Will Virtual Power Plants Ever Really Be a Thing?

The story of electricity in the modern economy is one of large, centralized generation sources — fossil-fuel power plants, solar farms, nuclear reactors, and the like. But devices in our homes, yards, and driveways — from smart thermostats to electric vehicles and air-source heat pumps — can also act as mini-power plants or adjust a home’s energy usage in real time. Link thousands of these resources together to respond to spikes in energy demand or shift electricity load to off-peak hours, and you’ve got what the industry calls a virtual power plant, or VPP.

The theoretical potential of VPPs to maximize the use of existing energy infrastructure — thereby reducing the need to build additional poles, wires, and power plants — has long been recognized. But there are significant coordination challenges between equipment manufacturers, software platforms, and grid operators that have made them both impractical and impracticable. Electricity markets weren’t designed for individual consumers to function as localized power producers. The VPP model also often conflicts with utility incentives that favor infrastructure investments. And some say it would be simpler and more equitable for utilities to build their own battery storage systems to serve the grid directly.

Now, however, many experts say that VPPs’ time to shine is nigh. Homeowners are increasingly pairing rooftop solar with home batteries, installing electric heat pumps, and buying EVs — effectively large batteries on wheels. At the same time, the ongoing data center buildout has pushed electricity demand growth upward for the first time in decades, leaving the industry hungry for new sources of cheap, clean, and quickly deployable power.

“VPPs have been waiting for a crisis and cash to scale and meet the moment. And now we have both,” Mark Dyson, a managing director at RMI, a clean energy think tank, told me. “We have a load growth crisis, and we have a class of customers who have a very high willingness to pay for power as quickly as possible.” Those customers are the data center hyperscalers, of course, who are impatient to circumvent the lengthy grid interconnection queue in any way possible, potentially even by subsidizing VPP programs themselves.

Jigar Shah, former director of the Department of Energy’s Loan Programs Office under President Biden, is a major VPP booster, calling their scale-up “the fastest and most cost-effective way to support electrification” in a 2024 DOE release announcing a partnership to integrate VPPs onto the electric grid. While VPPs today provide roughly 37.5 gigawatts of flexible capacity, Shah’s goal was to scale that to between 80 and 160 gigawatts by 2030. That’s equivalent to around 7% to 13% of the U.S.’s current utility-scale electricity generating capacity.

Utilities are infamously slow to adopt new technologies. But Apoorv Bhargava, CEO and co-founder of the utility-focused VPP software platform WeaveGrid, told me that he’s “felt a sea change in how aware utilities are that, building my way out is not going to happen; burning my way out is not going to happen.” That’s led, he explained, to an industry-wide recognition that “we need to get much better at flexing resources — whether that’s consumer resources, whether that’s utility-sited resources, whether that’s hyperscalers even. We’ve got to flex.”

So Why Aren’t We Already Flexing?

Actual VPP capacity appears to have grown more slowly over the past few years than the enthusiasm surrounding the resource’s potential. According to renewable energy consultancy WoodMackenzie, while the number of new VPP programs, offtakers, and company deployments each grew over 33% last year, capacity grew by a more modest 13.7%. Ben Hertz-Shargel, who leads a WoodMac research team focused on distributed energy resources, attributed this slower growth to utility pilot programs that cap VPP participation, rules that limit financial incentives by restricting how VPP capacity is credited, and other market barriers that make it difficult for customers to engage.

Dyson similarly said he sees “friction on the utility side, on the regulatory side, to align the incentive programs with real needs.” These points of friction include requirements for all participating devices to communicate real-time performance data — even for minor, easily modeled metrics such as a smart thermostat’s output — as well as utilities’ hesitancy to share household-level metering data with third parties, even when it’s necessary to enroll in a VPP program. Figuring out new norms for utilities and state regulations is “the nut that we have to crack,” he said.

One of the more befuddling aspects of the whole VPP ecosystem, however, can be just trying to parse out what services a VPP program can actually provide. The term VPP can refer to anything from decades-old demand response programs that have customers manually shutting off appliances during periods of grid stress to aspirational, fully integrated systems that continually and automatically respond to the grid’s needs.

“When a customer like a utility says, I want to do a VPP, nobody knows what they’re talking about. And when a regulator says we should enable VPPs, nobody knows what services they’re selling,” Bhargava told me.

In an effort to help clarify things, the software company EnergyHub developed what it calls the VPP Maturity Model, which defines five levels of maturity. Level 0 represents basic demand response. A utility might call up an industrial customer and tell them to reduce their load, or use price signals to encourage households to cut down on electricity use in the evening. Level 1 incorporates smart devices that can send data back to the utility, while at Level 2, VPPs can more precisely ramp load up or down over a period of hours with better monitoring, forecasting, and some partial autonomy — this is where most advanced VPPs are at today.

Moving into Levels 3 and 4 involves more automation, the ability to handle extended grid events, and ultimately full integration with the utility and grid-operator’s systems to provide 24/7 value. The ultimate goal, according to EnergyHub’s model, is for VPPs to operate indistinguishably from conventional power plants, eventually surpassing them in capabilities.

But some question whether imitating such a fundamentally different resource should actually be the end game.

“What we don’t need is a bunch of virtual power plants that are overconstrained to act just like gas plants,” Dyson told me. By trying to engineer “a new technology to behave like an old technology,” he said, grid operators risk overlooking the unique value VPPs can provide — particularly on the distribution grid, which delivers electricity directly to homes and businesses. Here, VPPs can help manage voltage regulation or work to avoid overloads on lines with many distributed resources, such as solar panels — things traditional power plants can’t do because they’re not connected to these local lines.

Utilities Can Build Big Batteries, Too

Still others are frankly dubious of the value of large-scale VPP programs in the first place. “The benefits of virtual power plants, they look really tantalizing on paper,” Ryan Hanna, a research scientist at UC San Diego’s Center for Energy Research told me. “Ultimately, they’re providing electric services to the electric power grid that the power grid needs. But other resources could equally provide those.”

Why not, he posited, just incentivize or require utilities to incorporate battery storage systems at either the transmission or distribution levels into their long-term plans for meeting demand? Large-scale batteries would also help utilities maximize the value of their existing assets and capture many of the other benefits VPPs promise. Plus, they would do it at a “larger size, and therefore a lower unit cost,” Hanna told me.

Many VPP companies would certainly dispute the cost argument, and also note that with grid interconnection queues stretching on for years, VPPs offer a way to deploy aggregated resources far more quickly than building out and connecting new, centralized assets.

But another advantage of Hanna’s utility-led approach, he said, is that the benefits would be shared equally — all customers would see similar savings on their electricity bills as grid-scale batteries mitigate the need for expensive new infrastructure, the cost of which is typically passed on to ratepayers. VPPs, on the other hand, deliver an outsize benefit to the customers incentivized to participate by dint of their neighborhood’s specific needs, and with the cash on hand to invest in resources such as a home battery or an EV.

This echoes a familiar equity argument made about rooftop solar: that the financial benefits accrue only to households that can afford the upfront investment, while the cost of maintaining shared grid infrastructure falls more heavily on non-participants. Except in the case of VPPs, non-participants also stand to benefit — just less — if the programs succeed in driving down system costs and improving grid reliability.

“I may pay Customer A and Customer B may sit on the sidelines,” Matthew Plante, co-founder and president of the VPP operator Voltus, told me. “Customer A gets a direct payment, but customer B’s rates go down. And so everyone benefits, even if not directly.” On the flip side, if the VPP didn’t exist, that would be a lose-lose for all customers.

Plante is certainly not opposed to the idea of utilities building grid-scale batteries themselves, though. Neither he nor anyone else can afford to be picky about the way new capacity comes online right now, he said. “I think we all want to say, what is quickest and most efficient and most economical? And let’s choose that solution. Sometimes it’s got to be both.”

For its part, Voltus is betting that its pathway to scale runs through its recently announced partnership with the U.S. division of Octopus Energy, the U.K.’s largest energy supplier, which provides software to utilities to coordinate distributed energy resources and enroll customers in VPP programs. Together, they plan to build portfolios of flexible capacity for utilities and wholesale electricity markets, areas where Octopus has extensive experience. “So that gives us market access in a much quicker way,” Plante told me.”

Hyperscalers Step Up

At this moment, there’s no customer more motivated than a data center to bring large volumes of clean energy online as quickly as possible, in whatever way possible. Because while data enters themselves can theoretically act as flexible loads, ramping up and down in response to grid conditions, operators would probably rather pay others to be flexible instead.

“Does a data center company ever want to say, okay, I won’t run my training model for a couple hours on the hottest day of the year? They don’t, because it’s worth a lot of money to run that training model 24/7,” Dyson told me. “Instead, the opportunity here is to use the money that generates to pay other people to flex their load, or pay other people to adopt batteries or other resources that can help create headroom on the system.”

Both Plante of Voltus and Bhargava of WeaveGrid confirmed that hyperscalers are excited by the idea of subsidizing VPP programs in one form or another. That could look like providing capital to help customers in a data center’s service territory buy residential batteries or contracts that guarantee a return for VPP aggregators like Voltus. “I think they recognize in us an ability to get capacity unlocked quickly,” Plante told me.

Yet another knot in this whole equation, however, is that even given hyperscalers’ enthusiasm and the maturation of VPP technology, most utilities still lack a natural incentive to support this resource. That’s because investor-owned utilities — which serve approximately 70% of U.S. electricity customers — earn profits primarily by building infrastructure such as power plants and transmission lines, receiving a guaranteed rate of return on that capital investment. Successful VPPs, on the other hand, reduce a utility’s need to build new assets.

The industry is well aware of this fundamental disconnect, though some contend that current load growth ought to quell this concern. Utilities will still need to build significant new infrastructure to meet the moment, Bhargava told me, and are now under intense pressure to expand the grid’s capacity in other ways, as well.

“They cannot build fast enough. There’s not enough copper, there’s not enough transformers, there’s not enough people,” Bhargava explained. VPPs, he expects, will allow utilities to better prioritize infrastructure upgrades that stand to be most impactful, such as building a substation near a data center instead of in a suburb that could be adequately served by distributed resources.

The real question he sees now is, “How do we make our flexibility as good as copper? How do we make people trust in it as much as they would trust in upgrading the system?”