Frontier Forum: Unlocking next-generation VPPs

In the mid-2000s, Ben Brown started his career designing demand response programs that relied on pagers and telephones. Today, as Renew Home's CEO, he's leveraging AI and tens of millions of connected smart devices to help households save energy and create an entirely new approach to grid management. Renew Home is building a new kind of virtual power plant that moves beyond occasional emergency events toward continuous, subtle energy shifts across millions of connected households. "The biggest evolution is connected devices," explains Brown, who previously led energy product development at Google after its acquisition of Nest. During his time at Google Labs working on large language models, Brown also witnessed firsthand the massive energy demands that AI would place on our grid. This realization, combined with his work on smart home technology, led Brown to envision a new approach to virtual power plants – one built on subtle, personalized adjustments across millions of homes rather than occasional disruptive events. “There's actually a lot more value continuously throughout the year, over hundreds of hours where customers can save more money by helping support the grid." With DOE projections showing a 200 gigawatt peak on the US grid by 2030, Renew Home's approach offers a compelling alternative to building new power plants. By focusing on customer control and personalization, they've achieved 75% opt-in rates, while creating a resource that is far cheaper than gas peakers. In this episode, recorded as part of a live Frontier Forum, Stephen Lacey talks with Ben Brown about the next generation of virtual power plants. How does Renew Home’s approach differ from demand response or battery-based VPPs? And what role can it play in addressing the grid’s urgent needs? This is a partner episode, brought to you by Renew Home. It was recorded live as part of Latitude Media's Frontier Forum series. Watch the full video to hear more details about next-generation VPPs.

• 0:00Speaker 00

  This is a Frontier Forum, brought to you by Latitude Studios.

• 0:08Speaker 01

  In the mid 2000s, Ben Brown worked at Constellation designing efficiency and demand response programs. DR was gaining traction in the residential and commercial markets, but the resource and the technology behind it looked a lot different than today.

• 0:20Speaker 02

  Well, I think one of the biggest changes, we don't use pagers anymore.

• 0:26Speaker 01

  I understand there are red phones still occasionally, though.

• 0:29Speaker 02

  Yeah, that part I don't work on as much anymore. But yeah, no, I think that the biggest evolution for sure is now more and more that we have connected devices as part of what homes and households have been adopting to help kind of reduce energy costs and have more control over comfort in their homes. And using this as a great way to engage and positively build trust with customers, I think has been one of the biggest changes of the last 15 years.

• 0:56Speaker 01

  There are now tens of millions of smart thermostats and smart water heaters across the US, and they've opened up a vast opportunity for managing demand that wasn't available in those early days. In 2012, Ben went to Google, where he worked on the company's smart home and Wi-Fi platforms. And later, after Google acquired the smart thermostat maker Nest, he led the development of new energy offerings. There, he saw the potential of another emerging tool: machine learning.

• 1:20Speaker 02

  You could see the power of these tools used towards a consumer product experience that was just really unique.

• 1:25Speaker 01

  The original Nest thermostat used machine learning to predict heating and cooling needs in response to some dynamic factors like weather, which set the foundation for Google's demand response offering. And the technology has radically improved since then.

• 1:37Speaker 02

  And that has evolved a lot, obviously, over the last 15 years, where now the capability is of anticipating when electricity prices are going to be going up or down, having unique insight in terms of like the thermal model of your home, having unique insight into like all the complexity around the grid and grid congestion and how that's going to evolve over time. That is exactly the type of things that AI is really, you know, kind of really helps it sing.

• 2:01Speaker 01

  That work brought Ben to Google Labs, where he worked on large language models for energy and finance applications. And that's when he saw the other side of AI: the energy demands.

• 2:15Speaker 02

  And you could absolutely see just how energy intensive these new LLMs were going to be, right? Like, if you think about the idea that maybe as each individual having kind of an AI co-pilot, that's the equivalent of like having hundreds of millions of new HVAC systems all of a sudden being put onto the grid. I mean, obviously, we'll find efficiency, but it is an interesting kind of like analog to think about from just the infrastructural perspective and what it needs to do, what we would need to do as a system to support all these new really compelling use cases.

• 2:45Speaker 01

  Today, Ben is the CEO of Renew Home, a company building a new kind of virtual power plant that helps households save money while offering reliable grid capacity at scale. And he says we're now at a moment when all these converging forces, ubiquitous smart devices, powerful AI, and spiking utility demand are setting the stage for a new generation of VPPs.

• 3:04Speaker 02

  There is going to continue to be, you know, more and more of a challenge as our system tries to struggle to grow to meet that demand, not just on the aggregate basis, but really on the distribution and locational basis. And I think that's also what's very unique about things like virtual power plants or DR versus just citing large central power plants.

• 3:29Speaker 01

  I'm Stephen Lacey. This week, a conversation with Renew Home's Ben Brown. It was recorded live as part of our Frontier Forum series. AI and electrification are driving load growth at levels we haven't seen in decades. The Department of Energy projects we'll see a 200-gigawatt peak on the US grid by 2030. So I caught up with Ben to talk about designing VPPs in this new era. What role can they play in addressing the grid's urgent needs?

• 3:53Speaker 02

  I deeply believe that we're in the infancy of how impressive and how powerful these use cases can get.

• 4:04Speaker 01

  You talked a little bit about the limits of demand response. Can you talk about why you think it offers diminishing returns as a one-size-fits-all solution?

• 4:13Speaker 02

  Yeah, so I think when we think about traditional DR as what was originally constituted maybe 20 years ago, plus where in which we had like direct load Control, the canonical example that like folks think back on is those old AC compressor switches where kind of essentially you'd have something sent out to a system to turn off the compressor. And so a household would be like turning up, turning down the thermostat to get cooler and nothing would be happening and it would just lead towards frustration. And so when we first launched Rush Hour Rewards 12 years ago, a lot of the intent there was to say, hey, if we could prove out that customers with more control, the idea that they can always change the thermostat, adjust, always get to a different comfort set point in their homes.

• 4:59Speaker 02

  With more control, you'd have a much higher opt-in rate and you would be able to get more folks enrolled while still providing similar incentives for peak shifting and peak reduction. That we could, we could achieve much larger rates of folks and through larger numbers, we would actually be able to deal with smaller opt-out rates across the entire population. And that was proven out, I think, really well over the last 12 years with kind of the evolution of DR in terms of what we thought for Rush Hour Rewards and DR 2.0 type of concepts. I think where we're at now, though, is that those were really targeted towards, you know, peak hours of the year, right? You know, maybe 40 to 60 hours a year where we were really just dealing with a really, you know, a heat wave. Now, obviously, we're seeing one, you know,

• 5:43Speaker 02

  more variable weather throughout shoulder seasons. Like the last couple of weeks, we've seen a lot of kind of heat waves or cold waves and that all of a sudden stresses the grid in unique ways. And so there's a lot more hours of the year where you're dealing with these kind of mini peaks on the system that the system really wasn't built for. And so how do we kind of, you know, leverage more of an ongoing, continuous resource to support that? But then I also think that, you know, the idea when we start to incorporate in some of the distribution and locational value of these things, where in which, you know, there's actually a lot more value continuously throughout the year, you know, hundreds of hours of the year, where in which folks can, customers can save more money by helping support the grid in unique ways that would be different from traditional infrastructure, is also really compelling and I think a really

• 6:28Speaker 02

  good, a really good direction for us to be going in.

• 6:31Speaker 01

  Okay, so tell me about how continuous demand response works, these subtle personalized continuous shifts. How does it actually work?

• 6:41Speaker 02

  Yeah, so I think a huge part of it comes down to this idea of control and personalization. At the end of the day, like, as part of its core, specifically for a smart thermostat, it wants to do its job. Its job and its promise to a household is to optimize comfort while minimizing costs. And so when we think about kind of extending that more holistically, traditionally, that was just really around efficiency, because we didn't have time-varying pricing. In the world we're moving to, where in which we have more now time of use rates, and we also have more stress on the system on a continuous basis, where in which households can benefit from reducing that stress and getting rewards for that, it makes sense

• 7:23Speaker 02

  for us to lean much more into this idea of personalization, and really going deeper around, hey, we understand the unique rhythms and habits of your household, like, kind of the flow of your day, and where you're going, and, you know, kind of what your preferences for temperature are during a given day. And we also have anticipation, because you have to remember, like, specifically, heating and cooling is an asynchronous optimization, right? It's something that, like, we want to get the house to the right temperature in 30 minutes or an hour, and so kind of have to be able to project what the next couple of hours of of cost and grid congestion, and your time to temp, or time to heating or cooling your home based on the thermal envelope. All of those things are really well incorporated into something that, you know, through deeper personalization and

• 8:06Speaker 02

  optimization, that we can do that's really targeted to households in terms of their preferences, and minimizing their costs. And so, back to this idea of continuousness, the only way to do that well is by really getting these things right, and being highly personalized for households, so that we can incorporate all these benefits, to be able to deliver the core promise, which is, you're comfortable all the time, and we are minimizing your cost, and helping supporting deeper aspirations that you may have around minimizing your consumption of of carbon-rich energy, or minimizing the stress on the grid, and these other things that you might also care about as well.

• 8:46Speaker 01

  And what are you looking at for opt-in rates so far?

• 8:48Speaker 02

  Yeah, so I think that's one of the biggest things that like at its core, we can talk about these things, right? Like it's a really, it's a cool concept to talk about it in the abstract and I think a lot of the industry has for a long time. And what has been really unique, I think, across the industry is what we've been able to prove out through opt-in rates. And so we have actually seen both opt-in and engagement that we're getting nearly 75% of all new Nest thermostat customers that install a device are opting into energy shifting. And what's great about that is that it's not only that happening in terms of engagement and opt-in, we're also seeing nearly all of those folks years later still engaged and still leveraging it and having really high satisfaction around it.

• 9:29Speaker 02

  And so to me, you know, it kind of proves out the central point that we've been trying to articulate for a decade, which is with users having more control and a better experience in their home, you actually can get to an order of magnitude higher engagement and opt-in rate than what we have seen traditionally, right? If we think about like the traditional demand response programs from 15 years ago, you know, the industry was probably seeing 10 to 20% opt-in rates. And so if we think about a world where we can engage a much larger population and then be able to have personalization, like there still might only be 20 to 30 or 40% that want to kind of maybe trade off more comfort for higher rewards or cost savings.

• 10:12Speaker 02

  And that's great and we can segment out that population, but it means we can, it would be great to still have the other 40 or 50% engaged, even if the optimizations are more subtle.

• 10:21Speaker 01

  So there's this very, sometimes very intense debate in the industry about utility control or customer control, you know, along the spectrum of distributed resources. And you firmly believe that customer control can bring better outcomes. Why?

• 10:39Speaker 02

  I think it's important for us to not look at that choice through a singular lens, which I also think can be somewhat flawed sometimes. So to me, customer control is a feature, not a bug. And the reason I say that is because the more customers feel empowered, the more they feel engaged in this overall part of supporting the energy system, dealing with kind of rising challenges to the system that had been increasing costs for a lot of folks in terms of, you know, the increased demand that we've been seeing across holistically, but also some of the climate-oriented challenges that we've been seeing. And so having customers engaged in part of this is a really important feature

• 11:22Speaker 02

  because they now feel more empowered and the collective part of the solution going forward. And so I think that's just one fundamental thing that I think sometimes gets kind of overlooked. The second important thing in my mind is if we're trying to build a resource, when we think about like the power plant side of the virtual power plant analogy, a lot of it is actually you're trying to build a resource that can perform for 15 years plus, right? If that's kind of us trying to say that a residential aggregated resource is as good as a natural gas power plant. And I think if we're trying to push in that direction, in order for folks to want to engage and be supportive of the system longer term, they have to feel like they're in control.

• 12:06Speaker 02

  If they don't, and we kind of, if we kind of don't have trust in that system and we don't kind of continuously engage and reward the customers around that, then it's one of those things that we could see folks peter out of that over time. And then it does not have the same kind of benefits that you would think about from a power plant in terms of longevity. And then I think the last part about it is the whole benefit of a distributed system, you know, from like, you know, kind of all of our backgrounds and statistics and whatnot, is like the power of large numbers, which is the reality is when you give individuals control or a long system and you have hundreds of thousands of customers or a million customers, you know, stochastically, we can predict, you know, how these things will perform in aggregate. And so actually having control the entire time

• 12:52Speaker 02

  leads towards less of these kind of asymmetric outcomes where all of a sudden something happens because people are in control and it's more stochastically predictable versus another outcome which would go the other direction.

• 13:03Speaker 01

  So there are many models that are approaching batteries as core to virtual power plants and you really think that HVAC systems represent a really important resource. So you talked about why you're focusing on HVAC first, that you're expanding to all these other connected devices in the home. How do you see these grid interactive devices evolving over the next five to 10 years as part of the foundation of the VPPs you're developing?

• 13:30Speaker 02

  You know, we tend to sometimes get really excited by the, you know, kind of silver bullet technologies, or we get excited about something that feels like is a new, a new evolution that is not, actually has not had a, you know, massive adoption yet, but feels like if it did have adoption, would have a bunch of, you know, beneficial characteristics in terms of supporting the, the needs we have going forward. This is a all hands on deck moment. We have, you know, it's very different than where we've been for the last couple of decades. As you mentioned, we have mostly had flat demand. And so now as demand is increasing, increasing substantially, and is going to increase substantially for the coming decade, you know, we really do need to look at all of the solutions that

• 14:11Speaker 02

  are in front of us. And so to me, the idea of not incorporating what is by far kind of the largest existing set of resources and infrastructure we have in the country, which is the 80 million plus central HVAC systems and 80 million plus hot water heating systems, that is existing thermal storage infrastructure that exists across our country, would just be kind of insanity to me, given the size and the scope of the problem. It doesn't mean that as we are adopting EVs at scale and as we are adopting more and more battery storage at scale on the resi side, that we shouldn't of course incorporate those. Those are going to be critical resources. They have a lot of different benefits and

• 14:54Speaker 02

  characteristics that does kind of thermal storage from HVAC or from hot water heating. But we need to leverage all of those because, you know, from the benefits of both the scale that we need and the adoption curves that we have to ride, it is important that we're having a solution that is, you know, benefiting from the fabric of those things working together in concert.

• 15:14Speaker 01

  I presume a VPP is a lot more cost effective than a natural gas peaker and a lot faster to deploy. You're looking at a moment when it's really impossible to get equipment. Goodness knows what's going to happen with tariffs to make build out more expensive. So tell me about comparing a VPP to like, say, an alternative gas peaker option for a utility. How should they be evaluating these things against each other?

• 15:40Speaker 02

  Yeah, so I think that there's three things. So one is time to market, which we've talked about. So like if you're, you know, if you're still waiting on a turbine for until 2032, that's one. Two is definitely cost effectiveness. You know, if you look at the Brattle report, it's a good example of that. We believe that a VPP similar capacity resource would be 60% of the cost of a natural gas peaker investment. And honestly, that's probably outdated because now we know that natural, building natural gas peakers are even more expensive over the last few years just because of the wait times and some of the additional costs that are going into it. And it would be, and a VPP would also be 40% of the cost of a utility scale battery investment. So we believe those, you know, for the first

• 16:25Speaker 02

  two, timeliness, cost effectiveness. And honestly, the third thing, which I think also just gets overlooked a lot, is the distributed nature of the resource. We've talked about that, and that's actually something that is not getting incorporated nearly enough. But we all know a huge part of our challenge going forward is updating the distribution infrastructure that we'll need on an ongoing basis, especially with new use cases coming online with more, you know, EV charging on the edge, more home electrification. These are challenges that are definitely going to be a key part of our distribution challenge going forward. And so I think that the decentralized nature of a VPP represents, when engaged properly

• 17:07Speaker 02

  and incentivized properly with homeowners, represents a huge benefit that honestly will be dependent on, you know, going forward.

• 17:15Speaker 01

  So you've got a plan to build out a gigawatt of capacity in Texas. What are the phases of scale for building that much capacity?

• 17:23Speaker 02

  You know, what's beneficial about the partnership we have with NRG in Texas, which we announced earlier this year, is that, one, we have a large existing customer base of Nest thermostat customers as well as other platform partners that we have, in addition to, you know, obviously a large, huge, large base of NRG retail customers in Texas. And so what's nice about it is that when fully engaged and brought together, that already represents a fairly large resource of where we're starting from in terms of, you know, getting towards that one gigawatt goal. But a huge part of getting to the one gigawatt goal will also be, you know, starting to give access to low-cost and no-cost thermostats, smart thermostats,

• 18:05Speaker 02

  across, you know, NRG's existing user base and growing user base. Also deploying other targeted technologies, whether that's batteries, EV chargers, really thinking through how we support folks along the home electrification journey. And so, you know, we definitely have a great scale of households and engaged households across Texas. And so a large part of what we're going to be pushing as quickly as we can on with a great partner in NRG is really just distributing these technologies to folks.

• 18:32Speaker 01

  So what has changed the most since you've been working on this in terms of how utilities and grid operators view these resources?

• 18:43Speaker 02

  So, I mean, I think, you know, the last 15 years has been, first phase, which we talked about, was proving, when we think about when we launched Rush Hour Awards, then it was about proving to utility partners that if we gave more control to users, we would actually build a much larger population of folks that were enrolled. That would be a much more reliable resource. I think we approved that up very effectively, getting to over 1.5 million thermostats enrolled in programs, over 100-plus utilities. I think that was a really important phase of like moving the model from non-controlled resources to a place where customers had control. So I think that was the first model. I think the next step of where we've been proving out and why

• 19:27Speaker 02

  we launched Renew was proving out that small subtle energy shifts with deep personalization across, you know, millions and millions and millions of households would actually have a even bigger benefit, both from a peak perspective, but also from a continuous perspective. And to me, like, that is a huge part of what we've been engaging with partners on over the last two years, which is, hey, let's prove out that this resource has huge benefits that are not necessarily what we had thought about before. That if we think about an even larger population, the 75% of folks that are getting into these, you know, now that are enrolling in energy shift, how do we think about kind of a more deeply personalized,

• 20:09Speaker 02

  continuous way of engaging with those folks that provides a much larger resource across the country? So like a good example on that one, and we've talked about this, the reason we've been really public about the numbers is to try to help with this, you know, kind of this chicken and egg question. And we've been really clear that saying, hey, we have over 5 million households that have enrolled, which represents a capacity resource that is shiftable across the country. So, you know, not just in Texas, not just across the programs that we've already enrolled folks in, but there's, there's capacity sitting across on the sidelines across the country. And I think the reason we've been public about that is just saying, hey, we want to let you know that most likely in your market, there are

• 20:54Speaker 02

  probably 50 to 300 megawatts of resource ready to go. We'd love to engage with you on it. And I think that, that to me is really important because a lot of what we've been talking about in the industry is if we support incentives for programs and we build certain capability sets with technology, then eventually we can bootstrap our way to a meaningful resource. And as we all know, Unless we're talking in the 25 to 50 plus megawatts in terms of actual resource, it's harder for energy partners to see it as meaningful. And so the scale is really critical to speak to. And that's kind of why, right out the gate, we've just been like, no, we already have a lot of scale.

• 21:39Speaker 02

  And now we want to scale it even more so that it is meaningful enough for energy partners to engage with.

• 21:46Speaker 01

  All right, so as we wind this down, I want to get your thoughts on either the tech or adoption trends that you think are going to shape the industry going forward. So you've been at this for a couple of decades now. What do you think is going to drive growth or change over the next half decade or decade?

• 22:03Speaker 02

  So I think thematically, we've talked about, I've talked about control. Could a good bit around like user needs for more control over our energy costs, more control over our homes. And I actually think control is a good theme. Control and empowerment is a good theme to think about when it comes to like technology over the next decade, because when we think about generative AI, my view, having worked on this stuff, is that most everyone is developing tools that are meant to augment our capabilities and give us actually more control as people doing work, people building businesses, people trying to operate their, you know, manage their costs in their homes, that like generative AI is a massive augmentation to skills.

• 22:45Speaker 02

  It's a great, very powerful tool that actually gives us more control and more empowerment than somehow trying to take it away from us. And so I think that is really critical, because I think generative AI could, you know, I think people worry that it would go the other direction. I actually think a lot of folks are investing in it. I actually think it's the same thing when it comes to its potential effects on costs for customers and costs for the grid, that we have to think about how we drive empowerment and control around giving households more control of their energy costs because of the effects of growing demand. That is another key thing that I think that if we don't push in that direction over the next decade of giving more control, more empowerment with solutions to customers,

• 23:28Speaker 02

  folks will get more and more unhappy with rising energy. Cost. And so I think there's like a really good trend there around giving more control, more empowerment with more demand side resources to households to help them navigate kind of a growing demand environment that we just haven't had to design for historically. And then I would say just lastly, I really believe because of the growing demand environment that the partnerships between, I actually feel like, energy companies, utilities are going to have better, deeper relationships with customers. I actually think it will go in the other direction that we've seen trends wise historically, where I actually think there's a huge opportunity to build, to build really positive reaffirming relationships between energy companies and customers because of this more control over costs, more, more

• 24:13Speaker 02

  empowerment to customers with how they're leveraging technology in their home to, to mitigate costs and to have more control over the experience in their home, like comfort. So all these things to me is like, these are me looking at these challenges through an optimistic lens, because I feel like that's what we're in the business of, is trying to build solutions here. Obviously, if we don't go in that direction, the, the other side of that coin can be very pessimistic, and I'm, so I prefer not to focus on that one. But I do think that these trends are coming. Like, I, I think there's a lot of folks who are, you know, maybe don't, aren't as aspirational about how much AI will grow in terms of its use cases. I haven't worked on this stuff. I deeply believe that we're in the infancy of how impressive and how powerful these use cases can get in terms

• 24:58Speaker 02

  of tools and customer empowerment and use cases. So I just think that's like, that is going to happen. I also believe that like, the benefits of home electrification technology around safety, around cost mitigation, around things like a better driving experience, which a lot of EVs just provide, these things are also going to be adopted. And so, these things are coming, and I think it's a great opportunity for all of us in the industry to be a huge part of that solution.

• 25:26Speaker 01

  Ben Brown, CEO of Renew Home. Thank you so much. I really enjoyed this.

• 25:30Speaker 02

  Likewise, even thinking so much for having me.

• 25:34Speaker 01

  This conversation was recorded live as part of Latitude Media's Frontier Forum with Renew Home, and there is so much more. This is an edited version of the conversation. If you want to watch the full video with lots of listener questions about the design and operation of VPPs, head on over to latitudemedia.com/events and click "Watch Recording." There's a lot more technical conversation there. And if you want to learn more about partnering with Renew Home, visit renewhome.com. Thanks so much for listening.