Maud Texier, Google’s Head of Clean Energy, in Conversation with Hugh Lawson, Managing Director at HPS

Colbert Cannon:  Hello and welcome back to Season 9 of the HPSCast. I’m your host, Colbert Cannon. If you’re new to the pod, HPS is a global investment firm that manages just over a hundred billion dollars in assets for a broad range of institutional and individual investors. That capital is invested across private credit and public credit strategies.

This week, we are going to go a little off format, as we have a special episode focusing on global energy transition efforts. The guest this week is Maud Texier, the Global Director of Clean Energy and Decarbonization Solutions at Google, who recently spoke at our annual investor symposium. We enjoyed her perspectives so much that we thought it deserved a broader audience. Maud is a thought leader globally in this area, having previously been the Head of Industrial Energy Products at Tesla and before that, was part of working on innovative technology for Electricite de France.

I will step away, as leading our conversation with Maud is HPS Managing Director Hugh Lawson, one of our firm’s leaders in clean energy investing. So now, without any further ado, please enjoy this fascinating conversation between Hugh and Maud.

Hugh Lawson: Maud, welcome. Thank you so much for coming all the way from Paris to join us today.

Maud Texier: Thank you for having me.

Hugh Lawson: And we’re excited to have the conversation about how bold leaps in energy transition are being unlocked. And you’re a great person to have this conversation with, as you’ve been a leader in energy innovation throughout your career. And your current responsibility is quite significant. Google has set an ambitious goal to run exclusively on carbon free energy by 2030, just in seven years. So, we’d like to explore today how you plan on doing that. And what’s interesting, I think, is the broader trends around that, that provide interesting investment opportunities.

But before we start, let me set a little context. And it’s interesting, last year McKinsey published a very long report – it’s about 200 pages – and they said that the transition is going to have a few attributes.

First of all, it needs to be universal, so it touches every sector. It needs to be significant in that something like $6.5 trillion dollars per annum will need to be invested in renewables and the infrastructure around it. And it’s rich in opportunity, of course, because there will be processes and products that are decarbonized along the way. Of the $6.5 trillion, which is an annual number again, only a fraction of it will come from public sources.

The IRA [Inflation Reduction Act] in the United States and other legislation is slated at something like a $1.3 trillion subsidy over a decade. So, we’ll call that $130 billion a year, which is significant, but it’s a drop in the bucket compared to the U.S. share of the $6.5 trillion. And by the way, that’s on top of the $2 trillion that will continue to need to be invested in carbon-based energy during the transition. So, these are really big numbers.

So, what’s interesting is the private sector is going to be the source of the bulk of the capital. And in this, Google is a leader, and I want to start with the pledge. So, the pledge is carbon-free energy by 2030, and you call it, for short, the 24-7 program. It’s more ambitious than the 2050 net zero target in the Paris Agreement. So, you probably could have set it for 2050, but you did 2030, and it’s more notable, given that the world’s demand for energy is projected to rise at least two or three times – some have even said five times. And Google, itself, uses something like three times more electricity than major cities, like San Francisco, per year.

And with AI [Artificial Intelligence], I can only imagine how much electricity you’ll need by 2030. So, Maud, could you share a little bit the scale of the 24-7 program and kind of your starting point with that?

Maud Texier: Yeah, absolutely. And maybe before digging into the program itself, let me start with a bit of context and history about where Google is coming from.

As you mentioned, we are a large energy user. If you look at our historical load growth, we are increasing our electricity consumption around 10 to 20% year-over-year. So, this has been a steep curve for us. And so, that’s why, almost two decades ago, when Google started thinking about its sustainability and carbon footprint, very quickly we zeroed in on the electricity consumption of the data centers, because we realized this was the backbone of the internet, and that was one of the core sectors where our carbon footprint was coming from. So, this is what started our journey, around 2010, around clean energy and how do we decarbonize the grids where we are operating.

And we really started in 2011 and 2012 looking for clean electricity that could power our data centers, but we realized quickly it was not that easy and actually was quite complex for an energy buyer like us to access clean electricity. Let me explain to you a little bit how grids work if you’re not familiar with that.

You can imagine a grid as a large pool where you have all the electrons. And so, you have a power plant generating electrons, and on the other side, you have Google, as a buyer, getting electrons. You cannot really trace those electrons, where they’re coming from and where they’re going. So now, Google as a user, we’re looking at this pool, and we’re looking for the clean electrons, and we realized it was impossible for us to do that.

And so, we went to the other side, and we found a wind farm in the Midwest of the U. S. This was our first clean energy project, and we signed a long-term contract with this wind farm – something that you call a PPA, Power Purchase Agreement. What PPAs are is, generally, you have a corporate buyer like Google, who is going to come and sign an offtake contract for all the generation of this wind farm. And this is going to be a long-tenured contract with a fixed price. So, it’s something very, very easy for the wind farm to take on, and something that a corporate buyer can de-risk from our perspective.

So, we went on with this first project in 2010 and 2011. And then, onwards, we started working into growing this portfolio with the objectives to gather enough clean electricity globally to match our annual electricity consumption of all our data centers across the world. And this is how we ended up today with a very large renewable energy portfolio under contract management. So, just to give you a range of the magnitude, we have around 80 projects globally that we are managing, and 10 gigawatts of wind and solar across five different continents.

Hugh Lawson: Wow. So, 10 gigawatts, which is a lot, across 80 different projects. So, let me make sure I understand. So, somewhere along that journey in 2017, I believe it was, you pledged to purchase as much renewable energy as you used, but that has a fungibility problem. So, it sounds like you were purchasing a lot of renewables in markets that had a lot of renewables, but in other markets that didn’t, you couldn’t, right? But you were over-purchasing in some markets to basically subsidize.

I think what you just described – and people often talk about the decentralization of the energy ecosystem – I think that’s an example. So, you’re going to have 80 projects where you’re trying to build clean energy or catalyze the building of clean energy quite proximate to the use. So, you’re going to generate it, more or less, next door to where you’re using it. Is that what’s different?

Maud Texier: Yeah, absolutely. You mentioned it very well. 2017 was a bit of a pivot year for us, because this was the first year where we actually achieved this goal of purchasing enough clean electricity to match our energy consumption globally.

So, fast forward, we achieved this for the first time in 2017 and in 2018 and 2019. And after two or three years, we kept growing our portfolio to match our energy consumption increase, but we started really digging into this inventory and trying to understand: are we really achieving, on one side, the decarbonization impact that we were looking for in the first place? And second, as our portfolio keeps growing, what does that mean from a long-term financial management perspective?

One thing we realized in 2017 when we did this inventory is that actually, like you mentioned, our portfolio was a little bit imbalanced. It was imbalanced from two perspectives. First, we had a lot of renewable energy projects in specific regions. In the U.S., those are the regions where you have a lot of renewable energy – Texas and the Midwest is an example.

In Europe, Scandinavia has a lot of wind as well, so historically we had wind farms. While we were purchasing those clean energy projects, our load was growing in markets where clean energy was not that available. And so, we ended up with this regional imbalance where, in certain grids, we were short from a clean energy supply perspective, and in other grids, we were long.

So, that was creating two issues for us. One, we couldn’t really claim that actually our data centers were running on clean electricity where we were running short. And two, this was also creating more of a financial risk exposure for us because now we have those long and short positions across different markets.

Hugh Lawson: Well, that’s interesting. So, I want to come back to the, sort of, portfolio effect, but let’s spend a minute on technology. So clearly, solar, wind, and to some extent, battery storage have been around for a while. There are new technologies that are on the horizon. In order for you to reach your goal by 2030, how many of these new technologies need to be scaled? Is it new things or the old things that will get you there?

Maud Texier: Yeah, great question. We just talked about, you know, the portfolio, and I would say the viability of this carbon-free energy content across our portfolio before. As we set out on this 24-7 objective for 2030, we started looking at all of our grids where we are operating.

And just to give you a sense, we’re operating across 24 energy markets or power grids across the world. And we realized that those grids were very different. In some grids, we had a lot of wind and solar but that also created some hours where we had a shortage of clean electricity because the wind doesn’t always blow, the sun doesn’t always shine.

And then we had other regions where we realized that those resources were just very limited in the first place. An example that you might have seen in our portfolio is Taiwan. We have a data center in Taiwan, but this is an islanded grid with very little real estate available on the island for clean energy.

So, for instance, solar or onshore wind is very limited. So, as we went through this process, this is how we realized that we’re going to need to expand a little bit the tools in our toolkits for clean energy and clean electricity procurement. And we started looking for: what are those new technologies that would be complementary to wind and solar and really help us increase this ratio between our consumption and our production on a real-time basis?

And so, for that, we kind of like went back a little bit to first principles of the grid. So similarly, if you’re not familiar with how we generate electricity on a grid, you have broadly two categories of power plants. One is a power plant that we call baseload – or firm generation resources – and you can imagine a coal power plant, for instance, in most of the grids today. It’s always on, it’s very cheap to run, and so therefore, it’s a continuous flow of electrons. However, it’s not very flexible.

And so, on top of that, you need another type of power plant that you call peaker plants. And those peaker plants run only a few hours a year. They’re much more expensive to run, but they’re very reactive.

So, on the very hot days, for instance, in the U. S. or even in Europe, right now, everybody’s running their air conditioning. You need those bursts of electrons that normally you do not use. This is when you use peaker plants. And traditionally, those have been gas turbines, for instance.

And so, what we’re looking for now is really clean alternatives to those traditional baseload and peaker plants. And that’s why we’re looking for clean, firm generation resources on one side and clean peaker plants on the other side. We have narrowed down our search to mainly five technologies.

Just to be clear, we’re not necessarily looking at picking technology winners. This is not our Google strategy, but we narrowed down the sources based on the requirements and the specifications that we had from our own goals and portfolio.

So, just to give you some specifics, our program is due by 2030. So, we need those technologies to be ready and at scale by 2030. We also looked for technologies that had a high potential of cost curve decline based on rate of deployment. So, the more you can deploy those power plants by 2030, the more the cost is going down. And then finally, we also took into account the fact that, when you build a new power plant as an asset, there’s still a lot of, I would say, criteria around social community integration and also environmental criteria. And again, we believe that those five technologies can bear with those criteria that we have.

Hugh Lawson: Well, it’s interesting. You know, as we were preparing for this, I was struck by this whole decentralization theme. So, I want to unpack that a little bit, because I think it’s important. So, it’s not just simply a case where in regions where renewables are easy, you necessarily want as much of it as possible, because of course, you’re not matching necessarily the loads in those regions with the generation, right? You might have too much, and then you have the variability problem.

And so, what’s interesting is, when you think about it, because you’re using more and more electricity all over the world, and you also want computing times to be quick, for people to do Google searches and so forth – so the data center needs to be responsive and close to the market – you fundamentally need to restructure the energy ecosystem basically as a business matter. Because it sounds like you’re reducing volatility around both the price and availability of electricity, you’re heightening security, and of course, you’re decarbonizing. Is that a fair way to characterize what you’re trying to do?

Maud Texier: Yeah, absolutely. I think for us, one of the key enablers on why Google took that target in the first place is it was really going beyond the traditional corporate ESG programs and perspective.

Today, you know, corporate energy procurements are mainly to drive carbon metrics, decarbonization, or greenhouse gas emission reductions. Our approach to 24-7, we realized very quickly, created a lot of business values from different perspectives.

One, as you mentioned, around energy price volatility and hedging, and we can talk more about that. But really having a clean energy portfolio with a fixed price on one side that is completely hedging your consumption on the same grid on the other side, especially in the current environment. What we saw last winter in Europe, for instance, brought a lot of value for us.

And then separately, to your point as well, with AI, and even before that with our cloud business, we had massive growth, but a growth that was a little bit different from what we had historically, because historically for search and YouTube, you just want a big data center in one place and run all your queries. With cloud and all that new usage, we need to be much closer to the customers. And that’s why we have decentralization – a little bit of our data center portfolio – and from there, we really need to tackle new markets and new countries. And with that, there comes new needs for us.

For instance, when we build new data centers, we need to make sure that we have net new electrons coming from that grid for our data centers. So, this is not even about how clean those electrons are. It’s that we just need those electrons to appear. And also, we are noticing, in certain regions, what I would call the more and more pressure around social license to operate for us and other industries.

As we want to bring more data centers and factories at the local side, we are required to bring a social and climate plan, and making sure that we are fully integrated in the local infrastructure. And this is where 24-7 really carries, as a story and an impact, because again, this is a local program for us that we are running.

Hugh Lawson: So I think it’s interesting, you know, in terms of the capital that needs to move. You want electricity where you need it, you might as well do it in a renewable format because that’s actually cost effective and more efficient, and it’s core to the business, which I think is really interesting.

So, maybe let’s turn to some of those projects a little bit and how you source them.  A lot of what you describe, in sort of a decentralized way, is creating investment opportunities because what it doesn’t sound like it is, is these giant utility systems expanding what they’re doing. What you’re talking about are projects all over.

So who are some of the actors? Are these developers, are these private sector people outside of the utility? Who’s involved, and what sort of investment opportunities in general does this create?

Maud Texier: Great question. And, before getting there, I should flag that, again, we are operating on roughly 24 grids today, and you can imagine that you have 24 very different markets to deal with. Every grid is slightly different, and that’s one of the key complexities, I would say, that for the energy industry to deploy globally. The local regulations and market mechanisms are always very different.

And so for us, what this means in terms of portfolio engagement is, we’re going to work with very different stakeholders across the globe. In certain markets, we’re going to be able to work directly with the local utility, develop with them tariffs, build together power plants, and really have a tighter engagement. In some other markets, we’re going to go work with independent power producers or renewable energy developers. And this is a very fragmented market, in general, but we’ve seen a lot of growth, especially in the diversity of, offering on that side in the U.S. and in Europe more recently.

Hugh Lawson: Right. Is the value chain too disaggregated to get there efficiently and quickly? Because you’re going to have to build a lot of these things at once. Can you talk to us about that?

Maud Texier: So, this is definitely not a quick process. Just to give you an order of magnitude, from the time when we sign a contract with a renewable energy developer to the time where we receive the electrons for the first time, there’s at least a three-year lag. And so, for instance, what the team is working on today is already looking at all we need for 2026 and 2027. And this lag is partially because this is infrastructure, and it takes time to build things, but also because this is a very linear and sequential process.

And just to give you a couple of examples around where the pain points are across this process. Broadly you can think about developing and building a clean energy power plant as two types of activities. The first one is what I call development and deployment of the projects, and the second one is going to be about commercial and financing of the projects. Deployment and development starts more traditionally. Like real estate infrastructure, you need to find the sites, you need to lease or acquire the site.

Where the difficulty comes, especially around clean energy projects, and more recently, particularly in the U.S., is it’s very hard for developers to understand which sites are going to be the best to integrate with the local grid. So, what I mean by that is, if you build a wind farm, you’re going to connect it to the local grid. You need to make sure that grid is sized to receive your electrons. You need to make sure that you have poles and wires available to host those electrons. And so, there’s a whole study that has to go through with the utility partner and the grid partner. That’s why we call it interconnection, and this interconnection phase has been a big blocker in the U.S. more recently.

A very specific example is one of the large operating energy market regions in the U.S., which goes all the way from Illinois to Virginia – so really a large region – over there, the grid operator has received a massive influx of renewable energy project requests for interconnection. And on top of that, those are new types of technologies, so they’re not used to studying those technologies very quickly.

And so, with that, the grid operator got completely overwhelmed, and they had to put on pause all of the treatment of those new requests. So basically, the whole clean energy development market has been stalled and is going to be stalled for a couple of years. So, just here you see how you get a lot of flags on how the grid and the stakeholders are learning about using those clean energy power plants, and how they can really integrate them into the system. So, that’s for the more technical side.

For the commercial and financing part, there’s also many steps, right? I mentioned earlier we are signing PPAs, Power Purchase Agreements, and I have to say this has been, kind of, a major game-changer for wind and solar projects, because you have to imagine those grids when you push electrons.

There’s an energy market layer on top of that, a financial layer on how you get paid for your production and how you pay for your consumption of those electrons. So, every minute really, you need to have this balance between demand and supply on these energy markets, and from there you’re going to get a resulting price.

Wind and solar is very challenging because they are intermittent by nature, and it’s very hard to predict actually when they’re going to be producing or stop producing. So, from price volatility, this is a lot of risk. Again, as a developer, you go try to get financing out of these projects. Your cash flow is not predictable at all, so it’s very hard to get financing, and that’s where things like our new transaction instruments, like PPAs, have been very helpful, because now, you have those fixed-term contracts with much more predictable revenues. And so, PPAs have been a core mechanism to really, I would say, boost those markets over the past 10 years.

Just to give you a sense, the number of corporate PPAs that have been signed last year is 10 times bigger than what we used to sign 6 years ago, globally, across all of the corporates. So, this is really a booming industry, and it keeps rising as well. But again, while we succeeded in doing that for wind and solar, we need to start thinking about how do we do that for this next generation of technologies we talked about earlier. What is a green hydrogen PPA? How do you define a green hydrogen project? Who takes the technology risk? Because those are still new technologies, and on top of that, you have market risk. And I think this is where corporate buyers have been doing a good job in the past, and we need to do more for those new technologies, but there’s also a dialogue that can happen with financing and investors and really connecting together on what does a bankable commercial contract mean for them to be able to bank those projects for us.

Hugh Lawson: Yeah. It’s interesting. I want to pivot just to come back to government policy a little bit. So, you know, many have heard about the Inflation Reduction Act. I referred to it earlier. It’s a pretty massive subsidy to stimulate the creation of clean energy in the United States, but it also has provisions that require domestic contents. So, a lot of the components have to be manufactured in the United States, and we’re actually, by the way, seeing a tremendous amount of opportunity of people wanting to build factories and create jobs and generate these parts. But how important has the IRA been to Google’s procurement?

Maud Texier: The IRA has been a big accelerator for us, especially as we think about this next generation of the type of assets and technologies. When we started the program in 2019 and 2020, and we were asking around – do you have green hydrogen or long duration energy batteries? It was very hard for us to get any tangible projects, and people were very worried around economics, asking is it really going to be a strong market signals? We were a little bit lonely, I would say, looking for projects. With the IRA, this has been a massive game-changer because there’s been such stimulus around those new technologies, and there’s been a big push in also creating hubs

So, for instance, in the U.S., you might hear about carbon capture hubs or green hydrogen hubs and really helping with those economies of scale. So today, for us, it helped a lot to find the supply and the projects that we need to find by 2030. I would say, though, even though the infrastructure investment piece of that is present today, one of the key pieces of feedback that we’re hearing from a lot of those projects is they still need the commercial of takers on the other side.

And one challenge that we have, as Google, for instance, is even though we’re big, we’re still not going to take off a one gigawatt green hydrogen power plants by itself. So, this is where you need to create this market demand and market signal, and how can you aggregate, how can you partner with other industrials, for instance, to create this market signal.

Hugh Lawson: Interesting. Okay. Well, let me pivot to AI for a second. It’s on everybody’s mind. The implications are pretty extraordinary. AI computations apparently are doubling every six months and training these models uses a growing share of electricity. It’s an extraordinary amount. You know, as it gets applied to more and more real-world situations, how much electricity do you think you’re going to need and is the world going to need to adapt to an AI world? Just give us a sense of the quantum.

Maud Texier: Yeah, this is an answer I would love to have. I think we’re just at an early stage of the cycle for AI. On one side, there’s a lot of discussions, there’s a lot of curiosity and activity, so we do see a spike in usage, that’s for sure. And, you know, long-term, we do see demand growth stemming from AI, but on the other side, we also start seeing much more efficiency on more traditional compute and services that we’ve been providing to cloud customers, for instance. You have AI, and you do not need so much traditional compute or storage to run the same queries. So there’s a little bit of a cannibalization, almost, of the usage on a certain side. So, that’s why it’s hard for us to really know how, you know, how steep the curve is going to be. The curve is going up, but again, if you look historically, our curve has been going up for the past 10 years, at a rate of like 15-20 percent year-over-year.

So, this is not necessarily a new challenge for us, but I think it is an interesting one, and, the one thing I would go even a little bit further with is, AI is stemming a lot of focus around electricity consumption of heavy electricity users. And I think that’s a very important dialogue because data centers are only one out of many industrial players who are going to face something like an AI growth over the next decade or two decades.

We are talking to a lot of more traditional heavy, industrial players like chemicals, and they all have net zero targets. They all need to decarbonize. And when you look at the very first step on how do you decarbonize a business like that, it’s all about electrifying your processes and manufacturing process. And so, right here, this is also a lot of net new electrons that they’re going to need.

On top of that, we talked about them already, but we see this new wave of climate technologies that are going to come in for the energy transition or just decarbonization of the sectors. All of those climate technologies, they need to be manufactured, and so, you see a lot of factories being built, bigger factories for batteries or other types of components.

We talked about green hydrogen, you need electrolyzers, big machines running out of clean electricity to actually produce the hydrogen. So, this is really the bigger problem, which is there’s a grid today and there’s a lot of focus on decarbonizing that grid, but from my perspective, it’s really more around what is the grid in the next decade or so that we need to build, and how do we make sure that it is clean and cost efficient and reliable in the first place. This is about the next power plants that we need to bring to the grids over the next 10 and 15 years, and this is where we have the real challenge in terms of scale and pace of deployment.

Hugh Lawson: Interesting. Well, I do want to pick up one thing you said on it. So, so AI uses electricity, but it also helps people save it?

Maud Texier: Correct.

Hugh Lawson: Its inefficiencies? How do you apply AI now, for example, to use less electricity?

Maud Texier: Yes. Great question. So, on one side, AI is going to be a vector for us to focus to make sure that we keep decarbonizing operations despite the business growth. On the other side, it is an opportunity to develop new solutions for climate change.

So, just to give you a couple of examples, again, in the context of energy and power markets and grids and how you can use AI. Personally, on our side, we’ve been using AI for our own operations. In the past, we’ve tested using AI to optimize how we are managing data centers and energy and some of our management systems. And we created a lot of energy efficiency within our data centers. And keep in mind, energy efficiency is very interesting, because on one side, it helps you decarbonize, but on the other side, it helps you save money as a business. And so, those are really strong vectors for us.

The second item I would mention is a little bit more broadly, which is, we talked about clean energy, we talked about how we need to better match our consumption and production in real time. This is really less about AI but more broadly around digitalization of the grids. And as much as we need to deploy and build that new infrastructure, if this infrastructure doesn’t have real-time controls or real-time data reporting, we’re not going to be smart about utilizing those infrastructures. We need to be very efficient about utilizing those infrastructures. And that’s where modernizing the grids around data – which data are available, and how do you use those data in terms of predictions.

We talked a lot about wind and solar and how intermittent they are. Can you predict, can you forecast when wind is going to blow, and therefore, when you do not need other power plants versus when the wind is going to stop? And you need a battery to help you procure the missing electrons during that time.

So, this is where AI is really entering, I would say, a real opportunity for us. But you still need this data, this data layer in the first place. You need all those data to be able to run AI.

Hugh Lawson: One of the other things that Google has been very vocal about is that your objective is not only to decarbonize yourself, but it’s to act as a catalyst to, sort of, help others do the same thing. Can you give us an example or two of how you’re collaborating with others?

Maud Texier: Yeah, absolutely. We have a couple of pathways for this catalytic role. One, we’re trying to de-risk technologies. By having Google signing a project with a new technology, we believe that it helps create a lot of credibility to those technologies. And also, we’re trying to be very transparent about the results there as well. So, being a little bit of a technology risk-taker, for us, is one way to be catalytic.

An example is, we actually have a project in Nevada – an advanced geothermal project – that is about to be commissioned and starting operations pretty soon. But back then, it was a few years ago, when we signed these projects about advanced geothermal, it was very early on for this industry, and today there’s much more activity in the U.S. around geothermal, so just as an example.

A second example is that we also realized that, very selfishly, if we’re the only ones asking for those type of products in the market, it’s going to be a high premium because it’s going to be custom. So actually, the more people we have embarking on that journey with us, the easier it’s going to be for us. So, it’s kind of this virtuous cycle for us.

So, we have been working a lot with partners across the energy value chain. You can think about utilities, energy suppliers, and we’ve been working a lot with them to create new contract structures and commercialization pathways to reduce the complexity for the buyers to be able to replicate the same thing.

Hugh Lawson: Well, we could go on and on and on, and I’ve, I’ve learned so much from spending a little time together, Maud. Thank you.

Colbert Cannon:  Maud Texier, Global Director of Clean Energy and Decarbonization Solutions at Google, Hugh Lawson, a Managing Director with us at HPS, thanks to both of you for that really insightful conversation. And thanks to all of you for listening.