This was an excellent episode, EnergyHub and SFT are doing interesting & important work, but as the saying goes: "if all you have is a hammer, then everything looks like a nail", and the "hammer" here is the VPP model.
What is the real goal? IMO the goal is to transition to a grid where all grid users (and their loads and DERs) coordinate with the grid in order to lower overall grid infrastructure costs, and minimize/optimize their cost & use of electric power. VPPs may be a means to that end, but in no way is the ongoing existence & role of VPPs a goal in and of itself.
The recent Volts interview with Saul Griffith contained this gem:
David Roberts: You think prices can do the job that VPPs are trying to do now? I have people who think that that are going to be on the pod sometime in the next couple of months.
Saul Griffith: They may be correct. There are people I know that you probably are going to have one of my mates on who believes that you should have a negotiated price between the toaster and the oven so that you truly optimize all the way down to the appliance. You can go mad.
Maybe the most liberal possible market is the right answer. Let’s talk about gatekeepers — not about price signals. You can do the calculation in a single spreadsheet cell that does everything that a VPP does. Does the grid need electricity now? Yes, discharge. This is not a hugely computational thing. You have to ask yourself, why do I need aggregation? All the VPP does is charge you a fee so that somebody who can write software can aggregate your home with a whole bunch of other homes to trade the grid. I don’t believe that step is necessary.
It adds a gatekeeper and it adds a rent seeker. If you had a well-designed electricity market that was fair, this stuff would just happen automatically at your smart meter or through your retailer or utility company. A VPP might be the mechanism — how you get from the past to the future — but it’s a band-aid that adds a rent seeker. Let’s just say that.
==== End of transcript snippet =====
IMO an accurate summary of Griffith's position is this: VPPs are unnecessary, rent-seeking, gatekeepers. I completely agree with that.
SFT's arguments against price seem incomplete, and possibly self-serving. He states that grids/operators that provide highly dynamic pricing don't result in successful VPPs, as if VPPs again are the goal/metric. They aren't! Are grid customers (and their loads/DERs) responding succesfully to these dynamic prices? I think the data supports that is the case. Richard Hart's comment below supports this also. And SFT seems to claim that price doesn't work, again, the excellent Volt's interview with Kate Slevin on NYC's congestion pricing program clearly shows that indviduals respond crisply to price signals, and congestion pricing is an amazing success. I'd like to see the data that indicates the opposite. And prices are not complicated. Everybody deals with prices numerous times per day, for every transaction they make. The real time (wholesale) price of electricity is easily available (e.g. gridstatus.io).
LSEs (utilities) have a direct billing relationship with each of their customers, and for the sake of argument, all utilities host a website that each customer has (or can) have a login account, in order to establish/terminate service, obtain their bill, pay their bill, and compare available rate/tariff plans and "programs" (e.g. demand-response, etc). There is no reason that these utillites can't provide Internet-accessible price-servers that their customers can use to obtain real-time pricing information, which will be used by IoT loads/DERs in combination with the expressed preferences/constraints of the user to automatically coordinate with the grid, to accomplish the top-level goal I mentioned above.
SFT remarks that VPPs florish in highly-regulated utility environments, again, as if that is a good thing! The reason being is that because utilities don't provide public price-servers using open/standard protocols (i.e. OpenADR 3) only VPPs that have the resources to make business deals with both appliance OEMs (Resideo, Nest, etc) and can undertake custom integrations with the utilities management systems. This is a "bug", not a "feature".
What we need are open/standard protocols that enable grid-coordination, supported both by grid-operators (utiliies) and their customers, again OpenADR 3 being an excellent example. Within the home, we need to push for open/standard protocols for appliances/DERs, Matter being a great candidate, and OpenADR 3 could also be used directly also/instead. What are VPPs like EnergyGrid doing to make that happen? VPPs can and should use these open/standard protocols, and it should be the customer's choice to coordinate directly with the utitility, or via an aggregator (VPP). VPPs that provide true value will florish, and VPPs that are only "margin-added resellers" won't. Case in point: anybody can run their own email system in their house on a $35 Raspberry Pi, but almost nobody does, they use Gmail (an aggregator!), because Google provides value and ease-of-use to its users. Open/standard grid-coordination protocols should support the VPP model, but they absolutely should NOT REQUIRE a VPP.
I agree with SFT that in addition to balancing grid-wide supply/demand ("macro" management), managing the distribution network, down to the individual distribution transformer and the N customers connected to that transformer is also important. This is the "micro" management challenge. Price is a great solution to the macro challenge, and we need to enhance grid-coordination protocols to include operating envelope constraints (import and export limits). OpenADR 3 has some operating-envelope support today, and enhancements are under consideration. As we add that support, grid-coordination will evolve from the management of individual loads/DERs to operating envelopes and dynamic pricing. The grid/utiility should not be concerned with each load/DER in a home, it should say: the price now is P, your import limit is I, and export limit is E. There should be a dynamic price for both I and E, P(I) & P(E) and ideally, they would be the same (one bidirectional price). There is no way that the grid operator, or a VPP will understand how best to optimize the power/energy usage of an individual home, there are many loads, possibly multiple BESS and V2H-EVSE-EVs, PV, and the detailed occupancy and usage patterns of the home's occupants. The level, detail, and privacy-implications of the information needed to make the best coordination decisions is only available locally, within the home itself, not in a VPP's cloud, or by the utility operator's cloud.
Want to build a shiny new data center next to Hometown USA? Finance community solar parking lot microgrid construction at all existing large apartments, condos, shopping centers, business parks, schools & municipal facilities in town, within 5 years. That's already happening in France & South Korea. Payback within 8 years. "Doubling down on distributed solar provides a predictable path for decarbonization while lowering costs for all ratepayers regardless of whether their home has a solar installation" according to a recent study of distributed solar in NY state. The primary reason we're lagging China & Europe on EV adoption, besides lack of affordable EVs, is our lack of widely distributed Level 2 EV charging. When anyone can easily charge while at work, or daytime shopping, in addition to night time hours, that problem is solved, and big EV batteries can fortify local distribution grids during peak demand & grid outages.
David asked Seth whether utility data was an obstacle and he said it wasn’t. David noted this contrasted with what he’s heard from others. I’d like to clear this up - we talk about VPPs as a monolithic thing, but there’s actually utility-facing VPPs (the niche energyhub focuses on) and ISO-facing VPPs (CPower, Voltus, etc). When it comes to data, there’s a world of difference - utility programs have whatever they need, since the utility gives themselves access to the meter data they collect. For the ISO programs, meter data is a never ending challenge, since the utilities don’t share in the value generated but are the ones who possess the meter data.
This should matter to everyone, if we want VPPs to fully solve transmission level problems in deregulated markets, which I think we do!
I think the concept of using VPPs to help stabilize the grid and reduce the amount of new infrastructure we need is great. I do have some questions and concerns, namely:
1. How much electricity is saved by changing a thermostat a couple of degrees?
2. In the Northeast, most heating is oil or gas-do they use enough electricity to be worth connecting to a VPP?
3. Since a given local area is all in the same time zone, and usually experiencing similar weather, is there enough variation in consumption and production to make the VPP practical--i.e., if most of the EVs are in use, or charging, at the same times, then they're not available to provide power to the VPP
4. If a house has solar panels and net metering, there's no financial benefit to having batteries-the extra power gets sold anyway--so I think the incentives would have to pretty much cover the full cost of the batteries to make it worth while
5. I think that security is a potential problem. We know that foreign countries are already targeting our utilities. I think that any system for sharing utility data, and enabling remote control of HVAC and other electrical systems, needs not only to have extremely robust defenses against hacking, but also to plan for the near certainty that, sooner or later, it will get hacked, and have a Plan B defense to minimize the damage when that happens.
David, can you check back in with Form Energy? News stories say they cranked up production last fall, and have already made tens of megawatt-hours in sales, plan on a big 2026.
Did they make it down to $20/kWh? Do they expect to?
Cheap-enough batteries, as per your own Vox story of 2019 that covered the MIT study saying $20/kWh was cheap enough to let intermittent handle base-load.
And are Form frightened by sodium-ion, which may get down into the same price range?
Another excellent episode updating us on both what's already being accomplished and what's left to get done to reach the modern grid.
One note of caution on getting hyperscalers to come bearing gifts instead of a bandit's mask, electricity theft is only half of the banditry problem, water is the other maybe bigger problem.
Someone needs to figure out how to turn district HVAC systems or something else into at least a partial solution to the water (used to cool the racks of humming servers) problem.
Did I completely miss it or is Energy Hub basically doing the same thing as the Kraken platform from Octopus? Competition is fine, but I hope someone is developing a universal standard that everyone abides by so we don’t end up with the nightmare we had with IoT before the Matter standard.
This was an excellent episode, EnergyHub and SFT are doing interesting & important work, but as the saying goes: "if all you have is a hammer, then everything looks like a nail", and the "hammer" here is the VPP model.
What is the real goal? IMO the goal is to transition to a grid where all grid users (and their loads and DERs) coordinate with the grid in order to lower overall grid infrastructure costs, and minimize/optimize their cost & use of electric power. VPPs may be a means to that end, but in no way is the ongoing existence & role of VPPs a goal in and of itself.
The recent Volts interview with Saul Griffith contained this gem:
David Roberts: You think prices can do the job that VPPs are trying to do now? I have people who think that that are going to be on the pod sometime in the next couple of months.
Saul Griffith: They may be correct. There are people I know that you probably are going to have one of my mates on who believes that you should have a negotiated price between the toaster and the oven so that you truly optimize all the way down to the appliance. You can go mad.
Maybe the most liberal possible market is the right answer. Let’s talk about gatekeepers — not about price signals. You can do the calculation in a single spreadsheet cell that does everything that a VPP does. Does the grid need electricity now? Yes, discharge. This is not a hugely computational thing. You have to ask yourself, why do I need aggregation? All the VPP does is charge you a fee so that somebody who can write software can aggregate your home with a whole bunch of other homes to trade the grid. I don’t believe that step is necessary.
It adds a gatekeeper and it adds a rent seeker. If you had a well-designed electricity market that was fair, this stuff would just happen automatically at your smart meter or through your retailer or utility company. A VPP might be the mechanism — how you get from the past to the future — but it’s a band-aid that adds a rent seeker. Let’s just say that.
==== End of transcript snippet =====
IMO an accurate summary of Griffith's position is this: VPPs are unnecessary, rent-seeking, gatekeepers. I completely agree with that.
SFT's arguments against price seem incomplete, and possibly self-serving. He states that grids/operators that provide highly dynamic pricing don't result in successful VPPs, as if VPPs again are the goal/metric. They aren't! Are grid customers (and their loads/DERs) responding succesfully to these dynamic prices? I think the data supports that is the case. Richard Hart's comment below supports this also. And SFT seems to claim that price doesn't work, again, the excellent Volt's interview with Kate Slevin on NYC's congestion pricing program clearly shows that indviduals respond crisply to price signals, and congestion pricing is an amazing success. I'd like to see the data that indicates the opposite. And prices are not complicated. Everybody deals with prices numerous times per day, for every transaction they make. The real time (wholesale) price of electricity is easily available (e.g. gridstatus.io).
LSEs (utilities) have a direct billing relationship with each of their customers, and for the sake of argument, all utilities host a website that each customer has (or can) have a login account, in order to establish/terminate service, obtain their bill, pay their bill, and compare available rate/tariff plans and "programs" (e.g. demand-response, etc). There is no reason that these utillites can't provide Internet-accessible price-servers that their customers can use to obtain real-time pricing information, which will be used by IoT loads/DERs in combination with the expressed preferences/constraints of the user to automatically coordinate with the grid, to accomplish the top-level goal I mentioned above.
SFT remarks that VPPs florish in highly-regulated utility environments, again, as if that is a good thing! The reason being is that because utilities don't provide public price-servers using open/standard protocols (i.e. OpenADR 3) only VPPs that have the resources to make business deals with both appliance OEMs (Resideo, Nest, etc) and can undertake custom integrations with the utilities management systems. This is a "bug", not a "feature".
What we need are open/standard protocols that enable grid-coordination, supported both by grid-operators (utiliies) and their customers, again OpenADR 3 being an excellent example. Within the home, we need to push for open/standard protocols for appliances/DERs, Matter being a great candidate, and OpenADR 3 could also be used directly also/instead. What are VPPs like EnergyGrid doing to make that happen? VPPs can and should use these open/standard protocols, and it should be the customer's choice to coordinate directly with the utitility, or via an aggregator (VPP). VPPs that provide true value will florish, and VPPs that are only "margin-added resellers" won't. Case in point: anybody can run their own email system in their house on a $35 Raspberry Pi, but almost nobody does, they use Gmail (an aggregator!), because Google provides value and ease-of-use to its users. Open/standard grid-coordination protocols should support the VPP model, but they absolutely should NOT REQUIRE a VPP.
I agree with SFT that in addition to balancing grid-wide supply/demand ("macro" management), managing the distribution network, down to the individual distribution transformer and the N customers connected to that transformer is also important. This is the "micro" management challenge. Price is a great solution to the macro challenge, and we need to enhance grid-coordination protocols to include operating envelope constraints (import and export limits). OpenADR 3 has some operating-envelope support today, and enhancements are under consideration. As we add that support, grid-coordination will evolve from the management of individual loads/DERs to operating envelopes and dynamic pricing. The grid/utiility should not be concerned with each load/DER in a home, it should say: the price now is P, your import limit is I, and export limit is E. There should be a dynamic price for both I and E, P(I) & P(E) and ideally, they would be the same (one bidirectional price). There is no way that the grid operator, or a VPP will understand how best to optimize the power/energy usage of an individual home, there are many loads, possibly multiple BESS and V2H-EVSE-EVs, PV, and the detailed occupancy and usage patterns of the home's occupants. The level, detail, and privacy-implications of the information needed to make the best coordination decisions is only available locally, within the home itself, not in a VPP's cloud, or by the utility operator's cloud.
Want to build a shiny new data center next to Hometown USA? Finance community solar parking lot microgrid construction at all existing large apartments, condos, shopping centers, business parks, schools & municipal facilities in town, within 5 years. That's already happening in France & South Korea. Payback within 8 years. "Doubling down on distributed solar provides a predictable path for decarbonization while lowering costs for all ratepayers regardless of whether their home has a solar installation" according to a recent study of distributed solar in NY state. The primary reason we're lagging China & Europe on EV adoption, besides lack of affordable EVs, is our lack of widely distributed Level 2 EV charging. When anyone can easily charge while at work, or daytime shopping, in addition to night time hours, that problem is solved, and big EV batteries can fortify local distribution grids during peak demand & grid outages.
David asked Seth whether utility data was an obstacle and he said it wasn’t. David noted this contrasted with what he’s heard from others. I’d like to clear this up - we talk about VPPs as a monolithic thing, but there’s actually utility-facing VPPs (the niche energyhub focuses on) and ISO-facing VPPs (CPower, Voltus, etc). When it comes to data, there’s a world of difference - utility programs have whatever they need, since the utility gives themselves access to the meter data they collect. For the ISO programs, meter data is a never ending challenge, since the utilities don’t share in the value generated but are the ones who possess the meter data.
This should matter to everyone, if we want VPPs to fully solve transmission level problems in deregulated markets, which I think we do!
I think the concept of using VPPs to help stabilize the grid and reduce the amount of new infrastructure we need is great. I do have some questions and concerns, namely:
1. How much electricity is saved by changing a thermostat a couple of degrees?
2. In the Northeast, most heating is oil or gas-do they use enough electricity to be worth connecting to a VPP?
3. Since a given local area is all in the same time zone, and usually experiencing similar weather, is there enough variation in consumption and production to make the VPP practical--i.e., if most of the EVs are in use, or charging, at the same times, then they're not available to provide power to the VPP
4. If a house has solar panels and net metering, there's no financial benefit to having batteries-the extra power gets sold anyway--so I think the incentives would have to pretty much cover the full cost of the batteries to make it worth while
5. I think that security is a potential problem. We know that foreign countries are already targeting our utilities. I think that any system for sharing utility data, and enabling remote control of HVAC and other electrical systems, needs not only to have extremely robust defenses against hacking, but also to plan for the near certainty that, sooner or later, it will get hacked, and have a Plan B defense to minimize the damage when that happens.
David, can you check back in with Form Energy? News stories say they cranked up production last fall, and have already made tens of megawatt-hours in sales, plan on a big 2026.
Did they make it down to $20/kWh? Do they expect to?
Cheap-enough batteries, as per your own Vox story of 2019 that covered the MIT study saying $20/kWh was cheap enough to let intermittent handle base-load.
And are Form frightened by sodium-ion, which may get down into the same price range?
Another excellent episode updating us on both what's already being accomplished and what's left to get done to reach the modern grid.
One note of caution on getting hyperscalers to come bearing gifts instead of a bandit's mask, electricity theft is only half of the banditry problem, water is the other maybe bigger problem.
Someone needs to figure out how to turn district HVAC systems or something else into at least a partial solution to the water (used to cool the racks of humming servers) problem.
I'm surprised that Seth Frader-Thompson didn't mention Denmark, where dynamic residential pricing has been shown to modify household behavior. There's a terrific study here -https://www.nationalbanken.dk/media/zrmbevi2/price-response-in-residential-electricity-demand-evidence-from-danish-smart-meter-data.pdf. It would be so interesting to have someone from Denmark's EnergiNet (the ISO) on Volts. They have a national energy database that is tapped into by Service Providers as opposed to the balkanized system we have in the U.S.
Did I completely miss it or is Energy Hub basically doing the same thing as the Kraken platform from Octopus? Competition is fine, but I hope someone is developing a universal standard that everyone abides by so we don’t end up with the nightmare we had with IoT before the Matter standard.