The Exchanges

Q the devices themselves weren't powerful enough? Was it That it was, so it was the wrong infrastructure? Was it just as a data collection device, the utilities didn't know what to do with that data? And they, you know, they, they had an inability to use it properly. Give us your broad view on what went wrong with AMI one point, which I think most people think wasn't that successful.

A Yeah, look, I, I think it was basically the wrong architecture. Look, it made sense back in 2008 or whenever that happened, but it hasn't transitioned since then. So, and by wrong architecture, look, back to my telecom example, imagine if the way Google Maps on your phone worked is your phone would collect 15 minute interval data of your location, send it up, ah, in batches to, to your telecom provider, Who would then make it available through, what, gray button instead of green button or something like that, and then applications like Google Maps could get that data a day later and then do something with it. Well, what would Google Maps on your phone be? It would be like a static map and, and maybe a monthly historical report of your traffic on your, your route to work and maybe compare your neighbors, uh, how you, your drive compared to your neighbors, but would you use that app? Like, you might use it every now and then, but you would not engage in it in the way you do with Google Maps. And what I just described exactly matches AMI-one.o architecture. Low-resolution data sent up to the service provider, made available, uh, later on, and look, there's some things you can do with that, but you just can't have a real-time, consumer-facing app, and then you can't see the grid in real-time from the edge either.

AI assessment note: “Yeah, look, I, I think it was basically the wrong architecture.”

Q So let's apply this to actual use cases out in the power sector. What were some of the successful use cases in AMI one point O and what do you think the new use cases for this next generation of smart meters are?

A So let's, let's start with the consumer side on the consumer side with AMI one dot O you could provide a kind of next day view, uh, you know, of how power was used in your home yesterday. So, so like this was used in these various portals that utilities provide that give People have some insights into how their bills tracking, which is a big leap compared to waiting to the end of the month before you have any idea what's going on in your home. Look, frankly, not a lot of people use those portals. I think there's pretty well defined metrics that these haven't got a lot of use. Once you get to this new architecture, applications like Sense become possible where we can provide consumers with their real-time, ah, detailed view of what's happening in the home, and this becomes super relevant for, ah, energy efficiency. People can see what's going on in their home and track down energy hogs, we call them in the home. It also is having a big impact on people's participation in Demand flexibility, load flexibility, you know, how do you deploy time of use rates or demand charges if you can't let the user see what's happening in real time in their homes. And then we're also using this high resolution detailed view of what's happening in homes for helping the electrification. Find the homes that are best candidates for heat pumps and so on.

AI assessment note: “with AMI one dot O you could provide a kind of next day view”

Q Greg, Tell me more about what the most, the highest value propositions are. Tell me about cost comparisons. How much are you factoring in, like, emissions and the, the, the, the move to factor in embodied emissions in the development and purchase of solar products? What are the most significant value propositions that potential customers are interested in?

A Steel frames. Is a very strategic way to de-risk the supply chain. It's not the whole supply chain for a module or for a project, but it is material being number two or three, uh, on the old bomb. And so I think if you look at the value proposition, it's stacked, but I'd say de-risking the supply chain is first and foremost, because when you have the existential threat of interruptions, Or arbitrary price increases because of, uh, tariffs or political decisions made on either side of the Pacific. That's really hard to manage. It's really hard to hedge. So that's it. And you get a great combo, given the domestic content that the IRA provides in terms of the tax credit. You get a material impact on increasing your domestic content as you De-risk the supply chain from Asia. The next one that's really starting to play out, and this has been coming, you know, I'd say the last six to nine months, there's an increasing focus on module fragility. Uh, the reality is the size of modules have increased over 40%. At the same time, there's been cost down efforts on both glass and frames to reduce the material And thus, the structural capabilities. And so, we have this perfect storm of cost reduction, the core structural elements. At the same time, we're seeing more and more weather, extreme weather events, be it wind, snow, or hail, that is requiring a higher performance structural support …

AI assessment note: “if you look at the value proposition, it's stacked, but I'd say de-risking”

Q How did you land on this 90% reduction in GHG, embodied GHG emissions in production of these frames compared with aluminum module frames? Um, yeah, this was a report that you put together with, uh, Boundless Impact, Boundless Impact Partners, is that right? And, uh, so tell me about like that analysis and the actual comparison and how you came to those numbers.

A Yeah, we did. We realized that, you know, just like with our testing results, we want to use third parties, very, very credible third parties, and Boundless Impact is a very focused, uh, engineering consulting firm that does lifecycle analytics, and so we basically just told them the, uh, the characteristics of our steel frame, and then they researched aluminum frames, and they developed this, and that 90% Reduction per average size module is based on their analysis, and we've updated it with actually our definitive supply chain, and it got a little bit better. But it really is a rigorous analysis, and that 90% reduction, which is a really big relative number, translates also to about a 90 kilogram carbon footprint reduction per module. And if you scale that to gigawatt scale, I think it's like a 180,000 metric tons of carbon avoided per gigawatt. And so, that is absolute materiality if you look at it. And so, we're also working very closely with the Ultra Low Carbon Solar Alliance, we're members, and they have been pushing the EPECO label. And we are literally geared up, and with literally third-party validated reductions, what we're able to do is provide to our customers that buy, the module manufacturers, a absolute data-driven reduction that they could go and try to achieve the EP criteria that I think is going to become increasingly more important going forward, because we…

AI assessment note: “we basically just told them the, uh, the characteristics of our steel frame”

Q We have another example of a solar technology that has, uh, been reassured, and that is trackers, and steel has been critical for the growth of that industry. Uh, Greg, any, uh, corollaries to the, the, the tracker market? Like, can module frames follow trackers?

A There's, uh, absolutely similarities. Uh, it's, one, it's the proof bank that says, How quickly you can reshore and create the local jobs and economic upsides without any material negatives for the solar industry. And in fact, reshoring and using steel, uh, accelerates and increases confidence in project development and construction by shortening the supply chain materially without concerns about logistics costs or uncertainties in delivery. And given the bottlenecks we have continued to have on all the ports and with the Random aspect of customs quarantine. It, it gives confidence. It actually improves the execution in the region. But the one difference between us and the tracker company is that we are taking advantage of a material cost difference. We're not just reshoring the same material to the U.S. We're actually reshoring using domestic steel, which is Traditionally been about the third the cost of, of aluminum. So we get to bring in a new metal, reshore the, the production of that, build the worldwide global competency center for it here, ah, instead of following and hoping to leverage from the Chinese. Ah, but how do we do that? And we can do that because of the inherent benefits and cost effectiveness of steel. We can do that at parity. And eventually, materially lower cost without sacrificing the performance and the decarbonization, ah, that, ah, comes with aluminum.

AI assessment note: “There's, uh, absolutely similarities. Uh, it's, one, it's the proof bank”

Q And are you talking about just heat, or are you talking about hydrothermal resources, or just hot rock?

A Yeah, great question. So I'm talking about, um, super critical areas where you can, you have enough heat that either using liquid that's in the ground, or, uh, you, or using above ground liquid You can do a thermal transfer and make electricity. So this is not about heating and cooling. Uh, there are a bunch of great companies, Dandelion and Bedrock and others who are working on that. This is about power generation systems that have power plants, uh, top side that can pump tens or hundreds of megawatts of power onto the grid. So Zanskar doesn't have special drilling technology. They don't have new fluids or new rink and cycle systems on, on the top. They literally just Have a better way to look for geothermal, because in America, there's only a few sites, and those sites were where we, like geysers in Northern California, where we see stuff coming out of the ground. In Iceland and Indonesia, there are a bunch of sites, because you can see the steam coming out everywhere. It's very accessible to get to that very, very hot area underground. Today, when we do, when we drill wells looking for geothermal, we'll drill 10 wells and get one hot spot. And, and that's not cost effective. Those wells are very expensive. With Zanskar, they have been proven that they can drill 10 wells and get 90% hit rates, and so nine out of 10 instead of one out of 10. That flips the switch on Zans, on, …

AI assessment note: “So I'm talking about, um, super critical areas where you can, you have enough heat”

Q What has surprised you about deployment trajectories? Did any particular trend stand out to you as you surveyed the landscape?

A I think we were struck by how bad we are at forecasting climate technology, but also how inconsistently we get it wrong. So there's some sets of technologies that we consistently overestimated, and those are the technologies that are always just around the corner, but somehow they haven't panned out to the extent that we predicted. And those are things like, uh, nuclear fusion or things like hydrogen, um, the CCS to some extent, where historic forecasts are much higher than what we're currently seeing or currently projecting. And, and it's surprising that we kind of didn't predict that correctly as a society. I think the second thing is indeed to what extent we keep underestimating the speed, uh, at which, um, solar, um, and batteries to some extent also wind, but to a less extent have, have achieved cost savings and actually are deployed. So somehow we seem to look at, at, at averages of climate technology more than really Uh, recognizing the differences and, and, and being able to forecast these correctly.

AI assessment note: “we were struck by how bad we are at forecasting climate technology”

Q that aren't necessarily designed to equip some of these, uh, peak loads. Others say AI is just different, you know, it's going to require exponentially more energy. Where do you, each of you fall on this spectrum of concern? Like, as people come to this and try to figure out just how substantial is this issue? Um, uh, Brian, we'll go to you. And then we'll go down the line.

A Yeah, as I, as I sort of alluded to before, I do think it's very real. And I do think that, you know, probably the most extreme projections are overblown. Um, I think on the other end, you know, the ones that I've seen on the low end, I think are, are pretty, pretty off base as well. Um, we've got a great indicator of investment and growth in this space, and it's the capex expenditure for all the big cloud companies. And if you look at that quarter over quarter, year over year, They are all accelerating investment. It's a great leading indicator of how much power, how many chips they're going to plug in in the future, um, and they're all betting really big on very large AI models, and so that's maybe one of the differences between AI and what we've seen in sort of native cloud is that there is very much an incentive to use more electricity. There's incentive to build bigger models to plug in More GPUs. And, you know, I had someone ask me the other day, well, what happens if, you know, if NVIDIA comes out with a chip that's twice as efficient? And my response to that is, well, then Meta or Microsoft or Google will plug in twice as many chips. If they have the power, they're going to build as big of a model as they can build. And so that's where I think AI is a little different than what we've seen. Not to say there's not incentives for efficiency, because there very much is, But…

AI assessment note: “I do think it's very real. And I do think that, you know, probably”

Q Yeah, and how does that actually work? Like, what compute workloads are you batch processing? What does that limit you to? And then how are you actually, like, contracting for those excess renewables?

A Great question. So, um, because the energy is intermittent, uh, we focus on compute that will be resilient to that intermittency. So we look for, uh, software and computing that essentially is batchable. So think about, uh, Uh, Netflix, uh, everybody watches Netflix, or your favorite streaming, uh, channel. Uh, that's a very real-time process. Um, you need the data center that's serving it up to you, or the network of data centers that's serving it up to you to be always on, because you don't want that movie to, to stop, or that show to end. We don't focus on those types of applications. So, you won't see Netflix, you won't see your ERP system, you don't see your e-commerce system in our data centers. What you will see is a, uh, machine learning process that is trying to, uh, find different applications for a molecule. You might find, uh, a secure system, security system that's, uh, supporting the, uh, addition of a new block to the, to the Bitcoin blockchain. Uh, you might find a, uh, a new inferencing process that's helping to improve customer service in the telecommunications system. Basically, anything that is, ah, resilient to that power loss is something we call batchable processes, things that you can do through a series, pause it, or put it to sleep, and then wake it back up when the power, ah, is available again. And it turns out that distributed computing has those el…

AI assessment note: “we focus on compute that will be resilient to that intermittency”

Q And so what does that mean for the engineering teams who are deep in this? What is, what kind of scramble does this set off inside Tesla?

A It means that, so the, uh, the Roadster, remember, was the first car that Tesla produced in 2008, and he had been promising a reboot of the Roadster for years, and now He's going to do it. Now he says, okay, now I'm, I'm going to do it. It's going to come out in 20, 26. And this means that the body crew, the battery crew, the electric motor, the whole powertrain, everyone has to be aligned in terms of the, of the specs that he establishes for that vehicle. In this case, the roadster. Now, if you're going twice as fast, That means that within the battery, when, when you accelerate, that means that the, that the lithium ions, the lithium that's in the battery is traveling from the anode to the cathode twice as fast, right? That means when you push down the accelerator, what that means is that the, the lithium is moving from one electrode to the other. Now you're, you're, you're saying, well, we want those electrons to move twice as fast, and it's not, You can't just say that, right? The, the, you, you have to set up the electrochemistry so that can happen, uh, without the lithium really getting gummed up in the anode or in the cathode or, uh, or, or getting lost in the, uh, electrolyte, and when you move like that, you create heat. You can't have the battery over, over heat. Uh, lithium is a highly volatile A highly flammable metal, and you, you have to, one of the big challenges…

AI assessment note: “everyone has to be aligned in terms of the, of the specs that he establishes”

Q it's wild because of what it does to the teams inside Tesla. Uh, there's some design crossover between the Model Two and the RoboTaxi, which you write about in your recent story, but they're very different cars. Um, why is that such a radical move for the teams working on these cars to suddenly slow or even pause the Model Two work and then put it all into the RoboTaxi?

A One thing to know is that a lot of people who work at Tesla are there because they believe in sustainability. They believe in the mission that was, uh, that was set 14 years ago to electrify the world and to help resolve climate change. The 25,000 dollar Model Two, the one they were developing, was The key, the, the key, um, to doing that, to getting there, right? That was going to be the, the, you know, ten million cars a year by 20 30 of those vehicles were going to be, were going to be sold, and that, that was how they were going to get there. When you cancel the Model Two You're messing with those people. That's the reason why they're at the company. But let's say that you're in it, you know, that you say, okay, well, you know, okay, Elon wants to do it this way, so let's do it that way. So you've been, you've been spending all of your time engineering that model to getting the battery, getting the powertrain, getting the vehicle ready for mass production next year. And That's no small thing, right? Because you're, you're, you're taking cars, right? The Model Three, the Model Y, these cost in the mid-thirties and into the forties when, you know, when you're adding other features to them, and suddenly you're saying, uh, you know, we want, we want a car that's 40 or 50% less, so you have to do all kinds of things, especially to the battery, to get there. All of that work is p…

AI assessment note: “All of that work is put on the scrap. On the scrap heap.”

Q confusion in the electric vehicle market. In mid-April, Musk announces major layoffs across the company because of lagging sales. A senior leader wrote An email to the supercharger team at that time saying must believe their work was critical, which you reported on. And then two weeks later, that supercharger team was gone. Uh, leadership was scrapped. The whole team was fired. Why was that such a head scratching move?

A The, this is the supercharger team, especially those are the crown jewels of Tesla. People are not going to buy an electric car ordinary. I mean, I mean, mainstream you and me, right? We're not going to buy one unless we think we can charge it up when we want and fast. Musk from the beginning understood that and built ahead of demand, this global Supercharger network across the United States and across the world, these places that, that not only charge fast, not only were, were convenient, but were pleasant to go to, right? You know, you wanted to go charge, I have friends, right, who go charge their car and sit in the back seat and, and, and play video games. This is fun, right? And, uh, and now, Musk is saying, well, actually, we're not going to build them at the rate we were, and I'm firing the entire team in charge, 500 people, and the person in charge of that team. I'm firing all of them. It, it, it's not just head-scratching, it's mystifying. You get rid of the, uh, team responsible for your most valuable asset, and You put all the bets of your company, you bet the company on a technology that isn't, isn't anywhere near ready for the market. All of these things together, they're not rational. They are not rational.

AI assessment note: “You get rid of the, uh, team responsible for your most valuable asset”

Q So, Sadia, you have, uh, over your career gone from building Renewable power plants to product management in wind, uh, for large wind turbines to deploying distributed resources. Um, what insights do you have about DERs now, now that you've been in this field for a while, that you didn't appreciate before?

A The customer engagement piece of it, you know, because I think when I started my career, we, I was focusing primarily on large utility scale, round-the-meter assets, and, you know, building that out. The National Solar Mission was announced back in 2009, and the very first, ah, plants that got built under Tata Power, which is where I was working, that, all of that was happening at the time, right? But now when you're looking at a lot of the, ah, innovation, the growth, The activity, the dynamism in the industry, it's all behind the meter. And anytime you're going behind the meter, as Apoor said, you are, um, touching something that's personal to a customer, right? But where the machine learning and AI piece comes into it is that we think about this as a network of distributed DERs that need to be orchestrated and managed, uh, to deliver flexibility. Now, whether that's an EV, Whether that's a smart thermostat, a water heater, a behind-the-meter storage, a microgrid, it's how do you, how do you do this consistently? How do you account for behavioral impacts? How do you use all of this data that you're getting back from the assets to be able to predict what you'll get, to be able to deliver it for network management use cases, for the system level use cases on extremely hot days or cold days, um, and then also as an alternate source of supply, which is where the virtual power pla…

AI assessment note: “The customer engagement piece of it, you know, because I think when I started”

Q what you can do when you have cheap, abundant, renewable electrons, and you said that this would not be a feasible solution even five years ago. I'm wondering, when we think about both the benefits to the grid and to the industrial customer, how do you frame this? Do you think about it in terms of grid reliability, or do you frame it specifically around the unique industrial use case?

A That's a great question, and the value to different participants in the energy system is different. We're working with EDP at scale in Europe. EDP is, I think, the world's largest owner of commercial industrial C&I Solar. They're a major player building utility scale wind and solar. They have a substantial electricity trading and grid operations capacity. They're able to bring those skills together to monetize the value of, ah, flexibility and grid services, to trade in day ahead markets, and to put together a very specialized kind of power purchase agreement for energy that's delivered when they're controlling the charging knob. So as to optimize the value of existing assets and new assets, and then put that together with their development skills, and go to customers and offer customers a fixed price, long term, heat as a service contract. During our announcement just recently, the EDP commercial CEO said, this will make our average CNI project maybe 10 times larger, and allow us To not just deliver 20% decarbonization to a customer, but 80%. Cutting into their giant scope one emissions and saving them money. So, there was grid value, again, in, in having large centrally controlled, and right now controlled, dispatchable demand. Jesse Jenkins has, is publishing a paper shortly exploring that. His paper findings really excited us because he found that in looking at systems that…

AI assessment note: “the value to different participants in the energy system is different.”

Q How should we think about the size and scope of the market for heat batteries? Like, what's the total addressable market? How much of industrial heat demand can these batteries serve?

A People have looked at this from two or three different directions. That's exactly the right question. From a temperature standpoint, so electricity is electricity. You can convert power from one volt to a million volts back and forth. Heat is not heat. If you're going to deliver heat at some temperature, you need to store it at that temperature or slightly higher. The temperatures that today's Rondo heat batteries deliver serve about 90% of all the Industrial heat used in the world. The one exception is about 40% of the energy used in making cement, and the steel industry is flipping over from using coal to hydrogen, and most of that energy is not actually heat, it's chemistry, converting iron ore. So it's about 90% at the current temperature regime. If you look at that Tesla's assessment, when they looked at the all-in energy demand, that's about 40 terawatt hours of installed capacity. Their assessment was that it's about three trillion dollars of total for the industrial heat that was in use in the world four years ago, I think. Obviously, this is a sector industry, the industrial sector is growing as the world becomes richer and grows, so decarbonizing it Getting it onto this track of decarbonizing rapidly so we can just stay even is critical, and we see the path clear with what we're doing to decarbonize most of that within about 15 years, given the constraints on how fast…

AI assessment note: “deliver serve about 90% of all the Industrial heat used in the world”

Q cutting U.S. industrial emissions by 85% by twenty-thirty-five. And the DOE also just unveiled Nine billion dollars in funding for industrial decarbonization projects, including a couple projects that will use Rondo's heat batteries for beverage production facilities. What do you make of this strategy generally coming out of the Biden administration, and what are the levers of government support that are most helpful for this particular slice of the market?

A The DOE decarbonization drive has just been phenomenal. The Biden administration is doing something that has Never been done before. DOE has assembled expertise from across other portions of the DOE and the government focused on industrial decarbonization. One of the things that DOE, in all the applications, materials, and the evaluations, they said very clearly they're looking at things that will step and repeat. We're trying to build first of a kind larger Installations that will then directly inform ongoing deployment of that technology in that sector. That is super valuable because many industries, a technology may exist in some other industry, but it's until it's used in this industry, it just doesn't get attention. It's not believed to be credible. It doesn't wind up in people's strategic planning for the next factory or for the next overhaul. So, we've been hugely excited to see the breadth of technologies that DOE is supporting, and yeah, we are being pulled into multiple industries by this program, and one of the things that was announced just this week was a deployment of Rondo heat batteries in the food and beverage sector with Diageo, who have been an innovator Looking at every other decarbonization pathway for food and beverage. They've deployed heat pumps. They've deployed electric boilers. This support has created the drive for our first two heat battery projects…

AI assessment note: “We're trying to build first of a kind larger Installations that will then directly inform”

Q This moment in 2014 is such an interesting one because it makes one speculate about what could have been. Uh, what do you think would have happened if Apple had actually acquired Tesla, which at the time was worth one 20th of what it is today?

A You know, I think if Apple ultimately acquired Tesla, it probably would have been, and this is just my opinion, a good thing for the company. I think they would have made it far more Uh, operationally efficient. They would have made it leaner. It would run, you know, an Apple-designed operating system. They would get some semblance of autonomy, and I quite frankly don't think Elon Musk would have lasted long, uh, at Apple, right, with Tesla. I think Elon Musk would have wanted control. I think Elon Musk would have wanted to run, you know, probably the whole company, obviously a very ambitious guy, and certainly I don't think that would have worked out. I think they would have clashed. I don't think his Culture, the way he handles himself, uh, you know, no judgment in either direction, you know, really meshes with how, you know, Apple as an entity operates, and so I think they would have taken, you know, a lot of the learnings that Tesla had, a lot of the underlying technology. Now, Tesla has some great tech from batteries to manufacturing, uh, to, you know, powertrains and electric car engines and such, right? That could have all been very useful for Apple, and I think they would have Taken Tesla as a basis and, and built, built on top of that. You know, what Apple does is they buy an entire company or some sort of algorithm and use that as a basis for a feature. Apple bought B…

AI assessment note: “I think if Apple ultimately acquired Tesla, it probably would have been”

Q Definitely a lot of disagreement over the design. I'm curious, what, what were some of the design changes over the years?

A So, the overall car design was pretty much the same from, from start to finish, ah, with major changes. I would say design one to the, to the final design, ah, was all in, in evolution. And I think that the, the very core of the car looked something like the Volkswagen ID Buzz, Which is their, you know, new generation concept, uh, reimagining of the 19 fifties, 19 sixties, uh, micro bus, right? Sort of that, uh, if you've seen Scooby Doo, sort of something like the bus that they, they drive around in. And the idea was to have reclining seats, uh, seats that can turn, uh, into places to put your feet up, uh, couch like living room seating, gigantic TV in the middle of the car cabin, A high tech cockpit with chairs that can swivel, uh, doors. One incarnation had sliding doors. Another had doors like in the Tesla Model X, the gullwing type doors. All sorts of crazy ideas. They had ideas about not having windows that can open. And so instead of that, they put microphones outside the car with a, you know, a high tech spatial audio system inside the car that could present The user with the feeling that they have open windows. They had a high tech air conditioning system where the air flow would flow through the sides and top and bottom of the car, sort of like what you see on some of the, the newer, I think air buses or Boeing airplanes. Speaking of, of Boeing, you had, you know, on …

AI assessment note: “One incarnation had sliding doors. Another had doors like in the Tesla Model X”

Q leadership that let the program linger while everyone else in Apple was cringing. When asked what went wrong with the effort, a senior manager involved in the vehicle's interior design replied, what went right? At what point did it turn from A Skunk Works project that could help, a very expensive Skunk Works project that could help Apple create its biggest product yet into something that was a little embarrassing.

A Well, certainly the, the first embarrassment started to, to come out around 2016, 2017, when they had to do the first major pivot of the project, which was refocusing around that autonomous driving system, paring back some engineering and operations related to the development of physical car hardware. In focusing all your attention, or some of your attention, or a greater degree of your attention on the autonomous system, hoping to nail that down, and then eventually pivoting back to building the whole car. And that was a big embarrassment. They had to, unfortunately, lay off hundreds of people at that point. Uh, you know, a bright spot of the project was when Doug Field returned to Apple. This was in the tail end of, of 18, and he, prior to that, was running the Model Three in all car hardware engineering at Tesla. Which had just got through the Model Three issues, and, you know, was on to, you know, the great success that it has seen in recent years. So Doug Field returns to Apple. Before going to Tesla, he was a Mac hardware executive. He comes back, and everyone thought, this is now the time when the Apple car is going to go full throttle towards development, and that was certainly the goal. He implemented a lot of the fail safes for the backup to the autonomy, but he ran into issue after issue related to disagreement, Indecision from top Apple executives, including Tim Coo…

AI assessment note: “the first embarrassment started to, to come out around 2016, 2017”

Q So, there are a lot of folks on the Special Projects team that are gonna move to generative AI work, to work on the Vision Pro. How important are those products to Apple's future?

A You know, I think Apple sees the Vision Pro as potentially a long-term replacement for multiple different product categories. I think they see it as something, ah, that has potential to replace the iPad as a consumption device, potential to replace the, ah, the Mac as a productivity device. And certainly something when you're shifting to augmented reality, uh, has the potential to be an iPhone replacement. Obviously, the Vision Pro would have to come down, uh, significantly in size, weight, and cost to be an iPhone replacement, but there is so much potential in sort of the head-worn, wearable space, and a lot of potential in the operating system they built there. A lot of potential, uh, in the application ecosystem that they've been building up there with ARKit for so many years, and the applications you're already seeing on Vision OS. So, I certainly think that's a critical category for them. I think generative AI is going to be at the very core of a lot of what Apple is building for both, you know, today and tomorrow. So I think you have some of the smartest engineers in the world working on the fundamental problems of autonomy at Apple on this canceled car project. And so placing those people in very important areas like generative AI, which is at the very core of what you're seeing from NVIDIA and Meta and Google and Microsoft, OpenAI, Anthropic, you name it, I think that w…

AI assessment note: “I certainly think that's a critical category for them. I think generative AI is going”

Q As you say in one of your stories, Apple saw this as the ultimate mobile device. Um, but as you reported, there were disagreements from the start. What were team leaders disagreeing about from the beginning?

A The big disagreement that, that Apple was facing from the get-go of the project was the level of autonomy to work through, right? Would it be something close to no autonomy? Would it be something like level two, like you see in a Tesla? Would it be level five autonomy, uh, like Apple ultimately landed upon as its ambition, right? And so that was a big Area of disagreement. Some people wanted to start with something like level two, or even, you know, level zero or one, and build upon that over time. And so I think that core fundamental technology is where the disagreements really began. There were also disagreements about who the car should be for. Should it be more like a traditional car? Should it be for people who love their cars? Or should it be for people who don't love their cars? And people who just want to sit in a microbus and have it act like a private jet or an airplane cabin or a living room on wheels, so to speak. And so you had differing opinions. You also had different opinions from the finance people at Apple. Do we really want to enter a market that is so incredibly competitive and so incredibly low margin? So anything you could really think of where there was room for disagreement or difference of opinion, there was Disagreement or difference of opinion.

AI assessment note: “The big disagreement that, that Apple was facing from the get-go of the project”

Q more training, and then not to mention regulations need to change, and consumer confidence needs to change, and suddenly this last decade of history shows us that it's almost a little ridiculous that they hung on to level five autonomy for this long. Why did it take so long for them to make this change? Decision when it was clear that there were a lot of struggles in the market.

A Well, it felt like they had come to some sort of solution, which is to build more of a Tesla Me Too product, which is to build something that would require less advanced autonomy, something in the level two plus three range, and they definitely were working on that. They had pivoted back to installing a steering wheel and pedals, but the problem is, is that you weren't really going to add much to the marketplace. Maybe you had a cool Apple design and user interface, but it certainly wasn't going to light the earth on fire, and I don't think they wanted to ultimately release something that was a me too product to Tesla. And so they, you know, ultimately made the decision to wind down the program. I think ultimately, to be honest with you, the biggest factor was probably, ah, the lack of profitability and the lack of strong margins that they saw coming from the car project. You know, on average, Apple's, you know, corporate gross margins are, are above 40%. You know, there are car makers like Lucid and Rivian that are losing Tens of thousands of dollars on every car unit shipped. And certainly that's probably not something Apple wanted to deal with on the bottom line. And if they were able to create a car without losing money, something that was profitable, uh, they would probably have to overcharge for the car. So, you know, in many respects it was a lose-lose, uh, for the compa…

AI assessment note: “the biggest factor was probably, ah, the lack of profitability and the lack of strong margins”

Q Well, let's go to a couple examples. You know, you've worked with a range of utilities and retail energy providers, and you said grid operators on using AI for asset management, for forecasting. What are the most compelling applications you're seeing today?

A When I think of all these use cases, I typically think of three areas. The first is assets and operations. So that has everything to do with the physical infrastructure, grid, power generation, et cetera. The next is finance and markets. So that has everything to do around power markets. And the third is everything around customer, customer three 60. So making sure that every individual who is part of this energy transition is treated in the right way. So first in assets and operations, two examples come to mind. One is we work with a retail energy provider in the Midwest that serves more than a million customers with electricity, gas, And distributed energy resources. So they have an offering, for example, for rooftop solar, and I believe about 20,000 of their customers have rooftop solar through the energy retailer. Now, sometimes issues crop up with rooftop solar. There might be soiling on a panel, so the panel might be dirty, or maybe the wires weren't connected properly and the panel isn't actually producing, or maybe a tree grew and now there's more shading. And Historically, this company basically had to wait until a customer called and said, I'm looking at my bill and it doesn't look like the solar is working as it used to be. Can you come and investigate? And they built, based on the smart meter data coming from the solar panels, a predictive model that basically says,…

AI assessment note: “they built, based on the smart meter data coming from the solar panels, a predictive model”

Q And do you have any numbers behind what we could potentially unlock?

A Uh, Well, yeah, in fact, there was a great RMI study that just came out, uh, very recently that looked at the PJM system and determined that you could unlock in excess of six gigawatts of extra capacity, and this is, I mean, PJM's a 90 gigawatt system, so six gigawatts is not nothing. That's pretty substantial, um, and, and this is for an investment of about a hundred million dollars that would return annually in excess of a billion dollars To rate payers. So, I mean, it's a huge return on investment without having to build a single new transmission line, right? And so when you think about the cost of doing that compared to the cost of simply putting better intelligence into the existing system, it's a massive return on investment.

AI assessment note: “determined that you could unlock in excess of six gigawatts of extra capacity”

Q And do you have any numbers behind what we could potentially unlock?

A Uh, Well, yeah, in fact, there was a great RMI study that just came out, uh, very recently that looked at the PJM system and determined that you could unlock in excess of six gigawatts of extra capacity, and this is, I mean, PJM's a 90 gigawatt system, so six gigawatts is not nothing. That's pretty substantial, um, and, and this is for an investment of about a hundred million dollars that would return annually in excess of a billion dollars To rate payers. So, I mean, it's a huge return on investment without having to build a single new transmission line, right? And so when you think about the cost of doing that compared to the cost of simply putting better intelligence into the existing system, it's a massive return on investment.

AI assessment note: “determined that you could unlock in excess of six gigawatts of extra capacity”

Q Let's talk about the scope of the problem. We have around a hundred million electric vehicle batteries that could get retired in the next decade. There's so much volume coming, so much expected volume of material that's going to need to be recycled over the next decade as lithium ion batteries overtake transportation and the grid. How do you wrap your arms around the scope of that challenge?

A Yeah, so that's a good rough estimate for, um, uh, what is on the horizon is the past 15 years plus the current growing waves of EVs that are being sold today ultimately make their way to retirement. But what a lot of people don't think about is, in addition, there's another major source of lithium ion batteries that need to be recycled now. Um, in addition to all the consumer electronics, you know, what do we do with all of our iPhone batteries? And laptop, and, and, you know, everywhere you turn, there's a lithium battery on a GoPro, you name it. But the real huge quantities that are coming today are from these gigafactories, um, and the gigafactory battery plants themselves, and that's in the form of production scrap. And so as battery components and these cathode, what are called cathode materials are cut and formed, a significant amount of that scrap is produced, and sometimes that's as high as 10 to 15% of the entire production of the plant. Especially in the earlier days as they're dialing in their processes. Um, so in the near term, this is a huge primary source of recyclable materials for companies like Aqua Metals. Um, and you know, from now until the end of the decade, uh, we, we think there's an estimated ten million tons or so of combined scrap material and end of life batteries that are really going to be ready to be recycled. And another data point that's interes…

AI assessment note: “we think there's an estimated ten million tons or so of combined scrap material”

Q want each of you to pick a number that illustrates a trend that you're seeing play out on the ground right now. We're going to be throwing around some big, big numbers. So let's try to make it a little bit more real, uh, For folks. Um, Catherine, what do you have? Any, any numbers jump out at you that kind of reflect what you were actually seeing play out?

A Yeah, so I have two numbers. One that someone gave me. So I reached out to Albert Chung, who is an analyst, a longtime analyst at Bloomberg. He's brilliant. And I said, what do you care the most about this report? And he said he was struck by the global offshore wind investment, reaching a record 76.7 billion dollars, which was an almost 80% increase in offshore wind. Now, there was a bit of a, of a decline from the onshore segment, but the offshore went, went up pretty quickly, and that's, that's remarkable, because the, as we have talked about on the show, offshore has had some issues, and it's good to see that there is still major investment in that sector.

AI assessment note: “global offshore wind investment, reaching a record 76.7 billion dollars”

Q control for these virtual power plants. Like, will there be one dominant point of control? And I'm thinking about two different scenarios, one where the utility controls the devices and the others as the aggregators are, are bidding these assets together and then bidding them into wholesale markets when eventual rules come together in wholesale markets. Um, do you see either of those points of control dominating in the future?

A It is interesting. The way that I think about the dichotomy that you set up is Will VPPs sell to utilities, right? If I'm Voltus, C-Power, Virtual Peaker, Swell, is my best bet to write a bilateral contract with a utility, or is my best bet to bid into a wholesale market that has implemented for a quarter two, two, two, two? I think when you have virtual power plants bidding into wholesale markets and earning a price that represents the value to the bulk power system, It's an incomplete price, because it's not taking a, it's not taking account of the value to the distribution grid. And when you have a utility contracting with a virtual power plant, a distribution utility, they can take into account what the value is of deferring investment in a substation, because they can flex the demand behind that substation, right? And so you're getting Not only the avoided, uh, peaking power cost at the wholesale level, you're also getting the value of deferring an investment in your distribution grid. So I think when it comes to, you know, where virtual power plants will succeed in the market right now, you have participation in wholesale markets because as, you know, a straightforward way to bid into basically an auction, and it's a bit more cumbersome to write a bilateral agreement utility by utility when we have thousands across the country, but that's where the real value lies. And so…

AI assessment note: “I think what you'll end up seeing is utilities recognizing the value”

Q So where do you think all the innovations in this industry are going to come from? Are they going to be largely technical? Are they going to be regulatory and enabling a lot of the, you know, the bidding of these projects? Will they be around customer acquisition and enrollment? Where do you see the most innovations happening right now?

A It really is coming from all angles. I think what's right in front of us from DOE's perspective is the regulatory innovation, I'll say, and I'll, I'll say a bit more about that, but I think what I'm most excited about is what's a little bit further afield from DOE is the, uh, consumer experience. So on the regulatory side, there is a long list Of improved regulatory measures that we're seeing utility regulators in particular adopt. So right now, a minority of states are doing integrated distribution system planning, where they're taking a look at, you know, what capacity they have on their distribution systems in relation to the bulk power system needs. But more and more states are adopting that. They're doing more DER adoption to understand those potential resources. You also have performance-based regulation or performance-based rate making, where commissions are realizing that the utility business model of compensating CapEx spending with maybe a seven to nine percent, uh, margin and passing through OPEX at cost to the consumer does not give utilities the right incentives to choose low-cost options such as VPPs, and so you have, you know, innovation happening within commissions who are, Uh, really better aligning utility financial incentives with, uh, what is optimal for the system. But when it comes to other innovations that are really going to transform this market, You ha…

AI assessment note: “It really is coming from all angles. I think what's right in front of us”

Q Yeah, well, let's just talk about where we think pricing will head this year and beyond. Are we on a continued downward trajectory?

A Yeah. Um, our, like our, our expectation at BNF is that prices will drop this year. Um, so drop about six percent. So, um, we averaged last year, just as a, as a pricing point, we averaged at about 139 dollars per kilowatt hour, um, on a volume average basis. So that includes all segments that we track. Um, and our expectation is that it'll drop about six percent this year. To about a 133 dollars per kilowatt hour. Um, and the way we actually do the, um, the forecasts or the expectation for battery prices for this year and the next three years is based on, um, firstly, like industry expectation on metal prices. Um, so this is taking into consideration the expectation that lithium, nickel, um, yeah, I think those are the two main ones, but cobalt as well. Um, though the prices for those major battery metals will come down, um, Um, and then essentially using those metal prices to inform and calculate what our near-term expectations for prices will be over the next three years. So generally, prices will come down, not as much as we saw between twenty-twenty-two and twenty-twenty-three, where it was a 14% drop, but still progressing downwards over the course of this year.

AI assessment note: “our expectation at BNF is that prices will drop this year.”