 Good morning from Stanford University. My name is Will Chu. I'm the faculty co-director of the StorageX Initiative. It's my great pleasure to welcome everyone to today's StorageX seminar. So today, we are delighted to host two speakers, Professor Dan Kamen from UC Berkeley and Professor Megan Madder from Stanford. So let me briefly introduce Dan. Dan is a distinguished faculty from UC Berkeley, and he has trained many of the world leaders today on energy technology and policy. And I just want to briefly highlight Dan's contribution on his service to the country and to the world. Not only is a world expert on decarbonization and sustainability, he's also serving many roles. For example, as the senior advisor for energy innovation at USAID, he is also the coordinating author for the IPCC, which received the Nobel Peace Prize in 2007. And he also has served as the science envoy under Secretary of State John Kerry. So Dan, we're really, really delighted to have you today as an honor, and we're looking forward to your talk. Well, thank you so much for having me on. Thanks to the audience and really the whole Stanford team. And as a short-term Stanford graduate student, I feel a little connection and pride in what you guys are doing. But I admit I am teaching across the bay. And for those who maintain the Stanford Berkeley rivalry, I'm certainly the other side of the friendly partnership. But I was a PhD student in physics actually at Stanford. So there's several things that I'm going to try to encompass together in my comments here. And really, the overarching one is the one that William already highlighted, that I won't be quite technology agnostic, I have to admit. But I'll really try to highlight different ways that storage is not only evolving and innovating on the chemical or potential energy or fuels side, fuels being hydrogen, as well as applications that range from large utility scale grids, parts of the United States, China, but specifically also what sometimes gets forgotten, and that is the energy access. And in particular, as I'll see later on, the health access part of the story. So this will be a more wide-ranging version. The chemistry and physics of some of the storage systems, well, you'll see them as some of the references and papers. But I'm going to really focus on systems thinking. And so anybody who wants either to access the materials, both the technical papers, some of the interviews, we just did a big series on CNN and then in the Guardian. And you can find that on the website of my laboratory. That's at the top. That's rail, the renewable and appropriate energy laboratory. And then the tweet feed at the bottom is not my own individual things. That's generally images, comments from myself and the laboratory team. But we use my name to tag them. So it's Dan, underscore Camman. And as we think about these projects, just giving you a shout out to my university team, these are all students, both graduate students and undergraduates in my laboratory. And it gives you a feeling for the kind of places we work just to highlight. And because you'll hear the names later on a little bit, Bemi Akinsepe is actually on leave working for the United Nations. She's working in the vice president of Nigeria's office on their decarbonization plan. John Arou is a South Sudanese energy researcher. And actually I was in South Sudan last week and you'll see some manage these storages. Alexander Grayson, Annalise Gilwile and Jess Cursey are all PhD students in my lab working on different aspects of either onshore storage, deployment, electric vehicles in one case, another case, hydrogen, another case, integration with the offshore industry. Jasmine McAdams is working on energy justice. Sam Miles is working on clean energy for clinics, particularly in the Congo. Nidhila Rasuliani, a visiting graduate student from the Ministry of Indonesia, a country that just signed a major clean energy decarbonization pact with the US, UK and German governments. Lin Shio is a graduate student from Kenya. Stephen Stack from Ireland. Hilary Yu is actually long-term living in the field in Eastern Congo. And then Sunay Dagli, Erma Duran and Brooks O'Brien are all working on different aspects of materials decarbonization and material science. There's gonna be a range of people involved in this project. And I don't think anyone needs, on this call, to really see again or be reminded of the generic picture of the historical emissions, the black we have done, the range of current policies that if we did them all and that is a massive if would only hold us into the two to three degree C range, an unacceptably high range. The commitments that countries have made up through COP 27 in Sharma Shake would limit it to about two and a half degrees. And then various pledges and targets would bring us down somewhat. The optimistic pathway really gets us to only 1.8 and that's it, everything goes right. And so far, I would say almost nothing has gone right. And then the pathway that we endorsed at the IPCC is the 1.5. And of course, that involves carbon negative strategies are now unfortunately a must because we waited so long. So that's the context. And what that means, of course, in practice is that all economies need to decarbonize. And I think people on the call will certainly disagree with me on this one, but to my calculations, which are pretty extensive, that means that any investment in gas going forward, any investment in gas is a risky and in many cases unwarranted investment. And that's largely because of the massive progress that's been made on clean energy and energy storage. And of course, unpacking that statement will reveal all our political views and differences. But it's important to state that when we think about the massive exploration efforts in Africa by US and European companies and Chinese companies, the efforts for gas in Southeast Asia, these as profitable as they may sound and as much as people want to argue that these will be turned off in some organized manner that we map. This is really a very, very bad bet. You would not make that bet if it was your healthcare that we're talking about. But somehow in an energy world, where most of us my age and older will not be around to see all of the worst of the damages. It is a really problematic decision, but it's one that we are making. Again, just to re-emphasize the story, we've been having these meetings now. I have been to 17 of the COP meetings of the 27, probably not proud of, and the climate change has continued because we've talked a great deal, but we have not done a great deal, which is really why this moment of speed and scale is so critical. The last piece of framing that you'll see coming back in multiple times here is the one that for many of us is the most uncomfortable. And that is the richest 1% of people on the planet contribute directly to 15% of emissions and the richest 10% contribute over half of global emissions, rich in the US, rich in Ghana, rich in Sudan, rich in Nigeria, Japan, South Africa, China, et cetera. The poorest half of humanity contribute less than 7%. And that's really the place where if you are gonna put fossil fuel investment in place before we think about the decarbonization of the storage pathway, that's where it needs to go. The problem is that when we make that argument and install these fossil fuel systems, generally the poor don't access them anyway. It's generally for industry, whether it's in South Africa or Southern California, they could probably afford to be on the leading edge anyway. Again, those are kind of complex words in terms of the politics. Jumping to the world that I feel a little more comfortable in, however, one of the tools that my laboratory has built to enable this conversation is an open source model called Switch, roughly stands for solar and wind transmission integrated with conventional power. Switch, you can see the website here, just Google my name, Berkeley Switch, it'll come up. And this is a now 10 year long project to build open source and open access models of power systems. In the vernacular, these are capacity expansion models, meaning looking from the power system in a given country or region going forward. And what we do is to minimize the net present value of the sum of all the assets. And by assets, what I mean are the capital cost and the operational cost of existing plants, both thermal and renewable projects. We generally have a reserve margin, you can see in the second line here, reserve margin for some places is as large as 15%. I know the places as small as four or 5%. And all capacity not used directly, we used to spill. Now, thankfully our equation of state is energy is equal to energy used directly, plus energy stored. And so we can either allocate energy resources to directly operate the energy system or to put it into storage. You can see the spill factor, which nowadays we can send to zero because we can capture it all if we have enough storage installed. This modeling also models the cost and the time to build new transmission lines, technologies that are very slow to build, such as nuclear plants or DC transmission, and those that are very rapid to do, such as lithium ion, lithium phosphate batteries, solar renewable energy assets. We also price in some energy efficiency maneuver, operations and maneuvers, but we'll get into that one today. So far, the switch model has been built and deployed in Western North America. We started with California, but we've modeled the whole WEC region, the Western Electricity Coordinating Council. We have switched for Mexico, for Nicaragua, for Chile, for the whole East African power pool, although we have a focus on Kenya, Uganda and Tanzania. The India version is not actually one that my laboratory is working on. Former graduate student who's at Professor UC Santa Barbara works on that. And then our biggest team is working on the China version. In Japan, I should have colored in and we released the Japan switch model last year. The ranges of partners are the ones that you'd expect when you build a modeling tool like this. We don't do it if we do not have a clear partner inside the national government. And so the Ministry of the Ministry of Mines, Ministry of Environment will always be a partner, a number of businesses, some that are Bay Area based and some that are local. CUBE is a mini-grid company in Norway that installs in East Africa. Varunga Power is the micro-hydro energy company spun out of Varunga National Park. Africa's oldest national park, Natel Energy is a Bay Area company that does micro-hydro, a number of energy councils. Just to give you a feeling of the range of the projects we do. As we think about pathways, essentially if you can dream it up and it makes some technical sense, we can build it. So SunShot was the US program to get solar lower than $1 a watt for large commercial systems and $1.50 a watt for small systems. That's a goal that has been achieved. We have scenarios with low-cost batteries. That'll be the one I'll highlight here. But we also look at those where nuclear plays a large role, where CCS works at scale, where we take into account the methane leakage that we observe from actual gas fields and pipelines where transmission costs are high and low, where we zero out or limit the amount of hydro to the environmental regions such as in California. But it gives you a feeling for the types of scenarios. And the outputs for all these models, this is kind of a typical, this one is actually for Western North America. So the black line is the demand for energy in 2050 consistent with a 1.5 degree scenario. And the colors I hope make sense, light blue is wind, yellow is solar, dark blue is hydro, gas with capture is this green color. And the most interesting one for us today is the orange is energy released from storage, the solid orange here and the negative going spikes below are energy going into storage. And it's really that landscape that we'll talk about as we go forward here. I think the most interesting feature is that we initially built this model to think largely about solar deployment to meet some of California's zero carbon numbers back in 2012, 2013, but increasingly we moved into storage. And so one of the big initial successes of using the switch model was that we had us sit down with all three California IO utilities, Pacific Gas and Electric, Southern California Edison and San Diego Gas and Electric, as well as the regulator, the public utilities commission, and we built a version of the model that was aggressive on the storage side. That dialogue and the publicly released aspects of our model led to a requirement passed by the California legislator in 2014 that by 2020, 2% of California's peak demand would be met by new storage that did not exist at that point in 2014. In other words, compressed air and pumped hydro were excluded from that analysis. 2% of California's peak demand is about 50 gigawatts. So we're talking on the order of 4% of that that was required. And interestingly enough, California basically met that target. It was a target for 2020, but we've now seen individual facilities. In Central California, we have one facility which is equal to about 1.3% of our total peak demand. And just to put it in context, in one facility recently opened 2021 in China was almost a gigawatt hour of capacity. Now I've been a little loose here, as you'll notice, between kilowatts and kilowatt hours or gigawatts and gigawatt hours. And that's because to really get the utilities comfortable with this in the beginning, almost 10 years ago, we talked about peak capacity. Now, of course, all of these assessments, whether it's here in California, whether it's the analysis going on at National Grid in the UK, or whether it's some of the PJM, the US East Coast Analysis, or thankfully, we've moved to the proper units at least kilowatt hours, megawatt hours, gigawatt hours. In terms of thinking through how much capacity we're gonna need, you're gonna get a range of answers, depending on who you ask. Some work that we've done with Imperial College, specific to the UK situation, because they are ahead of us in terms of deploying in particular offshore wind and storage. And you can see that as you wanna get to very close to 100% clean energy, the amount of storage ramped up really dramatically. And so roughly, again, this is an area of kind of fighting words, depending if you're a storage-only maven or if you're a storage plus nuclear or hydro person, we frequently find scenarios coming out of models like that call for between 60 and over 100% of peak generation capacity will be needed in times of storage. We in the Bay Area just had, for some of us had two days of blackouts, but back in August, we had a point in early September, we had this massive heat wave where it was touch and go, whether the grid would support it and California hit its peak demand levels, 50, 51 gigawatts. The grid did not go down, largely due to the amount of distributed clean energy and storage that's available in the system. I think I won't play with some of the simulation plots where we look at the amount of storage and which technologies are likely to win out over time. In this mix, there's pump hydro, compressed air, lithium ion, vanadium flow batteries, flywheels and hydrogen. And we do a lot of work to look at what we'll need, basically learning curves, the cost declines in different technologies. But I'm really just showing this, to show off a little bit about some of the images we do. So we'll skip over that. And really think system-wide. So California has really two key things. One is the set of storage mandates that I mentioned, but the other one is that California requires that we get to carbon neutrality. Right now, the letter of the law says we must be carbon neutral in the entire economy, not just the energy sector, by 2045. By 2030, we need to be 60% powered by renewables, which ramps up the amount of storage. And I think most interestingly, California was the first place to write environmental justice into its cap and trade requirements. 35% of California's cap and trade revenues that currently are $11 to $12 billion a year must be spent on underserved and marginalized and fence line communities. And you probably know that then vice president and candidate for presidency, Biden, did us 5% better and announced Justice 40 that calls for 40% of federal infrastructure spending to be for underserved marginalized communities. And in the Inflation and Reduction Act, the single largest allocation of money, $60 billion is for clean energy projects, energy transition, employment projects to that address the systemic and longstanding discrimination and marginalization of underserved communities, communities of color, women. And so it's a really remarkable push and it needs to play a bigger and more prominent role in the energy storage world. And I'll come back to that in a few minutes. Yeah, well, that's great to hear. California has these lofty goals. Again, that 2045 carbon neutrality goal is one that many of us are working with the state agencies. And to move that 2045 degree, that 2045 carbon neutrality date forward, I've proposed a number of places that 2035 consistent with when president Biden wants the national electricity sector to be carbon neutral would be a reasonable goal. And others are arguing for 2040 and we'll keep arguing. As we go forward just to highlight this, we've already seen points where California has done pretty well. What this shows here is April 24th, 2021, when California hit 95% of all energy consumed from renewables. And this is in the spring. This is our best time. It's lowest demand, best wind, best solar, not late August or early September. And you can see renewables spike up, natural gas use goes down, amount of hydropower goes to zero and California became a net energy exporter during that moment. If we're update to last year, we don't have the April date number between, but in 2022, we hit 104%, meaning that for more than a fraction of an hour, it was about a quarter hour. California met all of its electricity demand with renewables, something we've already seen in the UK, Costa Rica, Ireland, but as the world's fourth largest economy, I'm told recently California passed Germany, it may unpass, but right now we are the fourth largest. That is a significant number. And I would say as we think about the storage evolution, one critical feature here is that if in 2022, we can meet all our demand with renewables, the challenge is really on having storage catch up with that. That's really where my lab takes place. I will not bore you because I'm gonna assume that everyone has dissected and played with the so-called learning curve, just for those who haven't. What we observe for technologies that can be mass produced, deployed, such as solar panels, that we see a learning curve like this. This is log, log space. Time is a date stamp along the curve. It's neither the X or Y axis. Those are both log units of cost vertically and total cumulative production and installation. Then the so-called Moore's Law or Swanson's Law highlights here that we get about a 20% drop in price. That's the slope of this line for every time we double the amount of storage and that's a solar. And that's a remarkable curve, but it's not only true for solar. Here we have the exact same curve and dynamics for wind, solar and battery systems. And actually batteries have been improving each doubling of capacity slightly better than solar ever did at about 21 to 22%. And that's an aggregate number that's looking at cost of storage deployed, looking at lithium ion, lithium-bite flow batteries, all integrated together. And of course experts in different materials will dig in on their own favorite. But in terms of thinking about building out the energy storage resource, it's a really interesting and critical story. And it's led to the statement, Bloomberg News published this in 2021, that essentially everywhere, it is now cheaper to build renewables than to operate existing fossil plants. The next version of this statement needs to be, it is now cheaper to build renewables with sufficient storage so they can be base load or 24 seven, whatever your favorite news bit is than to operate existing fossil plants. We are not exactly there yet, but with these curves, we are getting close. So the analytics of learning curves are that we look at the cost at some future point, C2 relative to the cost today. And the models that everyone uses is to look at the volume of technology, solar panels, big pens, calculators, storage units, the volume of sales relative to the volume at that base point to that negative exponent, the learning curve. Well, one of the projects that we did in our laboratory now some years ago, five years back, six years back at this point was to take this apart. And so a large project we did in conjunction with researchers at Imperial College was to re-parameterize, to expand the data sets to make them open access. And what we found is that if you parameterize not only as a function of the volume of sales, a so-called one-parameter learning curve, a very dissatisfying feature if you're an academic because you'd like to think that research matters or something, what we find is you've got a far better fit, far better than just adding an extra variable. The whole collinearity issue is one we've dug into and this is statistically significant, not just the fact that we're adding additional variables in the story. So what we find is that if you have a second term or you parameterize the investment in R&D, which is the hardest number to get because even a grant or a company commitment to research in energy storage is often not a simple quantity to pull out. It's a bit tricky to identify exactly how much R&D money unless you're really literally spending a storage energy research grant. So we found this is a far better fit and actually our model, which has this R&D term in it by both the California Energy Commission and a number of individual groups, Bloomberg New Energy is actually the best fit model. Here's the paper, came out in Nature Energy again back in 2017. And we look at the learning curves for individual technologies. What we're seeing here is in the blue triangles are lithium ion batteries. We have flow batteries in here that are the circles. We have fuel cells on here. Pompidro doesn't show much learning because you don't get much better at doing something that simple. Lead acid batteries are here in the rust yellow color. And so you can see these learning curves are consistent but the slopes are somewhat different and our next paper that we hope come out in May now includes also lithium phosphate batteries which don't involve cobalt, one of the real bugaboos in terms of environmental justice. And it's that point that I wanna turn to the source of these materials to really highlight how we have to think a little more broadly about the story. So this is a picture taken at COP26 in Glasgow. This is the Power Africa, the US Agency for International Development team, Mark Corrado and myself, along with the Minister of Power in Nigeria, the Minister of Environment Nigeria and Damalola Oganbi, who is the UN Special Representative for Energy Access. She is the chair of Sustainable Energy for All. And one of the key issues is, how can we not only deploy more and more storage like through this innovation and policy landscape that I've described, but how can we also think very differently about energy services for the underserved in places that don't have big utility structures. And so again, I mentioned in the beginning, I'm just back from Kenya and South Sudan. And one of our projects there is a really a fascinating partnership. It's an effort where the German government, the UK government, the US government, as well as a new agency, an international group I mentioned in a second are focused on energy access for low income communities, focused on electrifying health clinics. Across Africa, there is about 170,000 hospitals, regional health centers and rural health clinics. And here on the right-hand side, you can see a deployment of one particular technology. This is a US Italian mixed company called Off-Grid Box. They make eight kilowatt and 15 kilowatt solar peak systems with four and eight kilowatt hour battery systems built in. They also do potable water. They provide Wi-Fi services and these are deployable containers. For larger facilities like you see in the center here, we need larger systems and we're working with many of the world's mini-grid providers, all of which that have to grapple with not only providing storage, but also with providing storage that is reliable. And in my trip to South Sudan last week, we went to eight rural clinics in one day on the road between the National Capital Juba on the Nile and Boer, the second largest city 180 kilometers north. Eight clinics, only two of them had a working system. In each case, the failure mode was the battery and in one case, it was the inverter and the battery. So as we think through a mapping of all the health clinics, this partnership effort to build that for both build data sets, to analyze what are the failure modes of those clinics and what can we do to get there has resulted in a process where our partnership has identified these 100,000 clinics without power and has set up a roadmap to build and maintain health clinic electrification systems at 10,000 clinics by 2025, a goal that I think will actually reach early. We're already over halfway there, but the monitoring, the maintenance, the training, being able to meet demands of new equipment, ultra-style machines, freezers, not just refrigerators for vaccines and other features as part of the story. My own lab is engaged in these build clinics to identify private sector partners that will be paid after the fact based on cold chains being maintained, radiological equipment being maintained and energy being made available for sale for local businesses and homes that are custard near many of the clinics. So that effort really highlights a huge cotton and white effort. Again, we're focused on my own lab in the Congo, in South Sudan, in Rwanda and Kenya. And if you go to the link down below, you can get to the data, many of the papers around these efforts if you wanna see how we're working to deploy. And again, batteries have been the failure mode in almost all of the systems. The other side of the story, I wanna say just a few words before I close is to highlight places where we are looking aggressively at other forms of storage, in particular, some longer term duration storage. And of course, give you this, but there are some places that will have sets of battery, have lithium ion batteries that you deploy temporarily in sequence. So maybe you go to six or eight hours by holding some of your batteries back, deploying later. That's a bit clunky. We're looking at flow batteries. We just came back from a flow battery, a vanadium flow battery installation in French Polynesia. But the other, of course, a piece of the story that's got a lot of attention and a lot of very wildly different views is on hydrogen. So Switch Japan is one of the models I mentioned. One of my doctoral students, Kenji Shirayashi and I recently completed this effort. And if you go online, you'll see there, we released a major report two weeks ago, highlighting the ways for Japan to get to a net zero future, even if the nuclear restart that they've announced does not end up handing out. And I think the most interesting feature if you can see nuclear in this dark maroon color, hydrogen and light blue coming in, renewables both wind and solar in the top colors, the green and yellow, is that there are scenarios that get Japan to zero carbon emissions. And in almost all of these cases, hydrogen plays a major role, largely because Japan's best wind resource is offshore around Hokkaido in the North and Okinawa, although Hokkaido is more likely to be a major national buildout. And what we find is that 10% of all energy in Japan would pass through hydrogen as a transmission or storage medium in these scenarios. I'm gonna highlight this because that was certainly higher than I initially thought myself, but it highlights where we might be going on this. Jumping back to the energy access story, this is just the time where we're going from where we are today, a few thousand electrified health clinics in Southern Africa, none of them would evolve hydrogen. These would all likely be lithium batteries, although lithium phosphate is likely the new preferred one. There's discussion about rust batteries. And I think you've talked about form energy and some of the new rust companies, some of the companies, some of the flow battery companies, ESS and others coming up. And in this effort to ultimately build these sustainable clean energy nucleation sites at a hundred thousand health clinics across Africa, you can see the partners in this effort, GAVI is International Vaccine Alliance, training and research efforts at the UN level, sustainable energy for all, my own laboratory rail at Berkeley. And what's come of this effort is worth noting for those thinking about storage, not just for utility scale efforts. And that is with a billion dollars provided by the Bezos Earth Fund and another billion by the Rockefeller Foundation and almost half a billion from the IKEA Foundation. This effort, which has a name called GAP, the Global Energy Alliance for People and Planet, now doesn't have all the money it needs, but it has about $2.6 billion to launch this mini-grid and healthcare deployment effort. And that really leads to my last minute or two of comments. And that is one of the features which anyone who has seen international development projects and has any healthy degree of skepticism will know, efforts that throw numbers around like 10,000 clinics by 2025, a hundred thousand clinics eventually know that unless the technical story works and the financing story works, this is another white elephant in the making. And batteries being the current weak link in that story is a reason why we have partnered with a Berkeley spin-out company called NLIME. They make these small devices, you plug them in to the address as you can see, you scan in the QR code, it uplinks and links the data from this particular device and it tracks outages and voltage and frequency. And so here's one that's just tracking the data flowing into this and the power flowing into this vaccine refrigerator here in the clinic in Goma in the Congo. We typically install two or more devices in case one gets unplugged accidentally or on purpose. At larger hospitals, we will maybe do four or five. Again, I just deployed 20 of these, 20 different clinics in South Sudan. There's programs to do several thousand of these in Eswatini, in Sierra Leone and DRC, but it highlights the data tracking part of the story. For those clinics that are on grid or have text access, this will upload the data directly. But for those that are truly off grid, you need to either drive by the clinic and they'll upload it automatically to a receiver. And in some cases, if you're outlying ones where using iridium and other satellite phones will upload the data there. But it highlights being able to track in detail. And so we're looking at voltage, frequency variations, tracking, this is data from clinics, small clinics in Eastern Congo. And the last thing I will mention is really the integration question on the storage and the vehicle side. And of course, everyone in California is familiar with this so-called duck curve. And this highlights that as we deploy more and more renewables, we hollow out the net demand so that in California now with a great deal of solar, roughly 11 gigawatts of solar behind the meter around the state, we reduce net demand to a very low level, which of course works against deploying more renewable energy, right when we wanted to deploy more, but then with evening peaks and high ramp rates, we really need to get power back in place. And so of course, schematically, this so-called abundance or problem in the belly of the duck, if we can transfer this to the evening, we change the story entirely. And of course, that's precisely the story where we think storage, both stationary storage and storage and vehicles comes into play. So one kind of exciting version of this is that a few years ago, we did a simulation for New York City to look at what if every taxi in New York City was electric. And so what you'll see here is a simulation run from midnight. A red dot is a taxi with a passenger. A blue is empty. I'll start running it from midnight. You can see they're moving around. It's now 1 a.m. Green is recharging in the middle of the night. And as we get to rush hour at about six, you'll notice a flood of red dots going into Manhattan. Here it comes. There they all flowed on in. First they were blue, then they were red. But now as rush hour passes, you're gonna see green dots appearing down here in lower Manhattan, as people want to recharge their vehicles without heading back to often where they charge or they live. So we did this simulation, worked through a plan to have New York City think about converting its taxi fleet. They didn't do so, but China called and China said, we'd like to do this at scale. The city of Shenzhen said, we would like to replace all 32,000 taxis in Shenzhen with EV taxis. For those of you with good eyes, you'll notice these are BYD, build your dreams, EV taxis. BYD is based in Shenzhen, so the partnership is natural. And while Shenzhen may have a gender problem in its drivers of taxis, they were able to deploy and my laboratory did the data science because as soon as they deployed them, even though at that point, Shenzhen had the world's largest EV charging station with 600 EV bays, the problem they got was this. Here we have the taxis waiting to recharge. After a 12-hour shift, everyone was required to recharge to hand over the taxi fully charged. So we did, of course, what anyone in Silicon Valley does, both at Stanford and Berkeley is build an app. And here's our app that each driver gets that says how many bays are available and you can schedule the time and the more prompt you are to rival when your time slot comes up, the more points you get. So you get priority seating, if you will. And you can see a picture of the driving bays here. And now Shenzhen is working to install inductive recharging of these batteries along the road here. I won't go into the details. I'm gonna really end there because I've gone right to the end of my time. But I wanted to highlight that as we think about storage, there are these technological material science issues. I highlighted some of that in the beginning with the learning but also the role of hydrogen and something I'm not gonna get into today but we might discuss in the Q&A a little bit. And that is the exciting world of really ramping in not just marine energy in terms of offshore wind but offshore metal and wave energy. And there are companies now emerging that do things like this that have wave buoys that are generating power, some for electricity directly, some to make hydrogen. And I think the last and most interesting feature of course is that the timing and periodicity of offshore energy is significantly different than onshore energy. And so thinking about storage but also ways to think about convenient partnerships and as Megan highlights the water role, I'm ending with a slide here that really just highlights the water opportunities as we think about what may be the biggest change in the world's energy sector the rest of this decade and that is the rise of offshore energy. So thank you so much. I think I'm ending right about on time and I appreciate it. I'm looking forward to our discussion at the end. Dan, thank you so much for that comprehensive overview. We have time for a few questions. I really enjoyed your highlight of the Africa energy storage solutions. So I thought we could start there and talk a little bit about trade-offs. So you highlighted many trade-offs for countries like US, Japan, China. How are these trade-offs different for Africa? Performance cost trade-offs, manufacturer ability? Could you talk about the Africa-specific trade-offs? Absolutely. So let me, I'm gonna take advantage of that to put a slide back up, I have to admit. And that is the trade-off in the Africa story is really simple on the technical side for these off-grid systems, on-grid big cities, places with well-functioning grids like Kenya, Senegal are exactly the same as we have in our situation. And that is that we have many groups that are keen to install renewable energy, California, Ireland, UK, et cetera, but not as keen to integrate in storage. Now that's an area where the UK and their national grid has taken a very aggressive role. The United Kingdom is converting its largest industrial center, Humber, to be a hydrogen-powered hub. So they are taking an aggressive role. I mentioned California with the storage mandate, but I think the real big story for this off-grid world is the other end of the life cycle value chain. And that is that much of our critical materials, I don't call them rare earths because they're not that rare. They are earth-based, but they're not rare in most cases. But the environmental and human rights abuses in and accessing many of these materials, in particular, Cobalt is a critical part of the story. This is an open-pit mine in the Congo. And as we think about the demand for energy storage, this clean energy transition that we must do is gonna be one where we transition from a hydrocarbon mentality to a metals and materials mentality. Increase in demands between now and 2050 of up to 1,000% per forecast. And again, lithium, which look massively scarce only 10 years ago, calcium is not so scarce at all. We're finding it in the Salton Sea. We're finding in aquifers in Wyoming and Madagascar in a number of places in Australia. And the demand for Cobalt, which is, again, the most human rights abusive of all critical materials today. If you go online, you'll see that Cobalt demand has dropped dramatically in part because people are moving away from their screens a little bit after COVID and the demand for electronics is down, but also in part because the shift from lithium ion to lithium phosphate batteries that do not require Cobalt has really gone, I would say at this point, I don't wanna say swimmingly, but kind of wonderfully well. We've seen a real ramp up. So much so that Rwanda and Kenya and Senegal are all talking about in-country first assembly and then manufacturing of batteries using local materials. So I'll just add a mine in the Congo where they have facilities to smell tin. They send their slag cedarite to Europe to be reprocessed to get out tantalum and neodymium and others, but they are now looking at doing in-country and they're appealing to the international community. Some of those billions I mentioned to really ramp that up. And I think the biggest part of the story will be if African governments are enabled, their private sectors are enabled to do local manufacturing, something that the West has been very, very poor at so far. That's really the story of how to make a sustainable storage industry in Africa and Southeast Asia. So thanks for letting me go back to a slide and I will unshare. Dan, thank you. Thanks for highlighting the importance of local manufacturing as a way for economic development. So I guess what you're saying there is the technology choices, my favor those that can be more easily manufactured locally. That's absolutely right. So materials that need to be sent off for secondary processing to get out the very low densities and often the toxic issues around cadmium, neodymium and others, that's an area where the international community needs to really listen when again the Kenyans, the Senegal's, the Rwandan's countries that are proven they can manage these, request that funding, that needs to happen. Otherwise we essentially replace a, I would say, in the crudest terms, a hydrocarbon colonialism with a material science colonialism and that's not gonna do any of us any good. Completely agree, Dan. Maybe let's take two more questions. You showed this beautiful collection of cost learning curves. Let's talk about two aspects. One is intervention. How can policy and others drivers be used to shift those learning curves? Especially- Yeah, so the beauty of these learning curves and I'll just jump way back to it, the beauty of them is that they have been so consistent for so long. When you look at these curves, first the one for solar that everyone's familiar with going back to the 70s and then the one that we kind of focused on today on energy storage. What it really highlights here is that if we take seriously what we find analytically to be true, yes, you could parameterize in a one parameter world just more sales, but really R&D plays a critical role. And again, we find that the fit is massively better. It's about 60% better if you include in an R&D terms. What this says is that policies that forward price that give benefits for storage such as right now in California, while we're not, we have not taken on the so-called NEM 2.0 or NEM 3, this net energy meeting story, particularly smartly. We are shifting California sub, you no longer get a subsidy just for installing solar, but if you install solar and storage as we have in my garage, not my car, but I have a stationary storage bank in the garage, you can get a subsidy still. So that says volume pricing is one thing. So if we forward price give a benefit to install at companies and businesses, that will push the price down, but it also says that we need to maintain and to track and to quantify the R&D dollars. And while we're having this meeting, the RPE summit is taking place. That's a place where R&D money for storage is going in, but increasingly cracking dollars that go to R&D or R&D or euros that go to storage is one thing, but storage is only useful as part of an integrated package with clean energy generation. And so really we're gonna have to get even more sophisticated adding other terms here and deconvolving. What about an investment in low battery plus microhydro or green hydrogen plus wind generation? Those are gonna need to be the frontier for bright young students who wanna kind of take the sort of work and evolve further. It's that deconvolution where energy science and data science are gonna have to become a real partner to highlight exactly what you're talking about here. What do we get by investing in different packages as we need to accelerate this transition? Thank you, Dan. We have time for just one final question. So again, on the learning curves, there is a little bit of spread between the different energy technology in terms of the learning curve exponent. Do you expect going forward there to be a diversity in the learning rates? And if so, how do we incentivize the system to pick the max learning rates? Yeah, so this is an area where I think your perspective is gonna vary. So I'm a physicist and one of the things that physicists learn and try to tell other people is that you should not get too excited about the fine structure in vague variables. So this is a long, long plot. And so if we look back here, for example, at the famous solar learning curve, there are people who went to town trying to understand this little dip here. Why did the prices drop here? Why did they go back up? This is chasing data that is probably illusory because what we know now is that this bump and dip actually happened when two big German PV manufacturing facilities came online, the capacity went way up, prices dipped down. But overall, over decades, that has been a pretty straight line. So I would caution people to read too much into the fine structure, but the broad structure is really interesting. And that is that we consistently see you centralized manufacturing, so-called Japanese or just-in-time source materials find ways to get leaner and meaner both on the amount of materials, reducing the amount of actual atoms and molecules and more efficient saws for solar cells, better thin film deposits, better use of reagents, all of that, it really makes a difference. And so I would say the lesson here is very broadly the case that if you mass-produce it and you deploy it, learning curves are gonna be your friend. But picking out the details on those bumps has historically proven to be an effort where it's garbage in, garbage out. You find and get what you want to find, not what's necessarily there. Thank you, Dan, for the great discussion. All right, and that's now, thank Dan, and we will have our second speaker and then we'll return to have a short discussion with everyone at the end of the seminar. Good morning, Megan, it's great to see you. How are you? It's you, I'm doing well, thank you. It's a great pleasure to have my colleague, Megan Mottor, speak today. And Megan is working in the area of sustainable water supply. And in the past couple of years, she has been extending this to the intersection of the water and energy system and specifically looking at the role of water for energy storage. So this is a new topic for us as storage experts, but one that is very exciting. Let me also just say a few words about Megan's broader contributions. She's working in the area of water treatment, water management, water policy, and she also is the research director for the National Alliance for Water Innovation that's a hub funded by the Department of Energy. So Megan, we're really excited to hear from you today on how water energy storage can be connected and new ways of thinking about it. Thank you, Megan. Well, thank you so much for the introduction. Well, and I think Dan set me up really nicely just talking about the host of real opportunities that exist across the energy storage space. I'm gonna talk today about energy flexibility innovations in the water sector and really the ways in which we can think about to the water and wastewater sectors as sort of industrial energy loads that have flexibility and can help support a really stable and reliable grid. So we're all here because we understand that solving the gigaton carbon problem is really gonna require what I call a 60 revolution. We need to diversify our sources. We need to decentralize a lot of our generation. We need to decarbonize that generation. We need to decouple energy storage capacity and supply by decoupling supply and demand. We need to continue to drive down the energy intensity of our unit operations and we're gonna digitize all of this because the future is not people turning on and off their switches manually. And I'll just say, I recognize and speaking to an energy storage audience here, the way in which I think we've increasingly started to think about realizing a stable electricity grid is certainly through storage as well as through industrial energy flexibility. And so really having control over the timing of loads is absolutely imperative. When you have a large taxi fleet, that's certainly one way to do it but there's still gonna be a lot of baseload demand throughout your system that hasn't historically had that same degree of flexibility. And so a big challenge is trying to realize energy flexibility in some of the more energy-intensive industrial sectors. At the same time, my own research realizes that a lot of the effects of climate change are actually gonna be felt through the water cycle. And so as well mentioned, a scholar that works primarily in the water cycle on water systems, both thinking about how we need to adapt our water systems to accommodate major changes in where precipitation falls, when precipitation falls as well as how water ends up being used in those places. And then in parallel with that, thinking about how we mitigate the effects of tapping non-traditional water sources, particularly reducing the carbon intensity of water treatment. And so when I am thinking through the process of developing new infrastructure for the water sector, I think I'm trying to keep at the front of mind the real importance of how we are ultimately going to integrate that water sector with the electricity sector. And so one of the ways I think we can potentially do that is to follow a similar pathway to the pathway that the electricity sector is headed down. I like to say that the water sector is 30 years behind the electricity sector, but we're all headed in the same direction. We are going to really require a similar set of 60 transitions. So we're diversifying away from conventional freshwater sources to a whole host of non-traditional water sources would be that seawater, wastewater reuse, industrial water reuse, brackish groundwater, and et cetera. We're also really starting to think through a comparable transition to the transition of large centralized coal fire generation to solar cells where it's on everybody's roof to a similar decentralization of the water sector, thinking about onsite reuse of water, even in one's home. Again, we need to decarbonize these processes and that's easier in some water treatment installations than others. There's opportunities to use what I think are a tremendous amount of storage infrastructure that is currently underutilized to help transition this water system to a future decentralized system, but also potentially to really think about integration with the electricity grid. We still need to focus relentlessly on it and water efficiency. And just as you're not going to be turning on and off, switch is manually all over the place. You're not going to run a complex water network using people and heuristics. You're going to do it certainly with people present, but with the assistance of a future digitized water grid. And so as this water sector transition unfolds, I think that it's really exciting, but it doesn't happen in a vacuum. And I think that a lot of my work is focused on really enabling these 260 transitions in the energy and water sector to happen in an integrated and synergistic way. And so my work largely focuses on how do you coordinate these future systems, particularly when they are more fully 60 to really support affordability of the energy and water sector as well as resiliency of the energy and water sector. And so I like to think about how to maximize energy flexibility opportunities across the water supply chain. Our group has done a lot of work on the transmission side thinking about source switching and pumping cessation in large scale water distribution and what should say water transmission systems. You know, the state water project in California remains a tremendously large demand response provider and really helps us to stabilize our grid here, but there are many other places around the world where this is also going to be pretty important. We have thought a lot about storage and how you use water storage. There's a lot of excess water storage capacity in our system, particularly, it might not be excess all the time, but it's definitely excess in some periods of the year. And so how do you use that more effectively? How do you start to run your treatment systems in new ways, intermittently, turning on and off really energy intensive processes? How do you use the distribution system where there's even more storage and particularly storage in those very visible water tanks, but ways to not only store water in publicly owned water tanks, but also thinking about how you flex consumer demands and use that to also provide some low shifting capability. And then we do a lot of work and water reuse. So again, how do you tie in the wastewater sector into this equation? Particularly on that front, there's a lot of effort in on-site electricity generation and wastewater facilities and on-site electricity storage even. So our lab broadly develops tools to do energy planning and investment de-risking in the water sector to help utilities quantify bill savings and carbon reductions from implementing some of these approaches and then ultimately to develop control platforms that help those operators manage their energy in real time while not seriously inflating their risk profile. So I wanna go into several different studies that we've done along this overview figure and give you a little bit more color behind each of those potential storage mechanisms. So I'm gonna start out with a study that we've done at wastewater treatment plants, really looking at how on-site electricity generation, electricity storage, water storage and intermittent operation can be used together to help shift load and save bills at these wastewater treatment plants. So why the wastewater treatment sector? Well, wastewater is actually really energy intensive at the moment it consumes somewhere between one and 2% of US electricity but that is rapidly growing as new water quality regulations come into effect and especially nutrient recovery becomes required. So we're seeing a major expansion in energy intensity of the sector. Another really particularly problematic issue is that even at facilities especially those here in the Bay Area, like East Bay Mud and Silicon Valley Clean Water which is the site that I'm gonna cover today that actually capture their bio-stallage, digest that on-site. You may see those big digesters, sledge digesters that generate biogas. That biogas generation is both incredibly unreliable. It's kind of standard combustion turbines that are being used and they do not always perform very reliably with what is middling quality biogas. And even more importantly, that generation tends to be pretty poorly timed with demand. So we end up flaring a very large quantity of the biogas that we end up producing. These facilities are also facing some stress in that electricity is 25 to 40% of their operational costs. That is a lot for a water facility and we've seen pretty rapidly increasing electricity costs over the past couple of years. And so this is starting to pinch these wastewater facilities and they're trying to figure out how to respond to that. In particular here in California and parts of New York, those demand charges end up being a pretty large fraction of the electricity bills. And the projection is that, given increasing adoption of renewable energies and that that curve that Dan talked about, people expect the electricity bills at these facilities to grow pretty substantially. So I think that over the next 10 to 20 years, there's going to be a real demand for understanding how to transition these wastewater facilities into more flexibly operated facilities. And I sort of scroll through some of the low profiles in this bottom figure as I was talking, but you can see this again is for Silicon Valley Clean Water where we have this base load demand is in the light gray in the back and the generation from these sludge digesters producing biogas and then being combusted on site in the pink. You can see it doesn't fully match those needs, but critically you commonly have these big outages. So we're interested in then simulating how load shift could help to reduce the energy bills that these plants at the same time end up helping to stabilize the grid. So what are the energy flexibility resources that we might deploy to do that? Well, the first is to intelligent, more intelligently deploy the batteries that have often been purchased to enhance plant reliability. I want to remind you that wastewater people require a lot of electricity. If there are power outages, that is very problematic. And so it's quite common for wastewater facilities to have some sort of either energy storage or backup generator present on site so that they can do their job when the power grid fails. And so we've done a lot of work to think about how to more intelligently deploy those batteries and help use them not just for resiliency applications but also to help shift load at these facilities. Alternatively, we can really dig into the nuts and bolts of how treatment plants work and uncover a lot of pretty diverse alternative storage options. So you may not know but wastewater treatment facilities actually have a tremendous amount of wastewater storage both upstream of the plant, Silicon Valley Clean Water again here in Redwood City is in the process of building an enormous tunnel to store wastewater and that's important because when you get big atmospheric river events or general storm events you need a place to store water and help equalize load through your wastewater treatment facility. This is also really common on the East Coast. I should say it's increasingly common on the East Coast because of combined sewer overflow regulations forcing sort of large interceptor design. So there's this amazing amount of water storage that is in the process of being built and commissioned. There's also primary effluent storage within the plant it's not visualized here. There's the ability to store biocellids and then there's actually the ability to store gas. A lot of sites do biogas storage to reduce flaring and effectively also reduce their local emissions. But facilities, even though they have all these things they really struggle with planning and coordinating these energy flexibility resources because A, there's a bunch of different options. B, the storage types are really interdependent if you store wastewater your biogas production is going to be changed and it's hard to anticipate some of those interdependencies. The optimal configuration for how to do this is really site specific. And there's also some sort of major issues with thinking about deploying infrastructure for multiple purposes. So we all love the idea of multifunctional infrastructure because you get way more bang for your buck and at the same time you expose yourself to slightly larger quantities of risk. And that risk might be very marginal and totally worth accepting. But I think it's really important when talking to operators and understanding what is going to enhance adoption to provide them with risk aware insight into their operation of multifunctional infrastructure assets. So a lot of what we've been working on this is a project that was funded by the US Department of Energy, EEREs, Wastewater Resource Recovery Program. But we've been working on this program in designing a design tool, building a design tool to help identify cost optimal storage upgrades to these wastewater treatment plants. And what this project does is build digital twins of those wastewater treatment facilities that combined both statistical learning and process-based modeling to really capture and evaluate the effects of water storage, gas storage and electricity storage upgrades and relate those to ROI numbers, return on investment numbers and the effects for these wastewater facilities on their bill savings. So I wanna go briefly over some of the results from this paper. This is currently under review and should be out soon. We've done a lot of work thinking about just generally what are the different advantages and disadvantages of these different storage types? So we've got battery, storage, gas storage, raw water storage and primary effluent storage. And they certainly all have very different characteristics in terms of the way that we would typically think about sort of a virtual battery. And I would say most distinctively is this kind of capital cost range. So this assumes new builds for all of those not using existing facilities, but we wanted it to get a sense of if you were just to deploy these resources kind of from scratch, what would their respective costs be and what would the expected bill savings be? And I should say this is all for one hour or one load hour equivalent at the plant. So one hour of standard plant electricity load. And what you can see is that here, there's value in battery that far exceeds the battery or say the storage value of some of these other assets. And yet there's a lot of diminishing returns on the size of that battery over time. So as you get larger and larger batteries, of course, you can flex into larger biogas generator outages, but at the same time, your effective cost of that battery size goes up. And so you're really trying to understand how much storage you wanna deploy at these plants. We've also looked at the synergies between these different storage types. But one of the interesting things, this is kind of a modified Shapley analysis, is that there's actually not a lot of synergies between batteries and some of the other, some of the other storage techniques, we expected to see much greater synergies than we did. And that was a surprise to us, but it also challenges us to differentiate between retrofits at plants and green builds. And I think a green build might help us tap synergies in ways that we were not able to at this particular plant. The next analysis we did was really looking at, okay, given Silicon Valley Clean Water, instead of looking at one load hour equivalent deployment, what is the optimal deployment of these energy flexibility assets, both in an instance in which a plant had a co-generator facility and one in which they did not have a co-generator facility. And the first takeaway is that when you have co-gen, there's much greater value in storage. So there's a real positive synergy between those two technologies at these plants. And also the battery continues to be a really valuable, the preferential mode of storage. But there's actually also pretty, if you already had storage on site at a facility, but you didn't have a battery, there are still some pretty positive return on investments and overall bill savings opportunities for these plants. We've gone ahead and done this, as I mentioned, at Silicon Valley Clean Water, but we've also explored deployment at Watsonville and Santa Barbara. And I'll just say that at Silicon Valley Clean Water, especially it was really interesting to look at how the existing battery differed from what would have been the optimal battery deployment. So both in terms of battery power and battery capacity, that existing battery is both a little bit, it's a little bit miss-sized for that plant. And so there was some effective loss in that present value because of suboptimal battery design. So it's really important to think about pretty site-specific design factors when going into a plant and thinking about helping them do energy flexibility, retrofits and upgrades. I'm gonna skip ahead there and just say that on the whole, this project has been really interesting in helping us to look at how wastewater facilities could couple on-site biogas generators and batteries to help provide some of that bi-directional energy storage. There's also these opportunities in biogas storage that effectively are providing chemical and fuel storage. And then we also have controllable loads by storing water, modifying or modulating aeration, et cetera. And so I think that really speaks to some of this flexible generation and controllable load question. I realized I was speaking to a storage audience, not a wastewater audience. So I wanted to relate that back to the areas of storage that DOE highlights. Okay, I'm gonna move on and just go much more quickly through some of these other studies that we've undertaken. One is particularly looking at water storage systems. And actually I should point out that this is water storage in the distribution system. So I'm sorry that the arrow should go over here, but a big piece of deploying storage as I mentioned is in risk-aware dispatchability of these storage resources or virtual batteries. And that's particularly important for the water sector. The water sector has a really clear mission of providing clean water in a reliable manner, not just for drinking water applications, but for fire. Fire control is really what dictates an awful lot of our water distribution system design. And you must plan for unexpected events in that distribution system design. And so what we wanted to do is develop a tool that would help distribution system operators and utilities to really understand how much additional risk they were taking on and also what was the additional value add to the utility from a revenue perspective in terms of accepting that risk. So many people have built water distribution system optimization models that minimize the electricity costs of pumping water from the treatment plant to consumers. And in this study, we recreated one of those and basically looked at an optimized pumping schedule for a city of about 20,000 people. You can see that you minimized your pumping load when you have high electricity prices and vice versa, okay? The next thing we did is asked, what is the additional load shifting capacity that pump scheduling and water tank drawdown might provide if an emergency DR call was issued? And I should say that we looked at a couple of different demand response timeframes. We looked at different notification windows, different shed windows, and then different recovery windows in the system. And in doing so, we acknowledged that utility might not want to fully draw down their water tank reserves because they wanna make sure that they had sufficient water supplies to meet demand and pressure throughout the system. If there was an unexpected event. And so the risk tolerance of the utility would probably depend on what level of compensation they would get for different notification, load shed and recovery period durations. So the upper plot characterizes a range of timeframes we looked at for each of those periods. And then this lower graph here just does this for a recovery period of eight hours. Which is a pretty long recovery period that would represent pretty serious drawdown of your tanks. And we look at what is the optimal load curtailment or shed fraction under different compensation prices. So the takeaway from the study is really that you'd need to allow utilities to make choices and form choices based on their own knowledge of the system and their own understanding of how long it's gonna take to recover. And that's gonna end up informing what price they're willing to bid in these markets. The sort of next study I wanna quickly go through is one of intermittent process control. So we've done a lot of thinking about how you would turn on and off treatment plants to modulate the duration and timing of particularly energy intensive processes. And at the same time, I think there's a fun story to tell about how intermittency in treatment plant operation is not actually all that uncommon. Maybe not on those short time scales but certainly on long time scales. So I wanna open with a positive story here. That is one, this is a picture of Santa Barbara's Charles E. Myer desalination facility in Santa Barbara. This facility is pretty energy intensive. They consume about four megawatts on an ongoing basis. And the plant operators are also very, very interested in trying to help support a fully decarbonized water system. And so I've been pretty progressive in thinking about how they might actually participate in energy flexibility and activities and DR calls. And so on September 5th, this past September, right over Labor Day weekend, there was a huge heat wave. It was kind of an emergency situation and all of the utilities started getting calls. Can you shut down some part of your operation? So Santa Barbara said, sure. And they ended up shedding load from 5 to 9 p.m. And they actually got paid a pretty decent amount of money for that, that this represents a healthy fraction of their monthly bills. And so they came to me and said, well, can how possible would it be to do this on a much more regular basis? And what would the value be? So we were actually just getting that project up and running, I'm trying to understand not just how they could flex, operate their existing plant, but they're also considering a desal expansion. So they are thinking about nearly doubling the size of their desalination plant to accommodate drought. And they're wondering whether they can, in building out that desalination plant capacity, also build in greater energy flexibility. Which brings me to this other interesting case of Santa Barbara, which this is the same facility, which is that it was initially built in 1992 because there had been a big drought. And then in 92, it ran for three months and then it started raining and raining and raining and raining and all the reservoirs were full. And they shut down the plant after only three months of operation. And the big challenge was, okay, what do we do with this huge piece of infrastructure that we are no longer using? And they ended up basically selling it for parts and totally deconstructed it. And then 30 years later, recommissioned it and I got it going again after another large drought. But the reality is that they spent a lot of money. I think it was 70, something like $72 million to build a facility. And if you just averaged that over the three months of water production it comes out to something like $104 a gallon. So a lot of money. And I think that this speaks to the fact that we are going to experience sort of futures in which we're gonna see droughts, we're gonna see large precipitation events and our infrastructure in the water space is going to have to become much more adaptable over a whole host of different time domains. So we can think about all of the different time domains over which these desalination plants need to be flexibly operated and the potential synergies in that flexible operation. So everything from the decadal scale of like, do you, should you expand that Santa Barbara plant? Especially given that we've had an incredibly wet winter and all of Santa Barbara's reservoirs are now full. There's also questions of, how do you operate much more on an hourly or daily scale where maybe you can modulate your, I'm sorry, maybe you can modulate your water recovery to not consume so much energy during that four to nine window or five to nine window and help shed some of the load that would typically be consumed by the plant. So we're trying to think about lots of different time domains in these facilities and I'm gonna go pretty quickly through this but in doing so, we're integrating short-term flexibility with long-term flexibility and in terms of drought intensity. And we've done some really nice work with my colleagues, Sarah Fletcher here in the CE department and a co-advised postdoc, Marta Zanielo really simulating what future drought scenarios might look like and how are different types of water treatment capacity. This red is baseload desal capacity might actually end up being deployed over those long periods. And I think the question is within this, maybe it makes more sense to develop facilities that don't have baseload capacity in the way that we historically think of it but they themselves end up being very flexible. So it's not just an on off but you're really thinking about on off during different periods of the day so that you're actually reducing overall water production capacity and really reducing the carbon intensity and electricity bills of operating those systems. Right now, we're depending on a lot of this kind of peaking capacity from smaller plants and we're using desal plants as a baseload. Okay, the final piece I wanna cover is just thinking about demand curtailment. So the electricity sector has done a lot of work on how do you shed load at the consumer scale. I think there are similar opportunities in the water space particularly focused on the timing of irrigation for our lawns and to support some of that we've built what we call a flow backtracking model that helps us determine what is the marginal energy intensity the marginal cost and the marginal carbon intensity of delivering a unit of water to a specific person I should say a specific node in the water distribution system as a function of time of day. And this again is a picture of Santa Barbara. You can see that the marginal energy intensity so this is kilowatt hours per cubic meter of water delivered is very high in the low flat areas and that's because they're receiving water from the water desalination plant. It is also high up on the hills. And so this is obviously very unique to Santa Barbara the inverse, I should say the inverse but sort of the inverse is true in Berkeley where the Berkeley Hills, the energy intensity of pumping it up to the top of the Berkeley Hills is incredibly high and the energy intensity is much lower in the flats. So we've used these marginal energy intensity models to help develop a couple of tools. One is a tool for helping to price water at different locations and times of day to incentivize irrigation water demand side management. This is not adopted anywhere but we think it will be. Sydney water especially is interested in actually having pretty dramatic time of use pricing in their water supply. We've also developed tools to help utilities realize carbon reduction goals without costly infrastructure upgrades. And then we've thought about where to deploy on-site water recycling units to help support energy efficient, I should say energy efficient water system. So I just want to wrap up by saying again that 60 transitions don't happen in silos and I think there's a lot of opportunity for the water sector and the energy sector to co-develop particularly the water sector to co-develop alongside a changing electricity sector and really play a powerful role in flexible load operation to stabilize the electricity grid. I want to wrap up by thanking all of my wonderful postdocs and PhD students as well as the collaborators on some of the projects from Rasha Kapal here at Stanford, Sarah Fletcher, also in the department in Silicon Valley Clean Water. And with that, I will end. Megan, thank you very much. That was a great overview of the water energy system. So we have time for maybe one or two questions and then we can come to a panel discussion with Dan. So Megan, in your talk, you highlighted both how batteries could be deployed to make the water system more efficient but also how the water treatment system can be used to decrease the need for batteries. I was curious if it can comment for the latter, how big an impact would it have? Sort of how much batteries would it replace if we properly utilize say water treatment, biogas generation, and how much battery would that replace? Yeah, so that's a great question. And it is incredibly site specific, right? It's hard to say that in a really general way because both of the local characteristics of your energy sector, but also the very, very site specific dependencies of your water sector. I think the easiest way to think about this is how much energy is consumed by the water sector. Again, that is very site specific. In California, it's incredibly energy intensive. Somewhere around 5% of California electricity is consumed both in pumping and treating water. In other places, it's incredibly low. I mean, you've got gravity-fed systems that really have minimal treatment. And so there isn't a whole lot of energy flexibility sitting there on the table. So I would say you really need to do site specific analysis to answer that question. But we're looking in a diesel facility at megawatt levels of flexibility delivered to the grid. Santa Barbara's usual base load capacity or say base load energy consumption is about four megawatts. So you think about, okay, could I fully shut that down? Could I fully shut that down for four hours? Could I fully shut that down? You have a lot of flexibility in the duration as long as you don't have a lot of constraints on your water production. And so again, I think it's how do you manage these coupled systems that becomes the real question? Thanks, Megan. Yeah, I'm just trying to understand the order of magnitude of how it can be impacted to half. You mentioned duration, right? There's a lot of complex coupling on the duration. Like you said, the water treatment needs to happen at some very specific intervals. So what are the range of durations that you think could be viable? I think you mentioned one hour in the park. Is that sort of the ballpark we should be thinking about? No, so I'm sorry I didn't make that more clear. We did analysis for like one hour battery storage, especially it is, and there's diminishing returns from the bill savings perspective at these utilities. But if you re-optimize this for carbon, it gives you a very, very different, if you're optimizing this for the grid, instead of optimizing this for the plant, you get a very different optimum outcome. And that was actually a really important insight that we had, which is that this is actually very, very low cost carbon reduction because you can do very long duration storage. So those wastewater interceptors, when it's not raining here in California, which is like I won't say all the time, but a healthy fraction of the year, it's not raining, right? And so doing storage, those can store massive amounts of flow for long, long periods. We kept it in our analysis at 24 hours because of some of the risk of going anaerobic in those systems and the risk of downstream water quality impacts. You don't want them to be smelly either, honestly. That's a big problem. Start to generate methane and hydrogen sulfide, but that aside, sorry. We're looking at much longer duration of storage, so 24 hours, six hours at very, very low cost. Thanks, Megan. I think that gives us a sense. So on the cost aspect, I think what's really exciting here is as you mentioned, much of the capital is already deployed and the equipment is on the ground already. What are the barriers now to implementing? Because most of the fiscal stop is already there. Yeah, so the barriers are, I think, twofold. The first is that we do need to deploy this. I mean, I think when you do anything at scale, you run into challenges that you don't anticipate. I don't think especially understanding whether a treatment plant would need to adapt any of their water treatment, downstream water treatment, unit operations is important. We already do a lot of water storage and so to equalize load. And so I really don't think that that's a huge risk, but we need to see it. And I think water treatment plant operators especially need to see that because there's an inherent risk avoidance within those operators toward new change to their plants. So the people side is a big one. The other big, big issue is just like, how many operators do you have and how are they going to think about controlling this alongside all the other things that they're trying to do? And again, this is a human piece, but the human element cannot be understated. The wastewater treatment plant today, it was staffed by a person of an average age of 54. There's not a tremendous amount of digital fluency. There's a lot of clipboards and paper. And so getting digital systems deployed where, again, the future is not people running around throwing switches, the future or turning on and off pumps. The future is to really have this be integrated and certainly overseen by people and people are part of making decisions on whether or not to participate. But I think the actual operation of the system needs to be digitized and for the wastewater sector, there's a big, that represents a big shift. So again, we're 30 years behind the electricity sector, but we will get there. Great, Megan, thank you so much. Let me invite Dan to come back to the stage as well. We have about 20 or so minutes in which we can have and hopefully a spirited discussion on energy storage. Dan, maybe I can ask you to start. I realize that we didn't really comment on the scale of energy storage that is required for the grid. So Dan, can you comment a little bit on approximately the terrible one hour that is needed to start to make a difference at the grid level? Yeah, I mean, obviously this is kind of a big picture modeling question, right? Because if, for example, you have more reliable hydro, which we don't have in California anymore, we're talking about the biggest dam in the U.S., the Grand Coulee Dam, for example, being shifted from a primary generator to a battery to only being used for load following. Swings and thinking about big hydro facilities. The governor, for example, asked me in the switch model to zero out hydro entirely from California's mix so we can save water for agriculture. And ideally, my pushback is both save water for nature and communities, not just agriculture, since agriculture consumes 70%. So that's Megan's territory, but that was my little rant on it. So if you think about a kind of a digitally super interconnected area, Northern California, the UK, where you're actually planning to have a more diverse set of renewables, the numbers I show you in that study we did for National Grid, where they found, using their constraints, we found that you need between 60 and like 110% of generation capacity available in storage. And of course, as Megan and I both highlighted, we had these outage events, near outage events in September, where we were cedaring on it in California. We made it largely because there was so much behind the meter solar and because people conserved a bit. So it really depends on where you are. There's no single answer, but I think this is a place where modelers are gonna get us into trouble. And I say that as a modeler. And let me explain, I'm sorry, it's a long answer. Right now, most places don't have that much renewables on the grid. California, we're an exception, Portugal, et cetera. So if you ramp up to have a good ratio of say, two to one, new renewables and storage, you're probably gonna do all right. Worrying about what we need when we're at 92% is a fool's errand because we may have a huge hydrogen or an ammonia economy that we might not. And people who plan for what we're gonna need at the end of the transition are missing. We need to transition for worrying about things. And it is hard to tell that story because certain investors only wanna be jumping ahead. But what we find here is that California's grid, we peak at about 51 gigawatts. We import about 10% and we only have about five gigawatts of storage. We are a super leading edge energy system and we're about to go to 100% electric vehicles about meeting over the next decade and a half. So we need a higher number than 10%. But we probably aren't gonna need these huge numbers, largely because almost everyone who's installing solar today, companies and individuals is being smart and they're installing solar and batteries because that's where the subsidy is. There's no more remaining subsidy here just to do solar. So I've danced around your question a lot of ways. The shortest answer I would say is that in all of those switch models, even when you think about the extreme events, a one week cloudy period or this and that blackout period, if you diversify your supply, offshore wind plus onshore wind, solar, you probably don't need more than about 25 or 30% of peak generation demand available in storage. Now, what I haven't said is how many hour duration? Clearly four hours is sufficient. I have a N-phase 11.2 kilowatt hour battery in my garage. I have two electric vehicles. They have more than 12 times that capacity, but in Northern California, vehicle to grid is not yet permitted. We just had a two-day blackout and I ran out of power late in the second day because I turned off my hot tub, very Californian. I turned off my refrigerator and so we struggled to make it. I think getting storage deployed, more or less part and parcel with new generation is where we should be thinking, not worrying about what's gonna happen between 89% renewables and 100%. That's a long way off and none of us are smart enough. Even those who claim they are smart enough, none of us are smart enough to give that answer. Thank you, Dan. I completely agree with you. So in the near term as we are deploying new generation capacity, new storage capacity, it's gonna be a very interesting interplay between the two. And the one thing that I am worried about is just overbuilding, say, of manufacturing capacity for battery technologies. And that could be a challenge here, especially as one thinks about how to go between battery for transportation and battery for the grid. Big decisions are being made and how to split resources. Any thoughts on that, Dan? Megan, I think has more thoughts on this, but I think it's really simple. And that is that if you enable batteries and hydrogen for transportation to be dual use, vehicle to grid with all the inefficiencies, there is no overbuilding. We have human rights issues, we have environmental issues, but we are gonna find good uses for all the storage. If we don't tie not one arm, right? I mean, like I said in my house, I have two EVs, a 70 kilowatt hour and a 60 kilowatt hour, and neither one can go into the grid. And yesterday when our power, we were down to the last 5%, like the Ford Lightning, if I had been able to plug my car into my house and send power in, I wouldn't have run out and I would have watched more Golden State Warriors, basketball and more Brit box mysteries, but instead I had to actually read a book. Well, Dan, at least you had electricity last week. Some of us, I know electricity at all. Oh, I know, I know. And look, I am a white male in my 60s living in the Oakland Hills, right? So the entitlement multiplies and I think that's the real worry. We are not building storage for low income baby hundreds point communities. We're not building storage for all kinds of low income areas. And that's where the Biden Justice 40 is so brilliant. And I know that some people didn't love Carter's presidency, but we look back on it very fondly. I think no matter what happens going forward, we're gonna look back on President Biden as someone who absolutely jump-started what we needed to do. Whether we continue or not is our own thing, but he really opened the door for what we need here. Thanks, Dan. And Dan, you made this really great point. I think maybe also let me ask Megan away in this well is having more and diverse use cases for assets makes a lot of sense, right? For example, vehicle to grid is another one. Some of the water treatment I'm making you talk about. So Megan, can you maybe comment more broadly maybe as we embrace energy transition, what are some of the other opportunity we can use assets on the ground in a smarter way to decarbonize rather than say developing to new technologies manufacturing them, which all have risks associated with them. Yeah. So I think certainly multifunctional asset deployment is really, really important. I know the water sector really well. So I focused on how you can make water assets multifunctional and particularly supporting the electricity grid, but it's transportation. Certainly it's manufacturing itself. I mean, all of our chemical manufacturing, food and beverage, pulp and paper, there's a tremendous amount of effort trying to think about how to incentivize energy flexibility in the manufacturing sector. And in some ways that's not a publicly held asset, right? It's a privately held asset, but there's still a lot of value in deploying those assets flexibly, especially if your overall production output is not sort of temporarily constrained. If you have more production output than you need, can you modulate that, right? So there's a tremendous amount of effort right now in chemical manufacturing and in pulp and paper and steel and cement to really think about energy flexibility there as well. And I think the water sector, in some ways the water sector is actually really blowing fruit from a deployment perspective because it's publicly owned. There's real incentive. And it's, I mean, quite frankly, like it's operated by the same public utilities commission at the end of the day. And I think that that incentivizes deployment and especially experimentation in a way that some of those private sectors are really only gonna respond to price signals, right? And so if we're gonna see the deployment of those assets, there has to be a clear price signal lead on that. And I can't emphasize enough the degree to which that isn't happening today. Like even in California's grid, the time of use pricing signals are not actually aligned with the carbon intensity of the grid and it's remarkable. Like, if I'm optimizing Silicon Valley clean water for carbon intensity, I run that system completely differently, right? Or Santa Barbara. And I'm getting like effective prices on carbon in the single digit dollar numbers, right? Like I can shed a very large amount of carbon from my operation, which at the end of the day is the goal, right? Like it's not shedding megawatts that we care about necessarily. It's how do we stabilize the grid and supply megawatts where it's needed, but at the lowest carbon intensity possible? And so we often like to look at this on a dollar per ton basis in terms of effective carbon savings or cost savings, carbon savings. And so it's really, really cost effective to do this kind of large energy flexibility deployment from a dollar per ton basis of carbon. Megan, let me build on this a little bit too and please, Osa, Wei and Dan. So you commented on this opportunity for optimization, right? To achieve, say the objective of really maximizing the effect on carbon. So what's missing today? Why are we so far from optimal? Seems to be a well-posed optimization problem. Yeah, I mean, I would say we are far from optimal. It's not the most well-posed optimization problem though, because you're not just optimizing for carbon, you are optimizing for resiliency of your grid. And that is highly local and very, very, yeah, it's just very, very place dependent and very grid mix dependent and very weather dependent, right? So I mean, functionally, it's just very dynamic. It's both place dependent and highly dynamic. And so it is, as academics, we can say, oh, yes, that's a well-posed optimization problem. I think when you get into the real world, it isn't at all. It's actually really complicated. And also, I mean, real-time electricity prices are really hard to plan around. Wastewater driven plant operator, again, coming back to my industry, they don't want real-time dynamic prices because how are they supposed to run their facility in a reliable, cost-effective way when it's completely uncertain what the price of electricity is in the next minute? So that doesn't work either, right? But I do think that, I mean, I don't want to be so negative, but I think there's a lot of challenges from both an optimization perspective because of the temporal and the spatial variability, but there's also real challenges from the human dimension of how do you provide enough price reliability so that the rest of your economy can function? I think that's right. I mean, it isn't a simple optimization. And there's a lot of aspects that make it challenging. When you think about, oh, that's optimized, the amount of electricity and the amount of storage, that sounds well-posed. And I mean, that's why we build models like Switch. But California right now is, for example, grappling with this huge debate over the Colorado River. It dramatically changed the near-term numbers even if we all suspect that this has to go over time. We have the challenges of rebuilding communities like Paradise, these are not small isolated cases, they depend on how much utilities are going to either decide or be forced to invest in hardening transmission. Utilities quite fairly feel that they do not get credit or and they can't get revenue out of certain things they do like what you and I, I think would probably say is natural. And that is everyone who has a smart meter which you need to have solar should also be able to put your energy storage from your vehicle to grid in. So, and those are not small numbers. In California probably has 11 gigawatts of solar behind the meter. That is a revolution for utilities that are not used to being super leading edge even though we have pretty good utilities here, almost everyone in California complains about them. So, it is a hard story. I think the biggest part of this is that what we have not done politically, not technically, is to incentivize utilities to be knowledge creators. And by that, what I mean is there are individual policies that tried it. My own personal favorite, some people love it, some people hate it, is something called decoupling where utilities are not paid in some straightforward, kilowatt is the same thing. This would make it mention, why don't we align prices with carbon since that's our state policy? But under decoupling, a utility gets paid an amount where you get paid more per kilowatt hour. If you sell less kilowatt hours and you get paid less per kilowatt hour, if you say more, based on a target, that the utility needs to analytically and scientifically set and then the utility and then the regulator, the PUC reviews. And that's a very, very subversive but very clever strategy, not just to reward energy efficiency but also to reward smart planning. And of course, there's variances that are allowed for extreme weather events like we just had or for changes in population. So it's not like this locks you in but we don't make our utilities be as smart as possible. And if energy and information are these complex quantities, not doing that does everything Megan said. We throw bad split incentives all the time at us. Well, I do think one place that, there's certainly a lot of opportunity to continue to do work is understanding how different durations storage and how different types of storage come together. The number of especially low flexibility options are expanding and are likely to continue to expand as we electrify more things. And yet, I think there isn't a strong understanding of the value add of deploying those different types of energy flexibility or energy storage assets after we've undergone that sort of transition. So certainly cars are, electric cars are like the big gorilla in the room as an example of like a tremendous amount of storage that should be deployed. But I think there's also a lot of other different types of storage and understanding how we expect different sectors to contribute there and how to provide value to those sectors is gonna be really important in then ultimately figuring out how this whole system works together. Thank you, Megan. Megan and Dan, we're at the end of the hour here and thank you for your talks and contribution. Maybe let me ask one final forward-looking question and get your thoughts. When I think about energy storage, we are playing with a lot of trade-offs, the trade-offs in efficiency, the cost, manufacturability and so forth. Can I ask each of you to pick one trade-off you would like to see broken? So in terms of innovation and technology, what would be the one trade-off you wish that doesn't exist? So we don't have the trade between the two. Do you mean breaking the laws of physics and civil engineering? Or do you mean like political laws? Neither, breaking what is going beyond what's possible today, what future technology can we develop that would have say right now, energy storage for the grid, you trade between cost and efficiency, right? If you want high efficiency to cost is high, low efficiency is cost is large. So I'll go first. Two things, I'm working on one of them, so I'm biased. I think actually in the next decade, we are very likely to have space-based solar. It makes huge sense, it then requires no storage. I probably would make an equally firm bet, but a little more than a decade to have fusion, not fission. If I were to bet where the future of nuclear is, I think it's gonna be light atoms, not heavy ones. But the other thing, you could do right away is already on the books, but there's lots of opposition and that is, it's crazy that we're not using a social cost of carbon. Yeah, well, Dan stole mine. I mean, I think- I'm sorry. As I said before, to the degree that, which we can price in that social cost of carbon and then start to make decisions around that, it's going to incentivize a lot of sort of smart deployment of technologies. We don't have that today. And so we're all operating on sort of uneven playing fields. The goal is to decarbonize here. And so I think that that to the degree to which we can really make that front and center in our deployment strategy of technologies and evaluate on an equal basis, that relative value, we need that. Dan and Megan, thank you very much for sharing your perspectives on a system level for energy storage for the grid and beyond. And for the very insightful discussions, many opportunities are ahead, both in more traditional energy storage, but also in unconventional storage as in the water system that Megan highlighted. Again, it's a great pleasure to host that both Dan and Megan, and please rejoin us in the spring quarter in a few weeks as we have a new series of speakers and stay connected with us on social media. Thank you very much for joining us today.