 Oh, I see. There's a mic. Okay. All right. But, you know, so Michael has had a really interesting career. He's the operational director, operational head of market strategy at Google. But he has done a lot of things within Google in the energy space, including some of the things on striking deals with utilities to go carbon, low carbon and carbon free, green tariffs, etc. And many, many other things he's going to talk to you about. The latest thing is this 24 seven carbon free, which he's going to talk about. But what is also interesting is in addition to the corporate background, he spent time in the White House in the Council for Environmental Quality in the Clinton years. And as a five year old, and he has a very diverse background, and he's got a JD from University Michigan, master's degree from Yale and BS from the University of the South. And in this navigation of the energy, I mean, he is now heading, he's the chair, the board chair of the Renewable Energy Buyers Alliance. And if you haven't heard of that, here it is. This is going to be the driving force of renewable energy, at least one of the key driving forces of renewable energy in the United States. It's in hopefully in the rest of the world as well. Great. Thanks, Arun. Thanks for having me. I'm really thrilled to be here. My voice is coming and going because I'm getting allergies. It's not the cold or anything else like that. So I'll do my best here, but I'll probably need to speak in a little bit of a lower voice. So can everybody hear me? Okay. Okay, great. So I sit within Google's global infrastructure team, and this is the team that manages Google's power consumption globally. And I'm going to talk a little bit about how we've been wrestling with decarbonizing the power supply at Google. That's where I'm going to focus most of my talk. How many of you here are familiar with the sort of 100% renewable versus 100% clean conversation and debate? Okay, a few. Well, we're going to get into that too a little bit. So why do we care about energy? When people think about Google, they don't think about energy, but it's pretty easy. People watch a billion hours of video on YouTube every day. We have nine products that have over a billion users, and we use essentially as much electricity as two San Francisco's. And that's actually a little bit old data, 10.1 terawatt hours in 2018. And we've been in a steady growth trajectory in terms of our electricity consumption for quite some time. So we're growing on the average of 20% year over year. And, you know, we are now making plans for how to manage our infrastructure when we're at 20 or 30 terawatt hours a year and maybe even beyond. It's all powered by a global network of data centers. These are the Google owned and operated data centers around the world. We're now at about 21. And then if you take into account the networking and the points of presence and sort of broader part of it, this is what our compute architecture looks like globally. It's essentially one of the biggest computers in the world when you think about it as a networked entity. This is just, I wanted to give you all a snapshot of what a Google data center looks like. Most people don't think about this, but these are actually huge industrial facilities. This is our data center in Monks Corner, South Carolina, just outside of Charleston. And these are facilities that can use up to hundreds of megawatts each. So they're very, very large and they run a lot of compute at each one of our Google data centers. This is just a little bit of a gif of what the inside of a data center looks like. Not surprising rows and rows and rows of computer servers. So I'm just going to talk really quickly about energy efficiency before I get to the work around carbon-free energy. But it is important, I think, to touch on this. So not surprisingly, the more we can make our compute efficient, the more we can save the company money. And also it helps with modulating our energy consumption. We're doing a lot of things to make data centers more efficient. We keep the temperature in the server rooms higher than what a normal human being would want. Humans are comfortable at 72 degrees, servers are comfortable at 80 or 85. We're doing evaporative cooling, a lot of work. We're also using machine learning to optimize how we run our cooling systems at data centers. And we're also looking at how we actually optimize the compute itself. And we're getting 7x the compute power that we did from a unit of energy that we did five years ago. And then also, this is actually an interesting paper that just came out last week in Science written by John Kumi, who was here at Stanford and a few others. And it basically looked at global data center energy consumption from 2010 to 2018. And they found that, even though the compute has grown exponentially, the actual energy load has been pretty flat, just a little bit of growth. And that's mainly because the ship to these hyperscale data centers, like the ones that Google owns and operates. So now it's really the sort of heart of this, which is really talk about, thank you, talk about how we are decarbonizing our energy supply at Google. If you look at Google's overall carbon footprint, about 85% of it is associated with our data center on energy consumption in terms of our the emissions associated with electricity that we use. So for the last 10 years or so, we've actually been going out and sourcing renewable energy for our data centers. Now done over 50 projects. And it's about over five gigawatts of renewable energy deals. We do that in really three ways. One, if there's a deregulated energy market, we could just go out to the market and say, we want to purchase renewable energy from a new additional resource. And a retailer will typically provide that offering to us and deliver us that energy. If we're in a vertically regulated energy market, we have to work with our utilities to actually find ways to source, but to have them go and source the power for us, or we've actually changed the law to allow us to go source the power ourselves. So for example, most recently in Taiwan, which is a vertically integrated grid with a national utility, we went and actually changed the law to allow companies like Google just to go straight to a developer and sign a deal. And we just did our first deal in Taiwan. We did the same thing in Georgia. We did the same thing in North Carolina. And it's slow going, but it's been, it's worked. And then the sort of third way that we do this is in markets where the wholesale side is deregulated, but the retail side is vertically regulated. We've gone also straight to the developers, signed a wind and solar deal, buy that power from those projects, retire the wrecks and then sell the power back into the grid. So none of these ways is perfect. Probably the deregulated option is the best, but we're finding ways to work around what is a system that is very outdated. And for the last couple of years, we've been able to match 100% of our electricity consumption with purchases of renewable energy. So you total up all of the electricity that we use at our data centers over the course of a year, and we've gone out and signed deals for new renewable energy projects that matches that on an annual basis. Does it mean that we are 100% renewable at all times? And I'll talk about that. But it is a nice sort of early milestone that we've hit. And a lot of other companies are getting into the game. We're the largest non utility renewable energy purchaser in the world. And you can see that there are a lot of other corporates that are now going out and trying to do these projects themselves. And there's just a lot of interest in the whole space around corporates sourcing their energy from renewables. So just to go back to the 100% again, we hit that for the last couple of years. So in 2018, we used 10.1 terawatt hours. We matched that on an annual basis. What does that look like at an individual data center site? It looks like that. So this is data center in Iowa. And you can see that there's a lot of variability in the energy that we're sourcing. So we've gone out and signed wind deals. We're using a little bit of what the clean electricity is on the grid from the utility, but there's still significant times of the day and of the year, where we're having to source whatever the carbon based power is on the grid. So this is a pretty good snapshot of how we're over procuring in some it sometimes. And then we're under procuring at other times, which gets us to our long term aspiration, which is to source 24 seven carbon free power. So everywhere and at all times. So how do we go about doing that? The first thing we've got to do is measure our our footprint, measure our energy footprint of our data centers. And so we developed these what we call heat maps, which is a 8760 perspective of what an individual site looks like from a carbon free energy perspective. This is our data center in Lenora, North Carolina. And you can see that dark green band is actually where we signed a 60 megawatt solar project with the utility under the green tariff program that we had to that we created. But only 67% of the hours are carbon free given here. So I'll just show you guys a few other a few other sites. This is our site in Finland. It's actually pretty clean grid. It's a lot of nukes, a lot of hydro. And we've also done some wind deals. And so we're up to 97% of the hours a year are carbon free at that site. On the flip side, this is our data center in Taiwan. They have a very, very small amount of renewables. They have a little bit of nuclear and a little bit of hydro. And but only 16% of the hours a year carbon free. So how do we start tackling that? And then also just looking across our global portfolio. There's a lot of variation. And this doesn't have all of our newest sites on it. But you can see that Asia is really, really tough. You know, the the the APAC grids are mostly coal and gas with a little bit of nuclear. And as you guys know, there's site in like Japan, the nuclear is being built back. So how do we how do we solve this from a transactions perspective? How do we actually start to fill in these dark hours? So this is a heat map from our data center in Chile. Pretty good grid. A lot of hydro already some renewables on the grid. And we were looking at how can we actually fill in some of these hours? Can we get to 100% 24 seven at this particular site? So we did a solar deal and 80 megawatt solar deal. It's been about three years now, I think. And like the North Carolina heat map, you can see it cuts out a big swath of the day. We came back on top of that and did more solar. Now keep in mind, in the middle of the day, we're actually over procuring it's generating more power than we needed the data center, but we are filling in the hours. And then we've added wind to it, we've gotten pretty close, you know, 96% but again, we were starting from a grid that was very, very clean to begin with. This is an example from a dirtier grid. So to go back to the Iowa example, before we started procuring wind in that market, it was only 17% carbon free. We've done a lot of wind deals there gotten ourselves up to 68% carbon free energy. And this is kind of where we are now. So this is sort of just to give you guys a picture of how are we how are we modeling out? How are we thinking about taking each one of these sites to 24 seven carbon free energy? So in this case, you could add solar to that wind. And you could add storage to that solar and a little bit of load optimization to sort of stretch those hours into the day. Yes, sure. And by the way, you guys, please ask questions at any time. Question about this graph. The color gradient shows that half that 100. The first graph you showed there were lots of variants regions where it was far above 100. Yeah, yeah, yeah. So that's this is a, you know, an 8760 picture of sort of looking at, you know, every hour of the day. And the the you go green when you get to 100% carbon free, it doesn't measure going over there actually to you can't do it in one graph we found. So the other graph was actually covering the overage. So you could do one word. Yeah, you just you have an even keel across the way. But we just found that this is a little bit easier way to to measure for Microsoft. Great question. Yeah, that's what you have been doing that graph in those times. Yeah. But when you plot it this way, erase that information. Yeah, yeah, it's a great point. And I think it gets back to what is the best way to even represent the data in the first place, you know, and when we started this, we didn't really even have a way of representing it. We hadn't seen these maps like this. And so this is what we've been using. But I think I agree. I actually like that graph too. I like the way it shows where you're over procuring because again, it's that you have to solve for that too. Exactly. Yeah, the trick is, is actually doing it in the balancing authority where we have a site. So on a on a global basis, then we're doing this now where we can over procure in these markets. And so we add up all of the electricity we're consuming at every site. We we we're going to add up all of the electricity that we're consuming at every site. And so we're going to add up all of the electricity that we're consuming at every site. We we procure over procure in certain markets like the Central US, the Nordics, but there's other markets like APAC, like Asia, where we're not able to there's just not cost effective renewables at that scale. And so we're matching our annual consumption with these purchases, but we're not actually doing it on the balancing authority. So what does that mean? I'll take for example, our data, we have a data center in Oklahoma, we've procured three X wind at that site. So we've gone out and signed deals that are three times what the load of that site is. Last August, there were days where we were 98% carbon power, because the wind wasn't blowing. And so how so if what we're trying to get at is, how do you actually drive the carbon off the grid? How do you ensure that every hour of every day in the balancing authority where you're operating is covered by clean power? So you're right, we can go out and so we could source three X our load. But what we're finding is is like, when you're at a grid that let's take a grid like Oregon, that's, you know, 90% hydro do you really want to go through a bunch of wind there? That's just stranded in that balancing authority if you can't sort of ship it out and let it and use it to displace and carbon, carbon freezing, carbon balance. Yeah. So again, back to in. Yeah, go ahead. Yeah. Yeah. So yeah, that's a great question. And it is it's pretty flat load. And but there is a little bit, there's a little bit of variation we can do in terms of how we shift how we do schedule jobs during the course of a day. So there's a it's not completely flat. It's got a little a few hills and valleys. And there's some compute that we can actually move and schedule at you know, it doesn't need to be done right away. You know, it's not associated with a Google search. It's backing up data files or that sort of thing. Actually, we'll have some we'll have more to say on this publicly here really soon. But it's a great question. But they are generally pretty high load factor pretty flat loads. Yeah, so we have gotten in the US, we've gotten authority from the Federal Energy Regulatory Commission to buy and sell power on the wholesale markets. And it's a it's a very imperfect solution, you know, and when we started it with our first 100 megawatt wind farm, I don't think we ever envisioned having gigawatts of wind that we're buying and selling into the markets. And so it just gets back to this, you know, when you when you are optimizing at a global level, we are not optimizing for each site, there are you start to distort that you actually introduce distortions into the market. So just back running through this scenario here. This is another one where we've run on the Iowa data center where you do wind solar batteries and load optimization. But even even then we couldn't get this to 100% carbon free in a way that was economically made sense. And so the big question is how do we get this to 100% carbon free? And I think that gets back to it's very hard to transact your way out of these problems, you really need to also be goosing the grids where you're operating and getting those moving towards carbon free. So how do we do this? If I'm sure we can do a little bit with transactions, but we don't that doesn't completely get us there. And it gets back to things that you guys talk a lot about here, which is advancing technology, transactions and finance and advancing public policy. And, and we're looking at sort of all these three areas in terms of how we as Google can move the needle. And this is a this is a pivot for us. We had sort of set these goals to be 100% renewable energy company matched on an annual basis. And we sort of see that as the start. And the real ultimate goal is to get to be carbon free at every location in every hour. And so we're really starting to turn our efforts in this direction. And we're finding that it's not something that's easy to do. It's it's it's arguably a moonshot, especially in places where the renewable resources of today are not very cost, it's cost effective as they are in some places. Can I ask a question? Yeah, sure. Are you from a project called the sea water dialysis of carbonate carbonic acid? Almost. Yeah, yeah. It was a map I saw in a book even provided how the research center, you mentioned the cost of both systems. And then he said it was too expensive. But I'm not sure that I agree with your math. I think that the astro teller has repeatedly used to describe the mathematics and so it is both people and I asked one about it, specifically here on campus, and he just refuses to say. And I think based on my reading of the research, it would be much less expensive than the proposal, which is often mentioned on this campus, which is to convert carbon dioxide into a super critical fluid and eject it back under underground. That adds about 18 cents per kilowatt hour to the cost of natural gas, for example, which is that's crazy. And if you want to keep making those economic purchases, you might consider that there's power gas solutions that you can get by leveraging the natural gas pipeline system as storage. So you don't have to build big tanks, you just have a shipping container size piece of equipment, probably a lot less expensive than other kinds of storage, batteries. Yeah, that's great. That's why I'm here today. So keep them coming. And actually, one of the things I've been talking about is ways that we can work together with you all to try and generate ideas and find solutions to these problems. And your point about gasses, I think last I saw there's $8 trillion of gas infrastructure in the world. And, you know, that's that's something that'll be hard to turn over in other ways that we can harness it in my region. So it's a great point. Okay, yeah, I don't have I don't have a whole lot left and we'll open it up to questions. So just again, technology transactions policy, how can we get work going in these areas to try to solve this problem? So the first area where we've really started looking like as looking at as technology, you know, we look to our strengths around data and machine learning and how can we apply that to solving some of the problems that we're trying to solve with 24 seven. And the first one that we've really gotten to work on is is this these large wind portfolio that we have where we're buying and selling power in the markets? Is there a way to optimize how we do that? For example, we have wind curtailments. We're not getting the full benefit and the value of that power because the way that a lot of power markets work is you have to schedule your assets a day ahead. And you get tend to get compensated higher when you do that than if you sell into the market real time. Well, how to variable assets like when when you don't know the wind is going to blow scheduled day ahead. How can you actually reserve your place in line? And what we've been doing is working in partnership with the deep mind team to use machine learning to take the weather data that's available publicly actually forecast what we think the wind production will be the next day and bid that wind into the market into the day ahead markets. We've been doing this now with 700 megawatts of wind and SPP, which is the Southwest Power Pool, one of the RTOs the United States. And we've been finding that we've been able to get about a 20% revenue uplift, both in predicting the production of the wind and also using machine learning to help us optimize how we actually schedule it how much we schedule and when. And so that's very promising. It's also it's challenging in that we've had to build the system to fit the RTO in which we work. But it is an area where we've been making progress and also can help reduce curtailments of wind. Yeah, I'm sorry. Yeah, so it's really actually getting more revenue. So when you sell into the day ahead energy markets, you tend to get compensated at a higher rate than if you sell in the markets real time. Most wind operators just sell real time because they don't know that they don't know what they're going to be the day ahead. So they can't actually go and reserve their place of line. So that's where we've been able to get more revenue from doing this because it's allowing us to schedule day ahead versus selling in the market. Transactions, I mentioned the Chile deal. This is another example of a way we're sort of looking at how we transact for energy. We've in the past just did transactions for wind or just did transactions for solar. Now we're looking at pairing up storage. Storage costs continue to come down. It's getting more economical to do these kinds of deals. And this is a deal that's actually not been approved yet by the Nevada Regulatory Commission, but it's under review right now. Not going to wood. And it's interesting in that it's a four hour battery. It's 180 megawatts of solar, 170 megawatt four hour battery. And we're using portions of the battery in the summertime to let the utility use to optimize the system. And we'll just take a hit on our CFP for those hours, but it's helping to actually pay for this very large battery. On the policy, there's a lot of work that needs to happen on the policy side that that can help us get to 24 seven and hopefully help others. Just information. I don't know if you guys are you guys familiar with electricity map that word that you can go online and you can actually see what the CFP numbers are for particular grids around the country. But it's actually this this information is not readily available for every electricity grid. The EIA in the US also releases some of this information, but it's sort of hard to get and it's not in real time. But if you had information, real time information on what the carbon performance of the grid is, you can actually, you know, you can have an app or you can you can optimize your appliances or your electricity use for carbon. So one of the things we're looking at now in terms of our data center compute is actually forecasting a day ahead what you think the carbon mix of the grid is going to be in any given hour and can we actually schedule our compute to align with that and save and use more sort of less carbon intensive electricity. We talked about a room touched on this. This is just like how do you empower energy users? You know, one of the, to me, one of the most interesting developments in the energy space these days is that you have this sort of new class of actors in the energy sector that beyond just the traditional players, you have these companies that have very large energy groups that are doing quite a bit on the purchasing side. We've formed an alliance with these other companies called the renewable energy buyers alliance that's trying to break down barriers to allow corporates just to go directly sourced to the market. And there's actually a huge demand for this. Like a lot of companies are now making pledges around sourcing clean power. And the question is can we use our market cloud and work together as a group to remove some of these barriers not only in the US but around the world. And that's something that we've kicked off here in the last year and a half. And it's super promising in terms of what we think we can accomplish. And I'm happy to talk about that in the Q&A. One example is we need to regionalize markets in the US energy markets. So for example, you need to, like one of the things we talked about we talk about is is greening the, greening the actual power grids. Well, how do you, what about changing the grids themselves, expanding the scope of the grid, covering a larger area. It lets you bring in more variable resources whether it's wind and solar and optimize over a larger distance. It also leads a huge cost efficiency. So the existing RTOs in the United States are now delivering on the order of one to $2 to $3 billion of savings a year just from being more efficient. And so how do we create RTOs in places where they don't exist now? And that's something that the Renewable Energy Buyers Alliance and others are looking at and hopefully we'll make progress on here in the next couple of years. And then lastly, how do you actually decarbonize the grids themselves? I don't know if people saw this study that came out of Colorado a few weeks ago but it really looked at what was a business as usual scenario for the state of Colorado. What was a scenario where you gradually took the coal off the grid over the next 15 to 20 years and what was a scenario where you took the coal off the grid in 2025. The most interesting findings from the study was that the BAU scenario was the most expensive scenario and the cheapest scenario was to actually gradually take the coal off the grid but actually taking the coal off the grid pretty quickly doesn't lead to huge cost impacts. And I think this is actually a trend you're gonna start seeing in grids around the country and around the world is looking grid by grid and taking really aggressive postures towards how we can decarbonize them. And with the cost of wind and solar and storage coming down, it's getting to be where you can do these things in a way that actually can make sense for customers because coal is so out of the money and so many markets. And then lastly, what do you do with the infrastructure that's already been there? So what do you do with old coal plant infrastructure? This is a picture of a site that we are, we've built a data center and actually it's just about to go live. It's in North Alabama. It's the site of an old TVA coal plant and we were looking to put a data center in the region and we were asking the utility, what are the really firm strong parts of your grid that are very robust? And it turns out that that's not surprisingly where they have power plants and this particular site had 300, 500 KV lines running out from it and also 9161 KV lines. It was incredibly awesome from a transmission perspective and a reliability perspective. So we're plopping a data center down right at that site using a lot of that embedded transmission infrastructure cut a couple of deals with TVA to source renewable energy from solar farms that are closed by and really trying to recapture the value from that coal plant. It's now been retired. They've blown up those smokestacks and we're using all the transition infrastructure including portions, that's what you're better there. So it's a super interesting way to redevelop sites. We've got another site now in Minnesota that we're trying to do the same thing at and I think looking forward are there ways we can incentivize and look for opportunities to redevelop these sites. Lastly, I know a lot of people here have probably seen this, but I love this slide. I think that the question is like, are we now in this sort of period for energy? Like we haven't seen this sort of radical transformative disruption in energy, but are we here? Because we have finally, the dog is called the car. The clean stuff is actually the cheapest option in many places and getting cheaper by the day. And so are we finally at this moment where we're really going to see rapid disruption and energy over the next 10 years? And I think we are. That's it. All right, let's open it up for questions. Maybe you'll ask the students to ask questions. So we'll then save you all the money. It's a great question. I think there are a lot of benefits. We are finding that in a lot of markets we can now go sign a power purchase deal for solar and wind as a blender with storage. That's actually cheaper than what's on the grid, mainly because of expensive coal and expensive nuclear too, which is what you can talk about. But there are also benefits to hedging power, signing fixed contracts in markets where you have variable power pricing due to coal and gas, I'm sorry, oil and gas prices. And so I think the short answer is we're still seeing, some markets it's cheaper and other markets we're still seeing premiums and some markets like Asia, it's quite expensive. Thanks a lot for the question. Energy storage, I think that's probably a part of it. What is Google doing in that area? And then the second question that I just wanted to I just want to mention negative emissions. Is Google really producing emissions for fossil fuels or are you sort of seeing that? Or are you starting to see that? So you're liking the way it sort of offsets. So on batteries, a couple of few things. One, we've done our first couple of transactions that I showed you the sort of Nevada deal. So just looking at how do we actually add batteries to the renewable energy transactions that we're already doing and we're seeing a lot of opportunities there. You're now like when you go out to the developer community and you ask for, hey, we need 2000 megawatts of solar, you're starting to see storage be blended into that. It's actually economical in a lot of places. So we're starting to do deals. We've got our first, it would be the largest corporate solar storage deal ever done and we're looking at other options. We're also looking on the technology side of like how do you optimization of batteries? Where does it make sense to put batteries on the grid? Can you put batteries in a grid on the grid in a way that that alleviates the need to build new transmission, for example. And then also data center backups. We use batteries now, like in the actual server trays, very small batteries, but can you use utility scale batteries to backup data centers? And that's another area that I think is super promising. Yeah, so you'll see more from Google this year on. You know by the standard student, what skill in that area is that you use your time, remaining time at the center. Listen to everything this guy says. You know, like I've been in this space 25 years. I used to hold out course slides when I was in the White House. It was my first job out of college and it was a slide carousel back then. And so I think this is the most exciting time we've ever seen in energy. And so I think there's like enormous opportunity and I think it's in all of these areas. So if you're interested in policy, there's a lot of work that's gonna happen with changing policies to get to faster decarbonization. If you're interested in technology, there's all sorts of things that need to happen on the technology side. And if you're interested in finance, how are we gonna build all this stuff and pay for it? So I think you follow what's of most interest to you. Which one of these areas is most interesting because there's so much happening right now that you don't need to sort of cherry pick which one you think's gonna have a future. They're all gonna have a huge bright future. There he goes. Student out there. Okay. Yeah, yeah, absolutely. I mean, almost every utility we go to now and every clean energy developer, we say give us a 24-7, give us your best 24-7 deal. So before we would just say, oh, let's go sign a 100 megawatts of solar. Now we're saying, okay, we'll do 100 megawatts of solar but what is like, we give them the heat map of our data center and we say, what's the best you can give? Can you get us to 70%? Can you get us to 85? Can you get us to 65? Depends on the site and where we are. But yeah, we're looking for it everywhere now. And back actually to the question on carbon removal, it's a great question. And I think there's a really interesting, there's a lot happening right now. Microsoft just made their sort of carbon negative announcement and I look at our sites in regions like Asia, Singapore, Taiwan and it's hard to see us getting to 80 or 90% renewable penetration in those markets unless maybe offshore or something like that. And so you're really gonna have to get into direct air capture, potentially carbon capture, that sort of thing. I have to say direct air capture, there's a mean pre-packed between law tools but I believe space is always gonna be more efficient because there's always going to be AHA. Yeah. Can I just say if you want to reboot Foghorn into a more viable company, producing carbon neutral fuel is a nice idea and it has a whole lot of moving parts and it does not fit into a single shipping. Yeah. But if you want to sell to the seawall construction market because a lot of cities all over the world and coastlines everywhere are gonna be interested in keeping the flood encroachment possibility in their near-term future. The seawall construction market for composite materials is something that uses a whole lot of carbon and CO2-chemistry.eu has a couple of CO2 to polymer projects. The one that, I like the one that's, that we'll send you an email about. Yeah, I definitely want your contact info. When we get into this space, we gotta talk to Scott. Yeah. Okay. Sure, what? Um, so in the markets that you talked about, where are you at? Yeah, it's all of the above. So in the, for example, the resource itself may not be as strong. So in very human environments, the solar resources isn't as strong as it is and the desert of Mojave or Nevada. So that's an issue. They are operating under antiquated regulatory regimes where they don't really have a good support system for clean energy. Some of these countries do have feed-in tariffs, but it's on a pretty small scale. And then also there's just not the market that's there to sort of help drive costs down when you have lots of actors and lots of projects and lots of things happening. So all of those things work together to keep prices high. It doesn't mean that we can't do it. The other sort of big limiting factors land and renewables for fire land. And in some places that's really challenging. Singapore, for example, they don't even swap power with Malaysia next door to them. They have one line I think that goes between the two and they have almost no land in Singapore other than the rooftops. But there's not enough rooftop space to cover the load of the country. So it's just those kinds of things. I think they're all solvable, potentially with other solutions. So it'll be something we look at. We're kind of going for the low hanging fruit first. So for example, solar and storage in Nevada, that's like where we're getting, it's penciling out the easiest and the cheapest. And so we're going for that right away. This is great. There's so many questions. What we're gonna do is to have like two at a time. And so that everyone gets to ask questions. So we're gonna ask them. Okay, sure. So what we're doing in this region, sorry. Yeah, it's a tough as region, no doubt. The first thing we've done is really just change. First of all, we don't have a huge presence there. So we're in Taiwan and Singapore. We've just started a building in Japan, but it's a pretty small site. So we don't have as nearly as big a footprint there as we do in the US or Europe. So it is a smaller market for us, even if the population is enormous. And what we really had to do right out of the gate is in Taiwan, we just haven't even been able to source power ourselves. And so like I said before, we weren't to change the law to allow corporates to do direct deals. And we just did our first solar deal in Taiwan and we're now looking at doing other deals. It's a little bit challenging because the resource is still pretty expensive there. And they have a feed-in tariff. And so the feed-in tariff is meant to sort of bring the market to maturity, but it's also adds an expense if you just wanna go and do a project. A project developer can go get the feed-in tariff price or they could sell to cheaply to us. So we wanna be careful about not paying too much for power. But that's an example. We're working very closely with the government there to really open up the space and allow more of these to happen and also make more land available. The government has a lot of government-owned land that can be used to develop renewable resources, whether it's wind or solar. And then also looking at offshore wind in Taiwan. Singapore is a little bit different. It's a deregulated market. So we can theoretically go out and do these deals. There's just no land. And so the big challenge there is actually importing power from Malaysia because Malaysia does have land and availability to do solar. And so that's something that we're working with the government there on now and we'll see. You had to hand out for a long time. I've been, mom and we wanna go together. It's new for us and then mom, I've been. And that's a question. Since many of these companies. So Wahilo has asked me to repeat the question. So the mic picks them up. So optimization of how we use, how our load in the data centers, is that correct? And then moving it, moving it. So yeah, wait, I'm sorry, what was? Market, market, market, market, market, market, market. Okay, so optimization of data centers. Yes, like I mentioned before that we, we optimize them so much now that the loads are pretty flat. There's not a lot of variability, but we are looking at ways to do more efficient and how we get more compute out of every unit of energy that we use. And so we are doing a lot of work around that around how we provision machines and that sort of thing. We are, there are opportunities to potentially shift compute location to location. And I think that's another interesting area that's something that we're looking at. So could you actually follow the sun? You know, or follow the wind with your compute migrating it from data center to data center. I think that's an interesting area that we're looking at. And then in terms of the marketplace, yeah, that we've done some partnerships. We did a partnership in Europe where we worked with a co-op in a local town and three other companies to do a group buy of wind deal. And there are platforms that are doing sort of connecting these corporate buyers with these renewable energy developers. So, you know, the easiest way to do that is to move towards more competitive markets where you just have deregulated markets and people can just, there'll be suppliers that are offering certain products and then, you know, the buyers can choose what they want. But it is something that we're looking at. And I want to say Goldman Sachs or one of the big banks has also done, where they'll go, they'll do a big project and then sell sort of pieces to companies. Yeah. Two questions. One you and then, okay. The source technology that you will use, is that developed in-house or you contract it with? And you mentioned vision as part of your carbon briefing. If you had your drillers, would you like more vision? And if the answer is yes, are you guys doing anything? Yeah. Storage and music. Yeah, yeah. So, what are you doing on storage technology and what are we thinking about on nuclear? So on storage, the answer is yes, mainly on, you know, optimization. I mean, obviously trying to sort of leverage the areas where we feel like we have strengths. So in some of the battery chemistry areas, no, we're really letting others sort of take the lead in that work. But when it comes to sort of how we actually optimize the use of batteries, I think that's maybe an area where we can, you know, actually have something to add. And, but I do think like one thing that is important to think about for companies is, you know, really trying to leverage your strengths the best you can. And being mindful of areas where you may not be as strong. And so I think obviously at the X department at Google has a lot more sort of freedom to experiment and tinker, whereas those of us that are actually running the data centers, we really want to try to leverage our strength as a buyer, as a large buyer, our networking ability, our ability to manage big data, the machine learning capability that we have, those sorts of things. Yeah. Can you see lithium battery technology with time quality? I'm sorry, what? Is it using lithium? Are we using lithium technology? The battery deals that we're looking at are lithium ion batteries, yes. And I can't speak too much about the battery chemistry we're using behind in our data center server racks, but yeah, we are using lithium. And then on nuclear, there are a lot of Googlers who believe really strongly in the opportunity created by, you know, nuclear power, whether it's fission or even fusion. We actually have done a little work in both areas. On the fusion side, we've actually used some of our machine learning capability to help speed up the plasma-based fusion testing. And I don't know a whole lot about that, but it's our research group. And they've just sort of working on the side doing that. And on the fission side, we have supported some increased R&D around advanced nuclear technologies in DC. We've done some work around that. And I think one thing we're worried about looking at our portfolio and looking at the heat maps is that, you know, if the existing nukes come offline, there's actually a pretty significant carbon hit. And what are we going to do about that? And do we need to keep those running a little longer than maybe we would have liked? And then how do you get new nuclear built? And what are the, you know, what do you need to have? What are those characteristics that you need to have in order to get it built? You know, small modular reactors. Everybody I talked to has a different definition of what advanced, what they think is a nuclear. I mean, your borrowers are on the line. Yeah. It's this first thing to support the guys down. We had a lot of debate around whether to call it a clean energy borrowers alliance and not a renewable energy borrowers alliance. So we ultimately decided that we're really trying to solve this specific market barrier right now that we just can't even get to the market. You know, like in a lot of places, you just can't go do a deal for the stuff that's there now. And we want to sort of solve that problem and then really sort of look more long term, especially if you start talking about, well, if you look at the utility, just decarbonize its grid as fast as possible, then everybody's buying clean energy. And so I think that's an area that you may see this group evolving into. I personally think that we need to be doing a lot more work on nuclear as a society. And that there's a lot of opportunity there and we shouldn't take it off the table. But I will say that for Google, politically, nuclear is really tough. There are countries where we just absolutely probably wouldn't be able to do it, or where the public backlash would be overwhelming. And then there's other places that we probably can't. And I think it's something that is we want to have on the table. We are looking into it. And again, it sort of gets to like, how can we really help significantly help the space? Is it by putting out an RFP for nuclear power with these certain characteristics? I don't know. It's someone looking at that. Let me just follow up on this, buyers and lines. As we all know, this problem is bigger than Google. And Google is doing it's part, but when you form this alliance together, my question is twofold. One is Google is 10 kilowatt hours. If you put the full alliance together, what is the 10 kilowatt hours? That's number one. And number two is that when you drive to renewable energy, obviously this is economically affordable. But there are issues of, I know you pushed that on trying to decarbonize the grid. And their corporations, they're obviously, as you know, very well, having lived in Washington, there are issues about social equity that comes in. And how do you address those issues at the same time, some things? So I don't know the number on terrawatt hours. We need to find out. Yeah, it's a lot. I mean, it's tens of terrawatt hours. On the social equity issue, it's something we're very mindful of. When we go and create a new buying program with utility, we're making sure that, we have to make sure that we're not getting subsidized at the expense of other rate classes. So that's one big consideration is that as you're sort of structuring these markets, you don't wanna make sure, you don't wanna set it up in a way where there's winners and losers among the buyers. And then the second question is like, how do you empower, should just the rich corporates be getting access to clean energy? Or how do we make it available for everybody? I think that's a question that we need to answer. That's a great question. Besides that, you also need to consider all your allies, members, and what you've managed on this good night. Yeah. Yeah, I don't think anybody thinks you can get to 24-7 carbon free energy without really getting smart about how you manage demand. What's the perspective on natural gas? I'm sorry? What's the perspective on natural gas? Our perspective on natural gas. Well, we're not going out and doing natural gas deals. You know, I don't like the term. That's the message. Which you'll, you know, but I think like our approach is really firstly, tackling coal and getting, I think there's a lot of opportunity to get coal offline very soon. Sooner is a lot sooner than we thought because it's really out of the money. Gas is tough. I think you're gonna start to see the same thing happening, but in certain markets, it's the only option available. A singing board right now, it's almost complete gas. Yeah. Yeah. Yeah. You know, it's a great question. And I know we're moving away from it to some extent. I mean, you're seeing this sort of trend where like even new building codes are not allowing for it. So, yeah. So I think it's not something that, I mean, it's something that the building team is aware of and are trying to sort of move their direction. Well, let's thank Michael for a terrific thing to discuss and talk and discuss with the students.