by Stephen Lacey
March 29, 2018

Stephen Lacey: This is The Interchange, conversations on the future of energy from Greentech Media. I'm Stephen Lacey. Welcome to the show. This week, how Google, Microsoft and other corporate giants are shaping energy markets.

Stephen Lacey: We've got a double header for you. Firstly, I converse with Google's Head of Energy Strategy Neha Palmer about hitting a 100 percent renewable energy. I wanted to know exactly what that means. Then, Shayle Kann talks with Brian Janous, Microsoft General Manager of Energy. Those two will go deep on how Microsoft structures deals. In both conversations you'll hear us talk about how corporates are thinking about matching renewables directly with demand in real time rather than just a yearly average. It's not easy but it's where the market is headed for the most sophisticated companies. Consequently, you're going to hear about how Microsoft and Google are thinking about storage which is crucial for those plans.

It gets us into this bigger question that corporate buyers are facing. What happens to markets when the gigawatts and gigawatts of renewables they're buying literally reshape how markets function? You're going to want to listen to both these conversations, they compliment each other really well so enjoy. Let's turn first to Google's Neha Palmer. After hitting a 100 percent renewables last year through a combination of direct purchases and renewable energy credits, she's preparing her team for a new phase of buying, matching supply granularly with demand. I first asked what Google means when it claims to be a 100 percent renewably powered.

Neha Palmer: That's a great question, 100 percent renewable means different things to different people. For Google, it means that for every kilowatt hour of electricity we use across the globe last year we bought an equivalent kilowatt hour of physical renewable energy. Again, this is a global and annual goal. Our next step is actually to make this a more regional goal to match our consumption with renewable purchases on every grid that we operate on and then also eventually to match it with our consumption. Every hour the energy production matches our consumption but for right now it means that in 2017 we bought a physical kilowatt hour of renewable energy for every physical kilowatt hour that we consumed.

Stephen Lacey: Matching supply to your actual local demand is incredibly complicated and there's a reason why companies are only just starting on this process. How do you work toward that? What are the steps that you have to take now that you've hit this net renewable energy goal? How do you actually balance it with local demand?

Neha Palmer: It's definitely a much more longer term goal than the renewable energy goal. To give you context, we started with our renewable energy purchases back in 2010 and it took us until 2017 to achieve that matching goal. We believe that this next phase will take a bit longer and has three pillars to it. The first pillar really is technology. Things like storage, battery storage, all of these will have to come along in capability and cost to help us get there. Also, new forms of renewable energy generation, things we may not even conceived yet will probably be required to get us there. The second is policy. The market structure can often dictate what we can do in terms of renewable energy purchases. You just see markets evolve to allow more consumer choice and consumers to be able to direct their energy purchases more.

The last piece is maybe a little bit easier and falls more into my camp here which is transactions. Are there ways to do energy transactions that match our consumption better? A simple thing might be just matching a solar and wind purchase together so you have more around the clock renewable energy, right? Wind tends to blow at night in some places and obviously the sun shines during the day. Those are the three main pillars but it will take longer than the seven year trajectory where we had to get to the 100 percent renewable.

Stephen Lacey: The big question is how much more diverse this procurement is going to get among big tech companies, big corporate customers, and people are just starting to take a look at storage. They're considering other sets of technologies but still mostly wind and solar. At what point do you think that will start to change?

Neha Palmer: I think storage is probably the next big thing to start to disrupt the market. As far as technologies go, I always say I'm technology agnostic but not cost agnostic. I think that that would be similar for most corporations, right? We have to run a business and do this in a cost-effective manner. Wind and solar are really the two that are setting the market prices right now. The storage, it really feels like it was back in maybe 2008, 9, 10 for solar. At that time, you really saw the only people buying large amounts of solar were utilities trying to meet their renewable portfolio standards. They were mandated to do so. You did see that that pushed the market to a level of uptake where the price was able to come down. We're seeing that with solar now.

It remains to be seen how fast that cost trajectory comes down and how that will actually interplay with the markets. Right now, I think they're still trying to feel out where is the most value for storage. Is it offsetting the curves? Is it ancillary services, regulation? I think that's still being vetted in the market but as the cost comes down, we'll have the ability to play in many more locations and many more parts of the market. We're hopeful that that will happen here soon and open up a world of different possibilities for different technologies beyond wind and solar as well.

Stephen Lacey: When you went to Google in 2012, you're at PG&E before that so you had seen the cost trajectory of wind and solar come down. Were you surprised at all about how much cheaper these resources got in a pretty rapid period of time?

Neha Palmer: I have to say it has been pretty astounding to just see the rapid decline. I think solar they said over the last seven years, it's over 80 percent and wind is 60 percent. That is pretty astounding for any industry to see price declines like that. Definitely surprised at the speed but when you look at the amount of uptake just between utilities and even corporations. I believe corporates have bought over eight gigawatts of renewable energy to date. You could see that the demand is there and suppliers are able to add more manufacturing lines and every time they do something like that, it just drives the market price down.

Stephen Lacey: What kind of internal resources can you use from Google to make data centers more efficient? For example, to make sure that you don't have to procure as much renewable energy or monitoring power pricing or what have you. Are there other internal tools within Google that help you in your job in achieving this ambitious renewable energy goal?

Neha Palmer: Yes, machine learning is something that we are developing rapidly at Google across many different platforms. Cloud is one area. We have machine learning offerings and AI offerings on that platform but we also leverage them internally. A huge emphasis of our data center program is actually what we call the negawatt, right? The megawatt hour, kilowatt hour that we actually never use. For us, this is the main way to reduce our energy consumption and it started way back when when we would separate what we call the hot aisle and the cold aisle of the data center. Now it's evolved to things like machine learning.

We were actually able to run some pilots within our data centers where we implemented machine learning to optimize our cooling systems and we were able to realize a 40 percent reduction in energy for cooling when we were utilizing this machine learning. These are tools that we're using across our businesses. I am also investigating different ways to use this machine learning AI technology to see if we can optimize around our energy procurement and pricing.

Stephen Lacey: Have you found any successful ways of using that data yet? Are you still in the beginning phases of that exploration?

Neha Palmer: We're definitely in the beginning phases. I think energy industry is very interesting and that there's so much data created. You have meter data, you have SCADA data, you have pricing points, you have hub settlements. You have all of these different data points. A big portion of ML and AI is getting the data into a format that's understandable. We are working on that. We're working with various partners within the company and trying to figure out how we can optimize around that. It remains to be seen how it will be used but I definitely think that this is something that we'll be using in the future.

Stephen Lacey: How do you see Google as an energy producer in the future, a participant in the energy markets? Not just signing contracts for renewable energy but as you outlined it using on site resources to maybe bid into wholesale markets to provide ancillary services to the grid, to not only serve your facilities but to be an actor, a beneficial actor on the grid. It seems to me that the next step in integrating a lot of these other resources beyond just wind and solar makes you into, for lack of a better phrase, the prosumer, this integrated energy provider, an independent power producer so to speak. Tell me about the role that you see Google playing as your operations get a lot more complicated and the technologies get more sophisticated?

Neha Palmer: I think it remains to be seen, the technology isn't quite there yet. The one advantage we have and many other companies do have as they engage in technology is again, this data. We are pretty expert at finding data streams, figuring out how to leverage that and then tying it to operations with our data centers. I don't think we have a strong vision of what that is right now but to the extent that we're able to meet our goals which is cost-effective, low carbon energy, we'll definitely engage with the markets to do that. We already have this large portfolio of renewable energy projects over three gigawatts of projects that we've engaged with and we do engage with the grid with those projects. We're actually scheduling the power from those projects onto the grid bidding into the market, bidding some ancillary services from those projects onto the market. We're already stepping the direction of doing that. New technology will just allow us new ways to engage with those markets.

Stephen Lacey: One of the dangers of more closely matching supply with your localized demand is that you could potentially over-engineer the problem. You procure more resources than you actually need. What are some of the issues that you need to think about as you attempt to serve your local demands with renewables?

Neha Palmer: Well, I think one thing that we're lucky in some sense is that our data center loads are pretty predictable. They're quite flat. In fact, from the utility perspective, being an ex-utility person I think of that as the ideal load, right? It's very flat, not to peaky and pretty constant throughout the year. From that perspective, that's a benefit in trying to solve this equation. I think the mix of resources that we'll have still remains to be seen. I think there's a lot of research going on. I think that there is a lot of assumptions around the different cost trajectories and capabilities of technologies that we're going to have to decode over time. I think we'll continue to focus on making sure that we're optimizing around cost. That will obviously be a key goal but I still think that the market is still vetting itself in how we're going to be meeting this goal.

Stephen Lacey: You said you're tech neutral but you're not cost-neutral. Let's assume that renewables hadn't dropped in cost as quickly as they have, do you think you would have achieved this a 100 percent renewable energy target in 2017?

Neha Palmer: Gosh, it's really hard to guess what would have happened.

Stephen Lacey: Counterfactual.

Neha Palmer: Counterfactual judgment here but I think it's interesting. There's so many different forces in the energy markets, right? We haven't talked about natural gas but that's been a huge driver in what's happening in the market as well. The [inaudible 00:15:24] fracking and this huge flood of cheap natural gas has really changed the energy markets. What you're seeing now is that it's a very energy-only focused market so you're seeing renewables to actually set that price of power in the market. If you're looking at things like premiums, who knows what it would have looked like? You might have just seen the price set higher but it would still be at the price of the renewables. Again, the counterfactual is hard to prove and hard to analyze but it's interesting to think about.

Stephen Lacey: 100 percent renewable energy has become a bit of a triggering word in energy circles. There's this fierce debate over whether we need a 100 percent renewable energy, whether we could power the world with wind, water, and solar as professor Mark Jacobson claims or whether we need this whole other set of technologies to decarbonize the grid and whether a 100 percent renewable energy should be the goal or a decarbonized grid which opens you up to a bunch of other potential technologies and the most notable being nuclear and carbon capturing storage. As Google goes forward, is 100 percent renewable energy going to be the goal or is decarbonization and flexibility equally as important?

Neha Palmer: Again, as we move towards that contemporaneous matching of our load, we're going to probably have to consider other technologies. We're definitely open to that. The debate that is out there is internal as well. We have the ultimate goal, it's always been to reduce our carbon footprint. With electric consumption being the largest portion and that's where we can pull that level. For us, we will look at whatever gets us along the path to decarbonization the fastest.

Stephen Lacey: I'm curious to learn a little bit more about how that plays into your internal conversation around energy. I've been following Google's approach to energy for a long time from way back when it first invested in a rooftop solar system at the Mountain View campus in 2006 and then a year or two later when it launched renewable energy is less than coal, there's wide-ranging investment initiative in which it poured money into enhanced geothermal, geothermal drilling techniques, concentrated solar power. Eventually, that was abandoned and Google reformulated its strategy. Years later there were a couple of engineers who worked on that RE is less than C initiative and they wrote this internal report about why focusing on renewables just wasn't enough to decarbonize the energy system.

When that went public it fed into this broader debate around the criticism of a renewables only approach versus maybe a broader decarbonization strategy. That was a really influential document. I'm just curious if that type of conversation if you're thinking about that as you're engaged in these long term procurements where you're actually having to evolve your business model and it's just getting a lot more complex for you over time.

Neha Palmer: Yeah, I think it does. A couple of considerations that I have. I am running a huge operation here. We buy a lot of renewable energy because we have a lot of electric consumption for our data centers. I have to do this today. We need to buy this energy for delivery tomorrow. I think a lot of debate does center around some of the assumptions of cost trajectories, availability of resources, a lot of different things. I think a lot of that still has to be played out. We have to see how the technology roadmaps develop. That's super exciting, as energy geek I find that very interesting. As someone who has to provide operational support for our business, I'm looking at the here and now. From that perspective, I think that's why you've seen us focus on the renewables, it's something we can do today.

It is offsetting a huge amount of carbon, what we're doing here with the renewable energy purchases. There is definitely debate on what is the best path forward. Again, the goal is to mitigate carbon and we'll continue to investigate and work towards whatever is the best path forward. I do think though that the industry needs to mature more and some of these technologies and policies and market structures have to be played out more before we can make a call on what is the best path forward. In the meantime, we're going to keep buying renewable energy because that's something that we can do now.

Stephen Lacey: Has there been any moment in your job when you've taken a step back and you've looked at pricing for example for projects and you're just like, "Whoa," just really surprised or where it's encapsulated how quickly this moment is changing?

Neha Palmer: Absolutely. We've been running RFPs for renewable energy now for the entire time I've been here and every single time we run RFP it seems like the price decreases by five bucks. It's astounding. You see things that are some $20 now. Especially considering when I started my career in this field at PG&E, the prices that we were seeing just astounding. It does make me think like what will the market look like. You have this marginal priced markets focused on the lowest marginal price. Will those markets be able to withstand super low pricing? It is astounding and it's interesting and fascinating to sit back and watch some of this unfold.

Stephen Lacey: Do you think they will? That seems to be the big question that finally people are asking, right? You're seeing 20, 30, $40 per megawatt hour contracts in wind and solar and now all of a sudden people are worried about wind and solar being a victim of their own success because they're price-takers. Can markets withstand all these cheap wind and solar, you think?

Neha Palmer: I do think they can. I'm [inaudible 00:21:40] market design expert, I think there are ways to support the market and support renewables but it is interesting to see as we see more and more come onto the grid how various services will be valued? How [inaudible 00:21:56] start plans? All of these stuff will have to be potentially rethought but it's very exciting. I think that from the customer perspective, we're getting great low cost low carbon energy so that's the way that I look at it. Maybe a period of disruption but the end result will be good.

Stephen Lacey: Neha Palmer is the Head of Energy Strategy at Google. Neha, thank you so much.

Neha Palmer: Thank you.

Stephen Lacey: Okay, let's get a bit more detailed on how these big companies are structuring deals, shall we? To get that detail, Shayle Kann talked with Microsoft's Brian Janous. In the remainder of the show, Shayle and Brian rift on risks, the details behind some big deals, and the future of electricity markets. As Brian explains, companies like Microsoft create their own weather systems on the grid which we'll at some point impact the economics of their procurement and everyone else's. Brian begins there detailing how cheap renewables create their own sets of risks.

Brian Janous: I think honestly the one that is probably least appreciated that is actually the most significant risk is shape risk. It's something that you've talked about on your show which is as we see increasing penetration of renewables there is a downward impact on wholesale market prices. If we're trying to put increasingly more renewable energy into a grid, then a lot of ways it's already fully supplied, it is going to have a downward effect on price. That I think is the biggest challenge that buyers really need to appreciate and that not all the corporate buyers really taken that into account. If you take a forward curve that just assumes that it's going to slope upward because prices always rise, I really think that fails to take into account that we continue to inject the zero variable cost energy into the market and so if you're not at least contemplating that wholesale prices in a market with 30, 40, 50 percent renewables will be lower than they are today then I don't think you're adequately assessing the risk and the cost in purchasing renewables.

Shayle Kann: In a typical traditional project structure, who bears that risk? If you put a solar project in Texas then all of a sudden a ton of solar floods that area of the market in Texas and wholesale prices wind solar is generating start to crash, who's bearing that risk? Who suffers from that?

Brian Janous: It's almost always the buyer. I mean, two different things can happen. You can have that in the context of a vertical integrated utility where they go out and they build a renewable project and they're going to guarantee cost recovery for the next 20 years on that project. Whether that project ultimately ends up being valuable, now, again, we're not talking about the carbon reduction value because there's a societal value there. We're just talking about the pure value of the electricity in the wholesale market. That will be born by the rate payers. If it's a PPA then it's going to be born by the offtaker. The offtakers really taking effectively that merchant position if you will in the marketplace. They're bearing all of the cost and risk of that project.

Shayle Kann: Right, actually, that's a good segue into you mentioned that you did a wind procurement deal in Ireland luckily last year and if I remember right, that one actually included some amount of battery storage within it which presumably is one of the things that you're looking at to try to deal with this shape risk. How much storage was incorporated in that deal? Do you see more of the procurements that you do going forward incorporating storage or some other way to be able to make those projects more dispatchable?

Brian Janous: We do. There was a relatively small amount incorporated in that particular project. It was really a pilot to test the interface with the market and how you would integrate those two together. We're starting to see more and more of that. We've seen some announcements recently in United States focused on the same thing. I really look forward over the next probably two or three years to start to really be able to increase the percentage of storage integrated into these projects because I do think that's going to be a critical piece of being able to integrate higher percentage of renewables. It's not a silver bullet. It doesn't solve all of the integration problems but it does help to get us to a higher level of renewable penetration.

Shayle Kann: I think there's two ways to think about why you might add energy storage to one of these big corporate procurement deals. One of them is what we're talking about now which is to manage that risk and to have the project sell at higher prices within the market. That sits alongside but it's not the same thing as the other reason you might do it which would be to try to shape your procurement according to your actual load. Instead of just saying, "I'm going to match every kilowatt hour that I use in my data center with one kilowatt hour of renewable electricity regardless of when it was generated," instead saying, "I'm going to try to either perfectly match or come closer to matching my load pattern with the generation pattern of the procurement." People have talked about that as the next step of a 100 percent renewables for really progressive corporates. I mean, you would obviously need energy storage in order to do that. Have you thought about that? Do you think that's the long term vision of where this procurement heads or is that just too costly?

Brian Janous: No, actually I don't think that's a long term vision. It's not so much about the cost though. First of all, we're absolutely investing in those capabilities that are going to enable greater participation for storage in the energy markets. Whether it is integrated with a renewable energy project or behind our meter, we're doing some pilots right now on utilizing data center, UPS systems which are integrated battery systems as grid batteries so we can have this grid interactive UPS. We're really excited about that project.

Shayle Kann: Wait, for people who aren't familiar. UPS is uninterruptible power.

Brian Janous: That's right.

Shayle Kann: It's been common in data centers for years. What's new about that is using it as a grid battery?

Brian Janous: Yeah. For every server I have, every megawatt of servers I have, I have a megawatt of back up generation and I have a megawatt of batteries as well. That battery is there to bridge a potential grid outage to when my generator can start up. Maybe that's somewhere between 30 seconds and two minutes. All data centers have, or most data centers at least have this battery capacity. It usually just sits there idle. The opportunity is can we tap into the spare capacity those batteries have and use them as a true grid resource. Now that being said, we're working on that and I think it's really interesting but going back to the ultimate question of should renewables purchasers be effectively responsible for trying to solve how to integrate that resource or match their load to the resource. I think the idea that the buyers need to be concerned about that is not ultimately going to be efficient at large scale. If I put my ECON 101 hat on for a second, I think this is a textbook case of comparative advantage.

The question is, am I as an individual consumer best positioned to determine how to integrate that resource or match my load to that resource? Is it better left to the market? While it's true I could match one to one my renewables with storage for instance, whether it's on my site or in the project itself, I don't think that's necessarily the best outcome in the long run if say there's a controllable load out there that would offer curtailment at a marginal rate that's far less than the cost of my battery. I think the value of the grid is we can design markets that lead to the least cost and most efficient solution. If several years from now I'm having to worry about time matching my own renewables, I would probably say as an industry we've clearly failed to develop the kind of market that is capable of integrating high penetrations of renewables and that's really the north star for us.

Shayle Kann: I want to get into the conversation about market design in a little bit as well but I think it's interesting, the notion that you don't have to think about that as a buyer at scale is particularly interesting because at scale meaning if every large energy purchaser does what Microsoft is doing and what Google is doing and all these other companies, then you have a lot of companies procuring a lot of renewable energy. If they're not thinking at all about how that aligns with their load shape then you're really putting a pretty big burden on the market to figure out the rest. You need pretty clever market design I would think and whether or not it's clever it needs to be a decent size departure from the markets as it exist today.

Brian Janous: Absolutely. I think actually that's probably the biggest challenge and honestly why we've really pivoted a lot of our focus to this issue of market design and working more at the federal level in thinking about how our market is going to evolve in the long run. I would say a few years ago we were much more focused on state level issues, what's happening in our state houses and public service commissions. We've really pivoted because we realized that without a more intelligent market design that really contemplates that the grid that we have 10, 15 years from now is going to be radically different than the one that we had for the last 115 years. This is the first time where we're really injecting a technology that has a potential to really cause us to completely rethink how markets actually function because this is the first time we've ever experienced a truly cost competitive zero marginal cost asset. That's pretty transformative for the wholesale markets that have been predicated on there being a marginal cost to operate.

Shayle Kann: Right. Okay, I do want to come back to market design. Before we do, I want to actually talk about a couple more of the specific procurements that you guys have done. One of them you mention and was just announced which is this 60 megawatt rooftop distributed solar procurement in Singapore. As far as I know, I haven't seen many other if any other large corporate individual procurements of a distributed rooftop portfolio, certainly not in Singapore. Can you just talk us through one, why you did this distributed rooftop deal? I'm not even sure if it's all rooftop but distributed solar deal. Two, how much more complexes it to get that deal structured versus just procuring from a single 60 megawatt project?

Brian Janous: Going back to what I'm saying before, we do face unique challenges in each market that we operate in. The market structures are different. In the case of Singapore, just the sheer lack of availability of land. There isn't somewhere where we can go plop a 150 megawatt utility scale project in Singapore. They're just not land available to do that. We were really limited to just looking at what is available and in Singapore it's rooftops. That was clearly the only clear route we had to get to a material amount of energy. It goes to the question we get a lot which is, "Why don't you just put solar on your own rooftop?" We can, but if I put solar on the rooftop of the data center and effectively supply the office building that's attached to it, I haven't really done anything to address the massive amount of server load just because energy consumption in a data center is so energy dense.

We build building that are roughly the side of a Walmart Supercenter but inside of that footprint we're consuming 10 to 20 times the amount of electricity just because of the nature of what we do inside that building. We have to get creative in thinking about how we can bring renewable energy particularly because we are concerned about doing it in the markets where we operate and if we're operating in Singapore then we need to find a way to get clean energy in Singapore.

Shayle Kann: What does that look like? Are you owning these projects that are on a bunch of rooftops? Who is developing them? How are you going to be sure that there are 60 megawatts worth of rooftop solar to be procured in Singapore?

Brian Janous: It is actually similar to any other project that we would do. We're effectively an offtaker but we're not going to own the assets themselves. In fact, we look at this very similarly to a utility-scale project that there's going to be X amount of solar or wind out there. We're not going to own the asset but we're contracting for the long term output of either the energy or the attributes or both and that's effectively how we do most of our deals.

Shayle Kann: Somebody else comes to you and says or you go to them and you come to an agreement where you say, "Okay, I'm going to procure 60 megawatts of power from this distributed set of arrays," then they go out and develop the projects over the next couple of years then you procure the power when they come online. Is that more or less how the process works?

Brian Janous: Yeah, that's right.

Shayle Kann: Singapore doesn't have a huge rooftop solar market even today. Is there any concern? Are you facing some risk in just development risk at this point? Which is bigger than the risk you would face in the centralized Solar Ray where you can check the boxes and permitting interconnection, all of those things and then you know the project is going to be there because you're a bankable offtaker. Do you face any additional layer of risk when doing something like this? I'm interested because I'm curious whether you think this is a model that is going to be replicable and scalable or whether this is a one off thing because it's Singapore.

Brian Janous: I think there is a small amount of volume risk, that's the full 60 megawatt get developed but what you don't have which you have in traditional utility scale PPAs is you don't have the risk that the entire project doesn't get developed. We've had other projects that had challenges related to permitting issues that may put in jeopardy the 100 percent of the output, the project may never get built. This one, because it is distributed in some ways you actually reduce that development risk. You may have a slightly more uncertainty around is it going to be 60 megawatts or 58 megawatts but you're actually distributing some of that exposure you have in traditional development projects.

Shayle Kann: Okay, let's transfer back to talking about the U.S. and I want to talk about two procurements that you've done in tandem because I think they present interesting contrasting examples of how you are working with or not working with utilities as you go through this process of procuring renewable energy across all of your areas of load. In Washington state, in your home state, there was a bunch of press I think last year when Microsoft reached an agreement to procure directly from generators like direct access in California rather than from your incumbent utility which is Puget Sound Energy. That was framed I think often in the presses as bypassing the utility or defecting from the utility and you did have a 24 million dollar break up fee to do that.

I think it's a little bit more complex than that. Then, in a contrasting example in Virginia you ended up I think partnering with Dominion in the state of Virginia on procurement of 20 megawatt solar procurement there that looks like a green tariff type program. I'm interested to get your take on how those two examples are similar and differ from each other and more generally how you're going about working with utilities when you start a new procurement or when you want to go about having a different source of energy wherever you are.

Brian Janous: I think like politics, all energy is local. We're pursuing the same strategy everywhere which is we want access to market-based rates, we want access to price signals whether it be real time pricing or ancillary service markets. All of these enables us to do the optimization behind our meter and ensures we're putting in the appropriate technology in the right place. Effectively what the market is signaling it needs as well as giving us access to be able to participate in large scale renewable projects. We really can't do any of these without the price signal, that's a key thing for us. If the market needs for instance more flexible fast start resources, but I don't have a price signal telling me to put that in behind my meter then I'm not going to incur the added expense which is what we did in Wyoming.

Another case I shouldn't mention but we put in fast start natural gas generators there as part of a [uber 00:41:07] tariff we created with them that gave us both access to market prices, the ability to procure renewables directly as well as install generation that the grid really needed, that utility needed in order to meet some of its capacity needs. In every case-

Shayle Kann: I'm sorry, can I just ask about the Wyoming example? Is that then you negotiated a Microsoft specific tariff in Wyoming that gave you access to those prices that the real time prices and price signals that other consumers don't have because they haven't negotiated the same tariff and then provided you something akin to direct access where you can procure directly?

Brian Janous: In this case, we're actually still behind the utility. We are still a customer of the utility. They're just providing us with a market-based rate. This is actually a tariff that is open to any customer but it's predicated on bringing some amount of back up capacity into the situation. That was the key for this particular utility because as we approach them a few years ago about our growth in their service territory, it was very clear that we were going to become a relatively large percentage of their overall customer base. They had some concerns about needing new capacity. We started talking to utility and say, "You know, there's two ways we can do this. We can continue to grow inside of your footprint. You can go put a megawatt of peaking generation on your side of the meter.

I in turn I'm going out to put a megawatt of generation in the form of my standby behind my meter, but in that case we're putting in two megawatts of generation for every megawatt of load I bring to the system. It doesn't actually seem to make a lot of sense. What is instead we put that peaking gas-fired capacity behind our meter?" I had my backup resource there but also, then could provide capacity to the grid if and when it's needed. That really became the more elegant solution that we came to is that, "Hey, as long as we're bringing our own backup capacity then we can grow inside of their footprint, with reduced risk to the other rate payers because now the utility does not have to go out and build a bunch of new assets." We're killing two birds with one stone and coming up with a much more elegant solution.

Shayle Kann: Right, that's a good example of what seems like an elegant solution that was negotiated between you and the utility in Wyoming. What about the Puget Sound Energy example in Washington?

Brian Janous: One of the key differences between Wyoming and the Puget Sound example is for Wyoming we were largely looking at that new load to the system. There was not a case of stranded cost. There was a case of the potential to build a new asset but we weren't talking about a bunch of existing assets. In the case of Puget Sound, again, we had the same ultimate goal which is we want to get to the market-based rate. We want to have that price signal so that even as a campus we can start to better optimize and think about how we use electricity. In this case of course, there were stranded cost involved and that's what led to the 24 million dollar payment. It got us to the same end. We're now going down this road of we're going to be ... This is true direct access which is not really what we're doing in Wyoming where there we still sit behind the utility. In this case, of course now we're also able to then control our own destiny as it relates to our renewable goals.

Shayle Kann: In that case, you are no longer a customer of Puget Sound Energy? Is that correct?

Brian Janous: Not for generation. I mean, we're still obviously T&D customer of theirs.

Shayle Kann: Right. Then how about the Dominion example in Virginia, was that a similar case of you had existing load and needed to fulfill that or were you putting in a new data center? What did that look like?

Brian Janous: Yeah, that one's even a little different than the other two in that it was really just driving to how do we get spur demand for more renewables in that particular market. Dominion had a project that they wanted to build that they originally brought it to the commission and it was rejected. We stepped in and say, "Hey, what role can we play in this project to help it move forward?" That was really just a partnership with the utility to say, "Hey, we've got this project that they want to build and we would love to see them build. What can we do to help?" That really led us to that project.

Shayle Kann: You know, one thing that I'm noticing from this conversation is just how variable these different procurements are and the different needs are and different locations both because of the difference in the facilities that you have in different locations but the regulatory structures and the tariffs that are available, the price signals, what the markets need. I guess one question I have for you is you have the capacity to do this because Microsoft first of all has a lot of load in a lot of places and so things about energy because I'm sure it's a big cost but also you have a team and you guys are all dedicated to figuring this stuff out. Do you think that ultimately every major energy purchaser is going to look more like Microsoft does just in its sophistication in these types of things? Is it going to be limited to a small number of companies that have huge load from data centers?

Brian Janous: I think if we don't evolve the structures and the opportunities in the market, it is going to severely limit the types of companies that can participate in this because you're right, not every customer has a big energy team. Not every customer is going to be comfortable in dealing with some of the ... whether it's the technical challenges, the accounting challenges, all the things that come along with doing these deals participating effectively at the wholesale level which is what almost all these deals entail. To some degree being a wholesale participant or at least understanding some of the risk and cost involved in wholesale market participation. I think as an industry, as I alluded to this earlier, we've made great progress in opening up the number of companies that are participating in these types of deals, but we're still dealing with generally very large customers who are doing one off transactions buying 100 percent of 150 megawatt project for instance or 200 megawatt project.

We really have to get to the point where new models start to evolve where smaller buyers are able to participate to some degree through some form of aggregation or being able to warehouse projects and break them up in trunches. That's really where I think the focus of the industry is going right now. There's been a lot of talk about that over the last couple of years. I think this is the year where we're going to start to see some of those new models start to evolve and be announced. We're excited about some of the progress we're seeing behind the scenes there. I think that's going to be the next step, how do you get the 5, 10, 20 megawatt buyer into the space because they can't do it on their own. They can't do it on this traditional model. It has to be a new model that's going to allow them to come into this market.

Shayle Kann: Right. I think you're right that there's a lot of effort and attention being paid to that right now, Rocky Mountain Institute through its Business Renewables Center. It's been focused on this community scale procurement for a little while. There's also this start up called LevelTen Energy which is basically dedicated just to doing that and raised a big series A last year. I think you're right. We start to see the deals get announced in more significant volume this year and presumably some of those deals you end up with a structure that look somewhat similar to what we've seen in community solar and a lot of places which is that you have a single project. Say it's a 100 megawatt project but you'll have an anchor tenant who can take up half of that. Microsoft could be 50 megawatts out of a 100 megawatt project and the other 50 megawatts can be dispersed amongst the larger group of smaller buyers who may be a little bit less credit worthy.

Brian Janous: That's right. We think that's a very credible model and one that's likely to have some success in the coming years.

Shayle Kann: One thing that I think often doesn't get talked about a whole lot when we're talking about all these corporate renewable energy procurement deals is the economics for the customer. I think it's a presumption that you know, you set these goals to get 50 percent or 100 percent or whatever it is of your energy from renewables but the presumption underlying that is that you'd only do this if you're actually saving money in the process. Is it true that you were saving money on all those procurements that you've done? Is it economically a net positive? Are we still in the position where especially given all these risks that you're bearing about market prices and so on, or is the overarching driving factor that desire for social good and you're not actually making money on them?

Brian Janous: I think the short answer is it's both. I mean, we're certainly concerned about the economics of these projects but we're not looking at that, the exclusion of the societal good component. I mean, there's no doubt that the cost of renewables has dropped to the point that no other I guess we'll call it traditionally conventional technology can compete from just a pure cost perspective. That's on a net new basis. I think sometimes it can be a little misleading when we see these studies come out to show, "Hey, the cost of solar and wind is now cheaper than gas and coal." That's not really the issue we should be looking at. Really it's about the cost of generation on a net basis. That's really the appropriate comparison. What we're concerned about is whether the cost of new generation will be adequately compensated in the whole sale market.

A market that really in many places is already amply supplied with energy. It's also a market that's experiencing very high rates of renewable penetration. Certainly, in pockets like Texas for instance. I think the mistake that many corporate purchasers make is they fail to take into account in their forecast the impact of their own actions in the market. I think again, this is something that you've talked about on your show and I think done a good job of surfacing this issue is that renewable energy cannibalizes itself. When you take a forward curve and just assume, as I said before it's just going to slope upward. I don't think you're really taking into account the fact that we are altering the market as we continue to develop more and more renewables.

Now, as it relates, that's really more of an issue to think about again the market design and how do we start to evolve in design markets that allow us to do this decarbonization we need to do, because technology is going to be a part of it. Technology isn't going to solve the problem. It also has to be a market piece. On the societal question, we recognize that we are building what we consider the electric grid of the 21st century. If you think back to the 20th century, nearly every innovation that we achieved in the society in some way can be traced back to electricity and electrification. With all that advancement, we created this legacy and this dependency on fossil fuels. In this century, the prime mover for innovation is going to be data.

Microsoft along with a handful of other companies are really building out that underlying infrastructure that will support that innovation. A core value for us as a company is that we're going to leave a positive legacy with what we build. That means the electricity we buy which is electricity is really the raw material of a cloud, that's going to be renewable energy.

Shayle Kann: Let's talk a little bit more about the market design then. I know you've been pretty active. You mentioned you had been more active at the state level and then now increasingly at the federal level as well. What should we be paying most attention to on the policy or regulatory front if we're in agreement and alignment with you on where the market needs to head? What are you focused on?

Brian Janous: I think we're going to continue being in a place at least for the next few years where we're playing both offense and defense on the policy front. For instance, we want to continue to support FERC in the upcoming technical conference on DERs. I think that was a really great order that we got a couple of weeks ago around you're really highlighting the importance of storage in this market. That's really important to us. We also want to continue opposing any proposals that are focused solely on propping up Out Of The Money assets that would result in really undermining markets in such a way as to distort price signals because the risk that we see there is that we know that flexibility is valuable to the grid both today and going to be increasingly valuable in the future.

What we don't want to see is a market where it's not valued. That's I think where a lot of people that are focused just on the technology are missing the market component that just because you can create a great new storage product doesn't mean the market is actually going to reward you for that. If we fail at the market design piece of it then what we're left with is a lot of great technology that the market is not valuing. That's why focusing on these policy issues is so important to us.

Shayle Kann: I'm wondering if you have a view on the right way to value flexibility? I've been having a bunch of this conversations lately and it seems like the direction the markets are heading right now are sort of two different paths you can go down or there's probably more but at least there are two that the markets are currently going down. One version is to say we recognize that flexibility is going to be the coin of the realm in electricity for the next decade or two. We're going to introduce a product in our wholesale market specifically to value flexibility. That's like PJM has proposed that the CAISO has a flexibility product and is revamping it right now. That's one version of what you can do.

Another version of what you can do is head toward Texas and introduce an energy only market that by having prices that fluctuate wildly inherently values flexibility. I've heard people make cases on either side. Do you have an opinion as one who has both going to be consuming and providing flexibility on which of those is going to make it easier for you to build that one the grid needs?

Brian Janous: I think we are going to have to get more sophisticated in the products that we offer. I'm not fully convinced. I can see myself in some ways a little agnostic on this point. I still think there's a lot we need to learn about how markets can evolve and much of it is a software question. How can we better visualize what's happening on the grid? Can we provide the type of price signals that result in actual investment? It's not clear to me yet that an energy only market provides that level of price signal. Is someone really going to put in a lot of new stationary storage in the taxes in the hopes that we're going to see some $9,000 megawatt hour prices in the next few years. Maybe they will but from what I've seen from the appetite of banks to back those investments, that's a big if. I think we're really going to have to experiment with the types of products that we're going to have to create in the market that actually result in this investment.

We're in somewhat a unique situation because as a data center operator, I mention that for every megawatt of servers we have a megawatt of generation and a megawatt of storage. We are going to be building assets. That's just part of building out our asset base. What we're looking at and saying, "Hey, if I'm going to be in the market anyway I want to make sure that I'm optimizing that investment to what the market needs." Of course, I need the price signal too but I'm a little different than say just an IPP that's going to go build some gas plant on a merchant basis. That's probably why we're so interested in these questions because we are investing a tremendous amount in all of these assets. That means we really are focused on getting the markets right so that we can have that signal to know that we're putting in the right thing at the right time.

Shayle Kann: I want to transition for a minute from talking about Microsoft as a large energy consumer to talking about Microsoft as a technology company and the types of things that you're working on that pertain to this energy transition. I know that you've given some examples publicly, working with an Australian startup on a residential internet of things, technology that relates to energy consumption. You've worked on aggregated Smart Water here and Hawaii. You've worked in Holland on a partnership that's advanced smart grid thing. Of all the ways that Microsoft's technology touches the electricity market, what do you find most exciting?

Brian Janous: Actually, that's not even part of my role because I'm focused on the energy to operate our data centers. Just because of where I sit in the industry I also do get involved in a lot of these other issues and how are we thinking about this as a company and the opportunity. As we've been alluding to all along, we need a new type of grid. We need a grid that is interactive where we have markets that support new technologies. There's two barriers to that. One is the technology, deploying whether it's ... We get to the whole debate about whether it's about deployment or about bringing the step changes in technology but it's a technology solution on one hand. It's also software. It's about how do we develop the right software architecture so that markets can evolve.

One of the limitations we have in developing say new products for flexibility is we have to have a better way of assessing things like participation. How are we evaluating the value that say a power wall in someone's garage is bringing to the grid? How do we compensate them for that? That's a data challenge that we have. In some ways I actually think the software side might be the easier part of it. How we then evolve markets around that software? Not the software itself but the intelligence we can gain, the ability to visualize what's happening on the grid. That's really where we see the opportunity as a company. How do we work with utilities, grid operator and other partner to help them better understand the grid as an organism?

Not this in some ways this simple machine that we created over the last hundred years although it's elegant and it's beautiful in so many ways. We're really just [inaudible 01:00:35] managing hundreds, maybe if not, a thousands of different points on a system. Now we're talking about potentially millions of different points on a system that need to be integrated and optimized in some way. I mean, that orchestration is an enormous challenge for us if we're going to get to the level of decarbonization that we need to see.

Shayle Kann: As the fact that big data and machine learning and artificial intelligence, all these other buzzwords have taken such a position at the forefront of the minds of companies like Microsoft, has that made electricity sexier to you guys as a company because electricity being one of the probably handful of existing markets in the U.S. that generate crazy amounts of data? Now with the additional, all these new resources present all these new challenges and opportunities. Has it elevated the role that electricity or more broadly energy plays within Microsoft? Is it still a point solution?

Brian Janous: No, it absolutely has. I mean, there's no doubt that electricity is way sexier today than it was 10 years ago. As a lifelong electricity guy, it's exciting for me to see that people start to think that our industry is a little bit sexy. A lot of it is is because of really the radical transformation that new technology is driving in the industry. If you go back over the 115 years of history of the electric grid, there was never a time that where technology both the hardware and the software is driving the transformational change that we're seeing today. It really is an area where five years ago if I'm walking around the halls of Microsoft and talking to people, what people were excited about was transportation, buildings, how can you automate those assets but today there's a lot of excitement about electricity because people see that, "Wow, this is a space that is transforming so rapidly and the foundation of that transformation is going to be data."

Shayle Kann: That's great. Brian, thank you so much for taking the time.

Brian Janous: Absolutely. Thanks for having me, Shayle.

Stephen Lacey: Well, here we are at the end of another set of good conversations. If you made it here after a significant amount of time, well then you're passionate about this subject. That doesn't mean the other people aren't passionate but you clearly you got a lot of value out of these conversations. If you think other people are going to get the same kind of value then pass a link on, tweet it out, share your favorite quotes from this episode, tell us what you think. You can find me and Shayle on Twitter.

You can also find The Interchange Show on Twitter as well, just @interchangeshow. Referrals are really good way to get us more listeners, passionate listeners because if you share your passion with someone else, chances are good they're going to apply that passion to the show as well and it just goes on and on and on. You can find us anywhere you get podcast so if you're not a subscriber already, please go ahead and do so. That's about it. Shayle Kann will be back with me next week. Until then, I'm Stephen Lacey and this is The Interchange conversations on the future of energy from Greentech Media.