Energy has been building for BlackRock, which I covered in our piece on Akaysha’s big battery bet on the NEM — those sites can be 100 MW-plus with four hours’ duration behind a single connection point and a single control system, not ten thousand separate ones behind ten thousand household Wi-Fi routers.

Who’s actually running these things #

South Australia has the longest-running example, the Tesla Powerwall programme originally rolled out across SA Housing Trust properties in suburbs like Salisbury and Elizabeth, later opened to any household willing to sign up. It’s become something of a reference case for how a state-scale VPP behaves under real stress, and I’ve gone through what that record does and doesn’t prove in our piece on South Australia’s renewable energy record. AGL and Origin both run national VPP programmes bundling batteries with retail plans, and in Western Australia, Plico Energy has built a retail model almost entirely around VPP-style battery coordination outside the NEM, in the standalone SWIS grid. Amber Electric takes a different tack again, giving customers direct wholesale price exposure and letting the battery software chase the spot price itself rather than acting purely on an aggregator’s instruction.

Every one of these operates under the same technical constraint: inverters have to comply with AS/NZS 4777.2 for grid connection, and the aggregator needs a live data link plus, in most cases, registration as a market participant or a sub-arrangement under one. That’s not exciting reading, but it’s the actual plumbing that makes a fleet of driveway batteries into something AEMO can call on.

The revenue stack, and where the real money sits #

Households are usually told a VPP earns money by selling power back at high prices during a peak. That happens, and it’s real — our explainer on negative electricity prices at midday covers the flip side of that same volatility, when batteries are paid or incentivised to soak up excess solar instead. But wholesale energy arbitrage is only one layer. The steadier income, for aggregators, tends to come from Frequency Control Ancillary Services, the fast, small, constant balancing signals AEMO buys to keep the grid at 50 hertz, and increasingly from network support payments where a distributor pays a VPP to avoid or defer a substation upgrade in a specific suburb. None of that shows up on a household’s bill in a way anyone would notice month to month, which is part of why the household-facing pitch stays focused on the once-a-year heatwave payout instead.

Where I think the branding oversells it #

Here’s my measured, not-especially-radical gripe: I reckon “virtual power plant” as a consumer-facing term does more marketing work than technical work. A retailer selling a battery with a VPP subscription attached is, more often than not, selling a battery — the VPP bit is a bonus revenue stream bolted on afterwards, not a guaranteed dispatch obligation the way a registered generator has under the National Electricity Rules. When aggregate capacity numbers get quoted in announcements, let’s be careful with that number; AEMO’s own VPP Demonstrations knowledge-sharing report from the programme’s final phase was candid that registered nameplate capacity and actually available capacity during a real event diverged meaningfully, partly because customers can opt out of dispatch on a given day, and partly because not every battery in the fleet is charged and available when the signal comes through. A VPP’s real capacity is whatever shows up on the day, not what’s on the sign-up sheet.

None of that makes VPPs pointless. It makes them one useful tool in a stack that includes the Capacity Investment Scheme, pumped hydro, gas peakers and grid-scale batteries, each suited to a different job. Our broader look at the Consumer Energy Resources roadmap covers how the AEMC and state governments are trying to knit household batteries, EVs and solar into the grid more formally, rather than leaving each retailer to build its own walled-garden VPP with its own app.

What it means if you’re weighing one up #

If you already own a battery, joining a VPP is usually low-risk and modestly profitable, and the Clean Energy Regulator’s rooftop solar and battery registration data shows uptake climbing steadily as more households make that call. If you’re buying a battery specifically to join one, run the payback maths first, the same way you would for any storage purchase, because the VPP credit is the icing, not the cake. And if a sales pitch quotes you a headline megawatt figure for a whole fleet, ask the duration question. It’s the same instinct I use sorting out a messy middlegame on the chessboard — the piece count on the board tells you less than which of them can actually move when it matters.

The technology is real, the households are real, and the grid genuinely draws on this fleet during tight five-minute intervals more often than most people realise. Whether it ever becomes the backbone AEMO’s long-term plans assume, rather than a useful supporting player, is still an open question, and one I’ll keep checking the numbers on.

For more detail on the underlying market mechanics, AEMO publishes its distributed energy resources work at aemo.com.au, and the Clean Energy Regulator tracks small-scale battery and solar installation data at cer.gov.au.

— Anjali Rao, Grid & Storage Correspondent

Photo by Romain Water on Unsplash