Google's Climate Tightrope: AI Innovation vs Emissions | Ep203: Kate Brandt

Bryony Worthington  

You are out there amongst the leading companies having set challenging targets for yourself. But you know, the latest sustainability report doesn't paint a glowing picture, because, I mean, as I understand it , based on your 2019 baseline, emissions are actually up quite substantially. Will you ultimately say we're going to move the date back? Or are we going to have to do offsets? Or will you need to say, you know, we tried and we failed. It's hard. 

Kate Brandt  

So this is what I'm saying. Every year we'll continue to share more about what that roadmap looks like. And in our last report, what we really leaned into was the transparency to say, 'here's where we're making progress in our Scope 2,' through clean firm, by the way, through not buying unbundled renewable energy certificates, but actually buying real power, which is something that we have a lot of conviction around. And also it's a very high growth company, right? We look very different today than we did in 2019 and that's okay, like that's why we set this ambitious goal. So you'll see our next report come out over the summer, and we'll continue to share more details on what that pathway looks like.

BW  

Hello, I'm Bryony Worthington, and this is Cleaning Up. My guest this week is Kate Brandt, Chief Sustainability Officer at Google. Kate has a fascinating background, having spent time in both the Pentagon, and advising President Obama on energy security and sustainability issues. In her current role, she oversees a team of people working to reduce Google's environmental impact. Google has set itself an ambitious company-wide climate goal of halving their 2019 emissions by 2030 but their latest sustainability report showed that they were off trac, as the company has grown and increased their investment in energy hungry artificial intelligence. As providers of curated information at a global scale, Google has a pretty unique intersection with the climate problem. Primarily, their impact comes from the electricity they buy, but they're also able to guide customers' decisions through the provision of information and through their parent company Alphabet. They've made some pretty bold investments in innovative, disruptive companies, so measuring their impact is arguably more than the sum of their Scope 1, 2 and 3 emissions. Nevertheless, missing a clear target can't be very comfortable. So, I wanted to ask Kate how she's approaching all this, and what she personally does to remain calm and positive in the face of hard problems and the dire climate science. Please join me in welcoming Kate Brandt to Cleaning Up. 

BW

Kate, thank you so much for joining me today. 

KB  

Great pleasure. 

BW  

Great to have you on this podcast. We're going to start, as we always do, by asking you to introduce yourself in your own words please.

KB  

Absolutely, I'm Kate Brandt, I'm the Chief Sustainability Officer here at Google, and I've had the great pleasure of serving this role for almost a decade now, having the opportunity to work truly with colleagues across the company on how we think about environmental sustainability in our own operations, as we operate our offices, our data centers, engage with our supply chain, and then also as an information and an innovation company. How do we think about utilizing our technology to enable businesses, policy makers and individuals to drive towards more sustainable outcomes? 

BW  

That's quite brief. But how did you get here? Talk to us a bit about your backstory of how you ended up here

KB  

Yes. So prior to coming to Google, I spent about eight years in Washington, DC, serving under President Obama. So in my last role — I was the first Federal Chief Sustainability Officer —  so I had the opportunity to partner with the entire federal government around its sustainability strategy. And as you may know, the US government is the single largest energy user in the world. So it was a tremendous opportunity to work at scale and to set out a really bold ambition for what was possible in terms of adoption of clean energy, electric vehicles, engaging with major federal suppliers on their own sustainability practices. And then prior to that, I had several other roles in government. I was a senior official at the Department of Energy as well as at the Pentagon, looking at energy security and energy technology, and had a couple other roles at the White House as well.

BW  

That's fascinating, and you had a little bit of a focus on the military, didn't you, specifically the Navy? So tell me a little bit about that.

KB  

I did. Indeed, I spent about four years in the Pentagon working for Secretary Ray Mabus, who was the Secretary of the Navy, incredible leader and still a dear friend and mentor of mine, and I was his energy advisor. And so at that time, we were really looking at how energy security could really serve a role in enabling the Navy to accomplish its mission more effectively. So the Department of Navy includes both the Navy and the Marine Corps. So looking at everything from powering ships and planes on advanced biofuels to bringing clean energy and clean vehicles to military installations. And even at that point in the war in Afghanistan, thinking about, how do we reduce fatalities from fuel supply lines by bringing advanced energy technologies into forward deployed settings? Really innovative work that the Marine Corps was doing. So that was a tremendous portfolio, and one where I learned a huge amount, too.

BW  

I can imagine, but what was it before that then that led you to being that energy specialist with this brief?

KB  

Yes. So before I went into government, my academic studies and graduate school were all focused on energy security and climate security, and I was always really fascinated by the role of natural resources in peace and conflict. And so it was a tremendous opportunity to take a lot of what I had looked at more academically and researched, and then really bring it into practice in this work with the Navy and Marine Corps, who were so interested in innovation, and really looking at this nexus of energy security and advanced technology.

BW  

And of course, energy security is now the topic. Is that something you still keep an interest in? Is it something that informs your work here today? 

KB  

Absolutely, you know, one of the things I love about doing this work at Google is we are such a global company, of course, we're US headquartered, but we have a presence across the US, across Europe, across Asia Pacific and absolutely, you know, energy is a critical input to our business. And we have, going all the way back to 2010, looked at how we can procure clean energy in a way that gives us supply for our operations, but also more broadly, enables greater clean energy supply on grids. And increasingly, over time, we've really focused on clean firm power solutions, so everything from advanced geothermal, advanced nuclear, long-duration storage, carbon capture and storage. Because, as we know, we're in a moment where, after several decades of load growth not taking place, that load growth is increasing, which is positive from an economic perspective, due to increased manufacturing to electrification of buildings and vehicles. Also in the global south, a lot of increasing air conditioning, which, of course, is a quality of life improvement. So we're seeing this moment where we need to bring more electrons online. And so how can these clean, firm technologies really play a critical role and provide more capacity and drive this next innovation and this next set of technological solutions? So it's very connected to the work I was doing in the military. We were thinking about similar things. We had a goal with the Navy, how could we bring a gigawatt of renewable energy online to support military installations, with a very similar philosophy.

BW  

And just sticking on that topic for a bit? I mean, it is such an interesting topic, because once you shift to an electron-based system, really, one of the challenges is that it does work better if you have more interdependence between countries, right? You know, big cables connecting grids leads to an easier to balance and cheaper system. But then there's this tension, isn't there, about countries wanting to remain domestically in control of energy, and I just don't know where we end up. We just did an episode with the deputy energy minister in Greece, and he was absolutely clear that Europe needed to do a top-down plan, interconnect all the markets, make it really efficient. But then at the same time, you've got Norway saying, 'hang on a minute, Germany,  you're forcing our prices up. We're not happy about that.' So there's a tension, isn't there, that it should all work, but there is this tendency back towards a more nationalistic view, to protect your own country's needs.

KB  

Yeah, indeed. And I think as a policy matter, in terms of ideal setup for grids, we've always looked at Nord Pool, for example, as such a great model, where you can wheel power across jurisdictions where it creates greater flexibility. The opportunity of demand response to bring renewables from places where they're more available to load centers where they're less available. So indeed, I think that is a more ideal setup in terms of how power systems work. And frankly, we have a lot of those conversations too in the Asia Pacific region, where we've been very upfront that we have this bold goal around 24-by-7, carbon-free energy for all of our operations on every grid, by 2030. And every year, we put out the map of how we're doing. And we, and everyone else, is struggling with this in the Asia Pacific region. And there too, it really seems like grid interconnection is what's going to make it more possible, right? So I think that's very much where things need to go in terms of grid decarbonization and, frankly, just having a wider variety of supply. But then indeed, you're faced with geopolitical realities. And this is exactly what I studied early in my life in school. So I think there is absolutely a tension there.

BW  

And maybe we just have to hold these scenarios in our minds. That is a scenario where interconnection becomes the default way of managing intermittency, that's one scenario. But if it's held back, maybe we do need much more firm clean power. And then another scenario is we do crack the small scale nuclear, for example, which I know Google has signed a very exciting contract. That's another scenario, right, where less interconnection probably means more reliance on these firm clean sources. 

KB  

Indeed. I think in all scenarios, we need more of the firm clean sources to complement large scale renewables, to complement the natural gas on the grid that's been with us for a long time. That, I think, will continue to be with us for many years to come. And indeed, I think small modular reactors are a very interesting, exciting technology. Speaking of my time at the Navy, I've been looking at this for a very long time. And in fact, Admiral Rickover and the nuclear Navy, frankly, pioneered small modular reactors. And they just put them in submarines and on aircraft carriers, which I've been on several of. But in this case, you know, I've always thought small modular reactors are such an important part of the grid mix. It would be tremendous to see the maturity of the technology, the ability for it to be deployed. As you were alluding to, we signed a first of its kind agreement with Kairos Power, which is a small modular reactor company, for an order book of projects — 500 megawatts over a period of time to really provide that certainty of demand. This is a model that actually the Department of Energy has proposed as being a really positive way to scale and deploy small modular reactors. And so we see that as a really important part of the grid mix. In addition to other clean firm solutions, for instance, advanced geothermal, we've also been very actively involved in that space, also have signed first of a kind agreements there. And yes, we see that this whole broader suite of clean firm technologies is going to be really essential.

BW  

There was an assumption that wind and solar and batteries... and in fact, wind and solar and batteries have done a huge amount to diversify our electricity mix, and we should make the most of those renewable sources. But is this focus on firm, clean a realization that whilst it's the fastest to add to the grid, it still takes a lot of resources to get through to a point where you're just really displacing fossils?

KB  

We need both, right? And I think that was very much the vision in 2020 when we put this goal out there for ourselves, but it's really a vision for the broader grid of 24-by-7 carbon-free energy. Because that's how you're able to better match renewable energy sources. And by the way, I think another area that's gaining more traction that we've spent a lot of time looking at at Google is, of course, load shifting too. It's absolutely critical. And you know, we've been in this space, of course, with Nest, which has been a part of the Google family for a very long time, which is enabling homeowners to more efficiently use load within their homes. But also, we've really pioneered that in our data centers. In fact, back to your point on energy security, we had already been looking at this in terms of, how could we shift load to times of day when the grid is cleaner? Or even to locations where the grid is cleaner for non-critical tasks, right? Because some compute tasks are going to be absolutely critical, they have to be served immediately, but some don't. So being able to have that capability. We really accelerated our capacity to do that during the energy security crisis in Europe, when we needed to support grid operators in doing that. And so that's something that we're continuing to pioneer, and even looking at that as it relates to AI loads as well. And I think that's going to be absolutely essential, especially in this period of time where grids globally are facing load growth. It still takes time to get steel in the ground, to get permitting done. If we can more effectively utilize the load we have, that's going to be really, really effective. 

BW  

Absolutely, and I think the crisis in Europe where   the reliance we had on gas coming in from Russia was suddenly removed, almost overnight, and it meant that there was a sudden need to increase the efficiency of the system. And it's interesting, moving to America and even to Silicon Valley, I've noticed that the sophistication of how utility companies interact with consumers is just not there, compared to little UK, where we have really sophisticated demand signals now coming from companies like Octopus. And you look at the incumbent here, and I still get notices that say,  'You seem to be really inefficient next to your neighbor, because your electricity consumption is super high.' And I'm like, but I have an electric car, so of course, my consumption is higher, but there's a lag in understanding this new landscape that we're moving into, where electricity is displacing fossil fuels in other places, and it's giving you flexibility. I'd rather get a letter saying, 'Hey, we noticed you have an electric vehicle. Would you like to trade that in and out of the market?' That would be super cool.

KB  

I know, indeed. Well, I was mentioning the Nest Learning Thermostat, which we still have. And in fact, we launched our latest version of that just a few months ago. But there is another component of that work called Nest Renew, which is kind of the next frontier of that thinking of, 'how do you create greater flexibility for homeowners, not only through thermostats, but through vehicles or having some storage at home. And so that company was spun out and a dear friend of mine, Ben Brown, is now the CEO. And I think they're really pioneering in that space and really trying to offer even more sophisticated tools for homeowners. And then that, in turn, is providing more flexibility to grids, and especially if you can pair that with storage on those grids, that can be brought online by utilities or in partnership with large energy users. I think that's where it starts to get really interesting. 

BW  

Yeah, totally. And I'm just coming back from a conference in Poland where I met a company that spun out of Alphabet called Malta.

KB

Malta salt batteries. Yeah, exactly.

BW  

It was great to meet this company. Because for a long time I thought — that's another scenario, right, where we start to store this, but store it as heat. Because, if you look at the Polish grid and the Chinese grid, these thermal operators are not just providing power, they're providing heat into the network. Combined heat and power (CHP) on coal is quite common. So you can't just throw up a bunch of solar panels and say, we've got a like-for-like replacement, you really need to think through how do we get that heat load and this molten salt technology, it's just super exciting. And I don't think people realize quite how much Alphabet has done in terms of those spin outs. But do you have an oversight of the Alphabet portfolio?

KB  

We do absolutely keep in touch with the companies that have spun out. And oftentimes, Alphabet maintains some investment share in those companies. And where Malta came from, and where a lot of the spin outs come from, is a sister company to Google within Alphabet called X, our moonshot factory. And indeed, X has been doing pioneering work in the space for many years, and we're increasingly seeing a lot of fruits of that labor. So another one of their more recent spinouts is a company called 280 Earth, which is doing a really novel form of carbon removal technology in the form of direct air capture, but co-locating it with data centers, also creating a fresh water supply through the way that they're doing the direct air capture, which then can support the data center if it's doing water-based cooling. So that's a really interesting technology that's come out of X. And then another that's still within the X family that I'm very excited about is called Tapestry, which you may have heard of, which is X's moonshot for the grid. And so they're really looking at how we can use AI, as well as working with Deep Mind and our weather models like WeatherNext to help grid system operators better plan for their grids of the future. So they've had a really deep partnership with Chile to do grid modeling for them, to help them model, for instance, what would it look like if they pulled off a coal plant? Or what are going to be the impacts of extreme weather events on your wind capacity? So seeing at least a 15% improvement in the way that grid modeling is being done, and there's several other grid operators that they're partnering with as well, and they have some upcoming announcements there. So really exciting applications, and it's really tremendous to see all the great work that's come out of X.

BW  

And the promise of this coupling of technology and data is to find out those optimizations, I suppose, of where things could be done more efficiently. And probably the grid is really fertile ground, right? I'm involved in an NGO in the UK, which does AI solar forecasting. And when they went to speak to the grid about their solar forecast, they were like, 'Oh, we don't do solar forecasting.' But we have 11 gigawatts of solar on the grid, so surely it really matters how much insolation you're expecting? And it just had never had a commercial value before, and so no one had done it. And there's quite a lot of old systems that are in place in the energy system that are, surprisingly, not digital. 

KB  

Indeed. And I think there's such a huge opportunity with all the advances we are seeing in weather modeling. We've had pioneering weather models that have come out of Google DeepMind that are enabling a 15% improvement in forecasting, and that's really tremendous for exactly what you're describing, whether that's solar prediction or wind prediction. We've done pilots in the past getting 36-hour in advance wind prediction, which makes wind more dispatchable, it makes it more valuable. So I'm very bullish on that. And we've also seen really powerful applications in this regard, more from a climate resilience perspective. So for instance, an area where our Google research team has gone very deep is around flood forecasting. And what the team has now done is built a world-class global hydrologic model, where they can now predict in countries around the world riverine flooding seven days in advance. And so that, I think they're now covering over 300 million people, and that's enabling governments to give alerts. We've even seen aid organizations giving out aid in advance of a crisis. So that's another really powerful application of this kind of prediction. 

BW  

How is this funded within Google, because you're a huge company, with nearly 200,000 employees, and a huge amount of work going on. But so something like that, where you'd expect the public sector to take on. What's the business case in the world of Google? How do you justify all this wonderful expenditure on public good?

KB  

So it sort of depends on the product area, but you know, for instance, we've also done a lot of work on methane mapping, so we partnered with the Environmental Defense Fund on something called MethaneSat, which is a satellite that Google Cloud is then ingesting that data, utilizing AI to provide better insights to the oil and gas industry on where methane leaks occurring, which as we know is very high global warming potential greenhouse gas. If you can detect those leaks, it's very, very valuable environmentally, but also to those companies. So that's going to become a service that Google Cloud can sell, something like what I was talking about with flood forecasting there. There’s a partnership model, where we're working with the UN with their Early Warnings for All initiative with the World Meteorological Organization to provide that service as part of this public good, which is around those kinds of early warnings. So it depends, and in other instances, we're building these features directly into products because they're providing something really useful to our users. So another great example of AI providing optimization is in Google Maps. Hopefully you've actually seen this, the little green leaf has fuel efficient routing. So in this instance, we're using our AI capacity in the background to better understand the gradient of roads, traffic conditions, and other factors to then show you or me, as we're driving home, or to pick up our kids, what's the most fuel efficient route. That's providing a really useful service within a product that we all know and love, where we can save money, we can save fuel, and then it reduces emissions. And actually, we've already enabled users of Google Maps to avoid 2.9 million metric tons of CO2e through that tool, like taking 650,000 cars off the road. So there's going to be a variety of ways that we bring these tools to to market. There's a variety of ways that they fit into our business model, just depending on the application. 

BW  

And for Google, how much of those products actually go into your consumer offering? Do they have to be global? Is that little green leaf here in the US, or is it global? Is there a  benchmark you have to get over before it can be really made into a product for Google?

KB  

Again, it depends a little bit on what the go-to-market vehicle is. But, for instance, for something like fuel efficient routing, we started that in the US, then we brought it to Europe, and then we've been rolling it out in other markets as well, just as there's demand pull and market fit. For other initiatives, we've done really excellent work using AI to do traffic light sequencing, and that requires deep partnerships with city traffic engineers. And there, we've picked a flagship set of 14 cities globally that represent different kinds of cities with different conditions and of course, where we have partners that are excited to do that work with us. And in that instance, actually, just to say that's been such a fantastic piece of work, we've been able to prove that we can reduce starting and stopping events by 30%. And as you may know, emissions at traffic lights are 69x what they are on the open road. So potentially, really large air quality benefits there, but also a reduction of emissions by 10%. So again, it will depend, but generally our model is we like to pilot, we like to find the right partner to work with, to learn alongside, and then scale from there.

BW  

This optimization stuff is really exciting, but I suppose if I were being a bit of a Devil's advocate, I'd say 10% here and there are marginal improvements on efficiency. And the really big shift is we're going to stop burning diesel and petrol or gas in our vehicles, and we're going to move to electric vehicles, right? And that will give you an order of magnitude more efficiency. What can you imagine being able to do to help with people's understanding of the benefit of electric vehicles, because there's a lot of disinformation around on the web and everywhere. 

KB  

Yeah, so this is an area where we've placed a lot of focus, because, of course, our mission as an information and innovation company is to ensure when people come to Google, we're giving them the highest quality information. And so we've invested quite a bit in ensuring our experience on Google search is equally good for an electric vehicle as for someone coming to Google Search to learn about any other form of vehicle. And so what you'll find is we have features now where you can see a calculator that shows you what are the incentives that are available to you in your geography? And then what is the difference in cost between an electric vehicle and another form of vehicle. There's also a feature that's a range calculator, because we know, of course, one of the things that, even in my family, we have this conversation, right? Can we take this vehicle easily up to Tahoe when we want to go skiing? And so there's a range calculator that enables you to see what that journey would look like for you. How many times would you need to stop, etc. And also, just in terms of your question on what are some of those really, potentially large-scale opportunities? Another topic that we've looked at really closely is, of course, the hard-to-abate sector of aviation. And as we know, I've been working on advanced biofuels for almost two decades, it's absolutely essential. It's taking a long time to come to market, it's expensive, challenging. Same with engine switching. But what we know, of course, from the IPCC, is actually roughly a third of all global aviation emissions — that's about 1% of all global emissions — are actually from the contrails, those plumes that you see behind the jets. So we said, well, what if we looked at that challenge? And so we partnered with American Airlines and Breakthrough Energy, and used AI to better understand why those contrails are forming, which is a combination of time of day and altitude and air temperature, and then how to avoid them. We actually were able to show through 70 test flights that American Airlines flew, that we could reduce them by more than half, by 54% and now we're working to scale that more broadly with airlines with Eurocontrol, who operates the most busy airspace in Europe, over Western Germany. So that's one of these examples that I think is so powerful. To say, here's a very hard-to-abate sector, expensive to decarbonize. If you could globally scale that solution, you could reduce 0.5% of all global emissions with just one AI solution. 

BW  

I want to go into that in a little bit more detail, because I saw that announcement — super exciting —  but I couldn't quite understand why a white plume of control, which we would think would be reflective, would cause that.

KB  

It's radiative forcing, so it's trapping the heat.

BW  

And that overwhelms the fact that it's creating a white surface that would reflect light? 

KB  

Yeah, it's the radiative forcing, exactly. 

BW  

Thinking more about corporate targets and Google. You are out there amongst the leading companies to have set challenging targets for yourself. But you know, the latest sustainability report doesn't paint that glowing a picture, because as I understand it, based on your 2019 baseline, emissions are actually up and up quite substantially. So how does that impact your role? And how are you feeling about this bold target that is proving very hard to hit?

KB  

Yeah, I think that the nature of these time bound targets is very helpful, because it gives us a framework to work within. And that was very much the vision, going all the way back to 2012 when we set this 100% renewable energy match target. And you know, that enabled us to really focus on innovation. We like to say at Google that we like hard problems, which is really true. It's one of the things I love culturally about this place, because I feel very similar. And by setting that goal for ourselves, it enabled us to really push the bounds of, how do we procure clean energy more effectively? How do we do it at scale? And so I think the same thing goes for our 24-by-7 carbon-free energy target, for our net zero by 2030 target. It's very helpful to help us to focus on what is the innovation we need. How do we solve this problem? And of course, our approach is not a paper based exercise. It's not how do we solve this for Google, just so the numbers look better, but rather, how do we have real world impact in a way that's actually meaningfully reducing our impact, but furthermore enabling the whole system? And so we were talking a moment ago about these clean-firm energy technologies, right? We've played this really valuable role of being an early adopter, creating an early demand signal, which brings down costs, which de-risks the ability of everyone to access these technologies. So that's, philosophically, how we think about it, in terms of the footprint itself. Of course, all the corporate footprints have different dimensions to them. For us, we've made tremendous progress on our Scope 2 footprint, which we've been working on for a very long time, electricity being a really critical input to our operations. And with our 24-by-7 goal, we're 64% of the way there, which is, I think, quite good in light of the fact of how challenging this is. Also being such a global company, we spoke earlier about the challenges in the Asia Pacific region. So we're making good progress on our Scope 2. On Scope 3, that's an area we've been working on for less time. Like most companies, you know, we set our first Scope 3 target in 2021 so we are making progress, but we have a ways to go. And there, I think that a lot of the solution is also going to be clean energy. So we've been working with our suppliers on clean energy addendums, those really asked them to commit to clean energy. We've also been able to do some innovative partnerships, for instance, with BlackRock in Taiwan to get a gigawatt of solar capacity built that we will have some access to, for our Scope 2 footprint, but also our suppliers will get access to. So we are making progress. We have a lot of work left to do. And also like so many companies, we've been looking at the concept of a Climate Transition Plan. What are all those key reduction levers that we're going to be utilizing? And what does that path look like? And I think increasingly, that has really come into focus, and kind of the year over year, as we report on our footprint, we'll continue to share more of what that plan looks like.

BW  

But here we are five years away from 2030, a 54% increase on a baseline that should be 50% down in five years time. 

KB  

49%. 

BW  

Okay, but still the wrong direction. 

KB  

Yeah. 

BW  

So will you ultimately say we're going to move the date back, or we're gonna have to do offsets? Or will you need to say, you know, 'we tried and we failed.' It's hard. 

KB  

So this is what I'm saying. Every year, we'll continue to share more about what that roadmap looks like. And in our last report, what we really leaned into was the transparency to say, here's what we're making progress in our Scope 2, through clean firm. By the way, not through buying unbundled renewable energy certificates, but actually buying real power, which is something that we have a lot of conviction around. Back to my point of, we need to actually be doing this work in a way that drives value and is not just passing around paper. So that's a part of the story there. And also it's a very high growth company, right? We look very different today than we did in 2019 and that's okay. Like, that's why we established this goal. So you'll see our next report come out over the summer, and we'll continue to share more details on what that pathway looks like? What are those reduction levers? And a lot of it is going to be clean energy, both for us and for our suppliers, but it's also, how do we have lower carbon building materials and things like low carbon concrete, lower carbon steel. 

BW  

So just in the day of the life of a sustainability officer, of a company this scale: Do you have an influence over, say, where the next data center gets built? Because you could build it in a state with a lot of wind and solar and excess capacity, and it would be a hard problem, but not an impossible problem. Or you can build it in a place which is running 90% coal just because of the nature of their geography or their history. Is there a decision making tree that says: No, we use Country X over Y because of their sustainability?

KB  

Yeah, it's always been part of the equation. And if you look in Europe, for example, where a lot of the earliest data centers were built — Norway, Sweden, Finland, Denmark, where there's already a lot of access to carbon free energy. Or our very first data center was built in Oregon in The Dalles, where we have a lot of access to hydroelectric power on those grids. So it's certainly a factor. We also need to go where our customers are for the cloud loads. We need to go where the business needs us to go. So it is a factor, and then, of course, when we do go to those places, you're seeing us show up and really working towards, how do we better plan with the utility that is in that geography for our load? And you may have seen we made this announcement last year on something called the CTT, the Clean Transition Tariff, or rate, which is a vehicle to plan directly with a utility for what that load looks like, what the shape of that load is going to look like over time. And so they can partner with us to say, 'how do we meet as much of that as possible with clean energy solutions?' 

BW  

There's also an educational role you can play with utilities, especially in some parts of the world where this transition is a little bit further behind. I do some work with a group that looks at, how do we repower coal assets? Because I think the assumption has always been, 'build out loads of renewables and the coal assets will just quietly drop off.' But they provide quite vital services in many cases. And, you know, they're powerful grid connections, and there's a workforce there. So we've been trying to think through the reinvention of those coal assets, rather than just hoping they go away. And things like Malta, where putting a molten salt store actually acts as a kind of adapter plug for that asset, right? So that it can take excess electricity, or it can convert it into different steam temperatures. But if you ask a utility company, maybe in the Asia Pacific region, they won't necessarily know about that technology. What they'll know about is, 'oh, yeah, we've heard about co-firing with ammonia or we're going to do CCS when it's cheap, which may not happen. So there's a kind of educational role about knowing what the range of technologies are that are emerging, that are just this close to commercialization, that then change the models.

KB  

Indeed, and being that voice into the customer and expressing our preference for clean load. And just a couple vignettes I would share with you, based on what you were just saying. One is, I don't know if you're familiar with the story of our Widows Creek data center in Alabama, but this is a data center that was built on the site of a retired Tennessee Valley Authority TVA coal plant. And so as you're just articulating, all of this infrastructure was already there, but the site wasn't being utilized anymore, so we actually were able to build a data center on that site to work with TVA to help meet a lot of the load demand with clean energy, and then take advantage of all of that existing infrastructure. And then, equally in the Asia-Pacific region, as you're articulating, they're just not as far along: in Taiwan, now going back many years ago, we were able to work with the government to put in place the first rules that enabled direct power purchase agreements for companies. And we were able to do the first ever solar procurement under that new policy regime. So, yes, I think there's always really nice opportunities for us to work with utilities under something like a CGT, to work with governments to help inform how they're thinking about their energy policies, so that it helps to unlock more access to clean firm power.

ML  

Cleaning Up is brought to you by members of our new Leadership Circle: Actis, Alcazar Energy, Davidson Kempner, EcoPragma Capital, EDP Portugal, Eurelectic, the Gilardini Foundation, KKR, National Grid, Octopus Energy, Quadrature Climate Foundation, SDCL and Wärtsilä. For more information on the Leadership Circle, please visit cleaningup.live, that’s cleaningup.live. 

BW  

Have you noticed any change in discourse in recent months and weeks? Because there's been a period of change in the US, we can't not acknowledge that. But as a global company, has it had an impact?

KB  

Yes, and I think we have been finding, especially on a lot of what you and I have been discussing, in this moment of load growth globally, but here in the US, how essential it is to see a wide variety of technologies that can be accelerated and advanced. And I think there's a lot of shared conviction around AI innovation, around the importance of that innovation happening here in the US, and then the potential of these clean firm power solutions like advanced nuclear, advanced geothermal, long duration storage, CCS, right? So I think this is where we're seeing a lot of opportunity for collaboration and a lot of opportunity to do this knowledge sharing and exchange. So I think there will be areas where we can be very productive, and keep advancing. And then, to your point, of course, we are a global company, so we do need to take a global view. And we are very active in many different geographies.

BW  

And so Google has soft power in going out and talking and influencing policy. Will you be going to COP30 in Belem and Brazil? Is that on your card for this year?

KB  

Indeed, yes. I've had the opportunity to attend many COPs in my time. And in addition to everything we've been discussing, too, I think what will be really powerful about the conversation that we'll be able to have at COP30, a conference in the middle of the Amazon, is the nexus between the climate agenda and the nature agenda. And we're seeing tremendous power of AI, not only to advance a lot of these more efficient carbon-reducing solutions, but also to really support nature and to support nature based solutions, and so that's something that I'm spending a lot more of my time on, and also that we already have a really great portfolio of work. I think some of our longest standing work has been through our platform called Google Earth Engine, which has been around for over a decade and is a geospatial tool that enables you to see change on the Earth's surface over the last more than 30 years — I think now 40 years — but also to have the ability to do real time monitoring. And so we've had great success working with companies like Unilever that have zero-deforestation goals to use this tool to help them better understand where deforestation is potentially happening in their palm-supply chains. To reduce that, to do monitoring and moving that into other commodities. We've done work on this in coffee supply chains, for instance. So I'm really excited to continue to lean into that space, and that's something that we'll really be keen to talk about when we're in Belem.

BW  

Yeah, because that sort of fits into your observational...  This is where the power of AI and digital data gathering and surveillance can really indeed change our understanding of what's happening. We just did an episode with the Scripps Oceanographic Institution down in San Diego. And it really struck me that the people who have been monitoring the makeup of the gasses in the atmosphere, they've been doing it on a shoestring. And it's an amazing data set, but it's not valued, because it's not new, it's not cutting edge. It's just this really solid data set that goes back to the 50s, and, you know, we could lose it, right? We take for granted this ability to know what's happening on the planet, but it takes people, it takes money, it takes institutions. And I do think that, fast forward to where we've got way more instruments out there, monitoring, listening, seeing what we're doing, coupled with what Google's doing to develop artificial intelligence or even generalized intelligence, there'll come a point where we are much more attuned to the state of the planet, right? At the moment, it feels like our data is very atomized and not very surfaced in terms of the risks. Ideally, we'd have a dashboard, a global dashboard that we're all very tuned into. Is that where you think Google might be able to add value? 

KB  

Yeah, I think the convergence of data sets is a very powerful use case for AI, even pre=dating AI. One platform that Google has built is something called Datacommons.org which is built on a schema structure that allows you to take multiple public data sets. They have over 200 different public data sets that then can be queried in natural language, but also could just be looked at in ways that would have been so much more challenging to look at: population data next to weather data next to famine data, for instance, and having this kind of knowledge graph that structures that data and then uses natural language on top, it makes it so much easier to make sense of these large public data sets. So that's something that we've been invested in for a long time. But then indeed, especially, as you mentioned, with climate risk modeling, for instance, or better understanding climate risks, indeed, I really see us, even just with the platforms that we've built, trying to drive towards more conversion. We talked a little bit earlier about flood forecasting. Separately, we have FireSat, which is a partnership with the Moore Foundation and several others, which is a constellation of satellites. In fact, the first one just launched earlier this week, and this will allow for much more real time monitoring on a 20-minute basis, 10 by 10, the size of a classroom-scale wildfire. So that can do much more early wildfire detection. Similarly, we've been doing a lot of using AI to better understand heat and heat islands, and so I certainly see us having the ability to converge a lot of those capabilities so that those then become layers in one interface, where you can see multiple different aspects, and then do really forward looking planning, sort of bringing in a lot of our weather and modeling capabilities. So yes, I see a huge amount of opportunity here. 

BW  

And in a way, the fact that this is being done at a time when the costs are coming down, the compute, the tools, the rate of development of the tools is incredible. And I suppose Google, perhaps people aren't aware — maybe many are — but the fact that you've open sourced a lot of your science in relation to artificial intelligence, that really gave birth to the current interest in large language models. You've got this great combined intelligence of Google Brain and DeepMind, these two amazing powerhouses of artificial intelligence. How much is Google shifting into that world of AI versus the core business, which is organizing the world's information, providing it in very user friendly formats. In your day to day life, are you talking much more about that sort of new interface with artificial intelligence? 

KB  

Absolutely, yes. And you know, Google has been an AI-first company since 2017, so now for a very long time. And we see that information and innovation mission, as you know, both entirely core to the whole business, but certainly to also how we think about sustainability. So much of what we've been talking about today is bringing together AI capacity, AI modeling in the background for something like fuel efficient routing, and then bringing that into a classic Google tool, like Google Maps and all the things, frankly, that I think that is going to enable too, from a sustainability perspective. I was experimenting the other day, as I love to cook, and I want to cook for my four year old, I was like help me make a sustainable meal plan. And Gemini  generated for me this incredible sustainable meal plan for five days, and then I was able to ask for my shopping list. And so all these AI capabilities are so helpful to us in our daily lives. But for those of us, like me, who really want to live more sustainably, they can make that much easier and much more accessible. So I think we're seeing a convergence of all of it, but certainly AI is very much at the core of what I'm thinking about. And I have so much conviction that I think AI is such a powerful climate solution. And you may have seen this, we partnered with BCG at the end of 2023 — speaking of COP — and put out, at the Dubai COP, this piece of research that pointed to the potential of AI to accelerate existing technology solutions such that we can see a reduction, in an additional 5, to 10% of emissions by 2030. Talking about this decisive decade — what can we do right now? I think AI presents a tremendous opportunity as a solution. And then, of course, we also need to be very thoughtful about its footprint and how we manage that footprint. But I truly believe being very close to this every day, that the benefits will far, far outweigh what that footprint is.  

BW  

And just thinking about that footprint, because it's both a footprint of demand, right, more energy demand at a time when we still don't quite have enough clean electrons, but we're going to get there with some of the advanced technologies you talked about. But there's also a kind of brain footprint, which is the ability to falsify or create a whole narrative or dominate... Like in thinking about AI, there's the information flows, there's the optimization, and then there's the prediction you talk about, but there's also persuasion, right? And I think at the point we're at right now, AI's ability to generate content at scale and distribute it in a very personalized way to consumers means it's a giant persuasion engine, and I feel like we might have lost control of that persuasion engine. 

KB  

Yeah. I mean, of course, you know, this far extends beyond my lane and the work that I do here, but certainly for the company, this is tremendously important in the grounding of AI models, and all the substantiation that you can see when you have AI generated information, watermarking of images. So this is something the company is really deeply focused on. I think when it comes to my work, I get to have a deep focus on these environmental applications, on the environmental footprint of AI. So not my area of expertise, but absolutely, it's a tremendous area of focus for the company, as we think about what it looks like to both boldly and responsibly manage AI.

BW  

And ultimately, we do know that most of the technologies we need to deploy are known, right? It's now just a question of the speed with which we...

KB  

Get accelerations, and deployment.

BW  

Exactly, and the political moment that we're in. I think, actually, because we're entering into this need for abundance and the recognition of the need for geo-stability and energy security actually could be interpreted as the moment we need to get us over this hump.  That we get those tools deployed, we get the range of technologies that we need out there in the real world. There's always opportunity in a crisis, right? So if we're entering into a bit of geo-instability, maybe that gives us the kind of drive to bring these to market. 

KB  

Indeed. I think as we've been discussing, the opportunity to accelerate deployment of these clean firm technologies. And then furthermore, we were talking about something like Tapestry, the role that AI can play, and actually helping with grid modeling, helping with deployment. We've seen experimentation with how to use AI to speed up permitting applications. I think there are so many ways that AI can also help as a solution, as we are in this moment where we do need to get more electricity off to the grids. How do we really do that in the most effective way possible?

BW  

Well listen Kate, it's been a delight, and I know we're running short of time, but I did want to ask you one personal question, going right the way back. In the times when you're really confronting climate science, and I personally can get a little bit overwhelmed, how do you centre yourself? What are your strategies for keeping optimistic and hopeful?

KB  

I think about that all the time, and I talk to my team a lot about how we need to care for ourselves. We need to support one another, because it can be really overwhelming. But I think it is so important to stay grounded in the problem and then really focus on the solution. And so for me personally, I'm so grateful — I was sharing this with you before we started the podcast — that I live in a state park in Mount Tam in Marin in Northern California. And every morning I possibly can, I go for a hike, and I have a particular tree I go sit with and I meditate. And for me, that's so critically important to ground me, to stay connected to the planet, which is such a huge part for me of why I do this work. And of course, the future of this planet, for my four year old daughter, for all of our children. So I really do think that personal sustainability and that sort of caring for ourselves, supporting each other, is absolutely essential as we do this really critical work together. 

BW  

And it's interesting to talk about why nature ends up being such a sustaining thing. I'm always blown away by how little I properly understand about the natural processes like the biosphere that's around us,  that keeps the oxygen that we breathe flowing, whether it's in the oceans or from plants, we still really don't understand how this all works. And I think it's good to keep a sense of awe and respect for the fact that we actually don't know everything right? And perhaps we're entering into an era where we have a greater ability to understand but at the same time, doing that respectfully, because the system we're in is so complex.

KB  

I couldn't agree more. And something I've become really personally fascinated by is also like, we discuss how technology has, in some ways, brought us inside more, away from each other, having more social interaction. There are, I think, starting to be signs of ways that we can flip that on its head. For instance, when I go hiking, I really like to understand the local flora and fauna? And there's tools, like iNaturalist, right, which you can take a photo and use AI to understand what is this plant. Or Merlin, right, a great app, where you can identify birdsong through an application using AI. So I'm very excited about the potential of technology helping us become more connected to the planet and even becoming more connected to indigenous wisdom. I've increasingly been hearing much more about projects where we're utilizing AI to create better libraries and collections of indigenous wisdom, better land management practices, agricultural practices that have been lost. But there are still people who understand how to do this, how to understand how to live more in balance with the land. So I'm excited about that. 

BW  

Well, I think that's lovely, and it's a form of environmentalism that I call Lovelockian, like James Lovelock. He was both a kind of tinkerer of technology and loved science and everything to do with the advancement of science, but ultimately, he was a preacher of Gaia. He felt connected to something, very kind of interconnected and biospheric. And so you can do both. You can be a techno optimist, but also really respectful of nature. And I think that's the kind of environmentalism I seek out.

KB  

I couldn't agree more, and that's why I go and sit with my tree every morning, to sit with mom Gaia, and to really feel that connection. And that's what fuels me to get up every day and do this work. 

BW  

And then spend your days talking about artificial intelligence. I can't think of a better thing to do every day. So thank you so much for sharing this time with us, really a pleasure, Kate. 

KB  

Absolutely. Thank you so much.

BW  

So that was Kate Brandt, Chief Sustainability Officer at Google. For me, Google represents a brilliant example of the power of collective human intelligence to create and deploy new tools. And, as Google Brain and DeepMind continue to develop artificial intelligence, the rate at which these tools emerge seems certain to increase. As Michael pointed out in our January Audio Blog on AI and energy demand, the increase in emissions from energy that AI drives could and should be offset if it increases efficiency and our ability to problem solve. And that growth in electricity demand could be just the thing we need to bring a range of clean technologies to market faster than they otherwise would — only time will tell. My thanks as ever to Oscar Boyd, our producer, Jamie Oliver, our editor and the team that make the Cleaning Up podcast and Leadership Circle possible. Thanks to you too for listening. We hope you enjoyed it, and please join us at the same time next week for a very special final episode of season 14 of Cleaning Up.

ML  

Cleaning Up is brought to you by members of our new Leadership Circle: Actis, Alcazar Energy, Davidson Kempner, EcoPragma Capital, EDP Portugal, Eurelectic, the Gilardini Foundation, KKR, National Grid, Octopus Energy, Quadrature Climate Foundation, SDCL and Wärtsilä. For more information on the Leadership Circle, please visit cleaningup.live, that’s cleaningup.live.